First of all I wish to thank all who gave me such wonderful and prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did speak to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the company
not to publish them for marketing considerations. He is to snail mail me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2 ohm
current capabilities is the more important criteria. That's all he would say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to Patrick,
Fabio, John, et al.
west

west wrote:

First of all I wish to thank all who gave me such wonderful and prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did speak to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the company
not to publish them for marketing considerations. He is to snail mail me
some tech specs, whatever that means.

You will get a letter in the mailbox. Remember them?

But Infinity is now part of Harmon Kardon, so this Infinity tech
may not be all that familiar with 1982? speakers.
Your'e lucky to have got a promise for official info for
something so old when the sales folks and techs these days are mostly
unable to comprehend items that are 20 years old, and still in use.

But I guess they must have had specs, including dB/W/M,
and impedance graphs, both of which you desperately need right now
to establish your amplifier needs.

The nice gentleman did say unlike a large voltage swing amp, let's say for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2 ohm
current capabilities is the more important criteria. That's all he would say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to Patrick,
Fabio, John, et al.
west

I have a pair of Reference 20 here for repair with seized voice coils on the
woofers,
and foam surrounds just about to crack and fall apart.
The foam coned dynamic tweeters have seen better days and someone
has poked holes on one tweeter foam dome, exposing the wool packing underneath.
So total re-drivering is what will be done using Peerless drivers, now easily
the equal or better than what Infinity lumbered us with 23 years ago.

Patrick Turner.

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the company
not to publish them for marketing considerations. He is to snail mail me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2 ohm
current capabilities is the more important criteria. That's all he would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best, cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was not all
that inefficient. It was about average. Tubes need not apply. Transistors
are the best option. Whilst not the IIIb, here is the impedance curve of the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg



--
Trevor Wilson
www.rageaudio.com.au

"Trevor Wilson" wrote in message
...

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail mail me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say
for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2 ohm
current capabilities is the more important criteria. That's all he would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best, cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was not all
that inefficient. It was about average. Tubes need not apply. Transistors
are the best option. Whilst not the IIIb, here is the impedance curve of
the
lineal descendent of the IIIb:


Hi Trevor,
The pre-owned ME-850 you advertise, states that it puts out a prodigious 75
amps! Can that be true? Also, why don't most other amps advertise their max
output current?
west
BTW: your Kappa9 link (below) is dorked.
www.rageaudio.com.au/kapp9.jpg



--
Trevor Wilson
www.rageaudio.com.au

"west" wrote in message
...


Hi Trevor,
The pre-owned ME-850 you advertise, states that it puts out a prodigious
75
amps! Can that be true?

**Indeed it is true. It can deliver those 75 Amps for a significant period
of time, too (many seconds). It has a 2,500 VA power transformer and
265,000uF to back it up.

Also, why don't most other amps advertise their max
output current?

**Many do. However, their maximum output current is usually limited to a few
milliseconds.

west
BTW: your Kappa9 link (below) is dorked.
www.rageaudio.com.au/kapp9.jpg

**D'OH! (smacks head) Try this one:

www.rageaudio.com.au/kappa9.jpg


--
Trevor Wilson
www.rageaudio.com.au

OK, now I got the point. It is true that SS is the more sensible way to go
to get many amps on 2 Ohms. But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to provide
the performances You need, a tubed amp shall use a "large" number of low
internal resistance power valves, like series-pass triodes (6S33, 6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s. I'm not in favour of
OTLs, particularly if You need 2 Ohms capability. A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS easier
designing a suitable OPT to 2 Ohms. This load can be further reduced using
more (or different) tubes.
All "giant" tubes like GM-70 are definitely not Your choice, 'cause they all
need some 10K as anode load, and this makes the design of the OPT a real
mess without any advantage.
My only modification with respect to Patrick's designs, just due to Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.
Unfortunately they all have top caps, which are not the best for safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very low
impedance which works at about 250V B+, it could be a good one but You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

Ciao

Fabio


"west" ha scritto nel messaggio
...
First of all I wish to thank all who gave me such wonderful and prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the company
not to publish them for marketing considerations. He is to snail mail me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2 ohm
current capabilities is the more important criteria. That's all he would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to Patrick,
Fabio, John, et al.
west

"Fabio Berutti" wrote in message
...
OK, now I got the point. It is true that SS is the more sensible way to
go to get many amps on 2 Ohms. But this is a tubed group and the sand
state of matter is considered only to make beaches, not amps.

**There is FAR more sand in any tube, than in even the very largest of
transistors.


--
Trevor Wilson
www.rageaudio.com.au

Trevor Wilson wrote:

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the company
not to publish them for marketing considerations. He is to snail mail me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2 ohm
current capabilities is the more important criteria. That's all he would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best, cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was not all
that inefficient. It was about average. Tubes need not apply. Transistors
are the best option. Whilst not the IIIb, here is the impedance curve of the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg

--
Trevor Wilson
www.rageaudio.com.au

I cannot see any evidence of you having given West any good advice.

The impedance curve is most misleading.
There is no identification of exactly what speaker it is for,
and is it impedance?, or phase shift?

If it is impedance, lowest Z = 0.8 ohms at 25Hz,
which seems utterly impossible,
since the DCR would surely make Z minimum at least = DCR, no?

You went on to say that a 20 watt Rotel or NAD amp would be OK
and that to use tubes would be unwise.

This is utter BS.

Even if the curves were the right ones for Infinity RSIIIA, and
i don't see any proof that they are, then any 50 watt tube
amp with a 4 ohm outlet match would
be OK, since Z averages 4 ohms from 80Hz to 6 kHz.
Z averages 6 ohms between 120Hz and 1 kHz, the band which has the most signal.
Tubes are more than capable of coping with load mismatches
where the amount of energy is low, ie, below 80 Hz and above 6 kHz.
Many tube amps have their maximum power output
for a load below the nominated load for a given load match.
So that where one uses the 4 ohm outlet, often maximum power at 3 ohms
at clipping is 30% above the maximum 4 ohm power at clipping.
Below 3 ohms the max power may well roll off, but it
certainly does not mean that there would not be ample
power, voltage and current to cope with the low load
at extremes of bandwidth.

A pair of KT88 is ideally set up to
make 35w to 8 ohms, lots of class A, 55 wattsAB to 4 ohms, including about 15
watts
of class A, and perhaps 65 watts into less than 3 ohms, with not much class A.

The 20 watt level is between about 1 ohm and 12 ohms.

And this would be when using the 8 ohm outlet.
If there was a 4 ohm outlet, you'd get all the above powers at loads 1/2 what I
said,
so 65 watts absolute max into 1.5 ohms, which is 6.6 amps rms.

Now if you were to take advantage of a high current capable amp using SS,
and you were abole to use 75 peak amp capability, we are talking
53 A rms, which is 79.5 vrms into 1.5 ohms which is 4,213 watts.
You'd bleed from the ears.
Huh? But I thought we could do 75 peak amps?
Unless the rails of the amp were up around +/- 125v, forget
ever seeing anything like 75 amps, and what's more,
you won't EVER EVER need such current ablity.


Unless anyone can get the exact and actual impedance
curve the factory prepared all those years ago and for the exact model number,
there is only one proper way to find out how hard it is to drive Infinity RSIIIA

is to measure the speaker impedance properly,
and do the measurments on peak voltage and peak
current for your loudest
listening using a borrowed powerful SS amp, of least 200W.

The impedance curve which Trevor has guided us to looks
quite wrong.

But if it were correct, there is a worry that if a tube amp has Ro = 0.5 ohms,
then where Z 3 ohms the response would fall and a dB of response
would go missing above 2 kHz, because Z averages 2 ohms above 2 kHz.
This would be the case with some tube amps using their "8 ohm" outlet.
But the 4 ohm outlet will have Ro = half the 8 ohm outlet Ro,
or Ro = 0.25 ohms, so negligible losses will occur at HF.

People should always remember Trevor Wilson strongly favours
solid state amps, which he trades, especially those made by ME Technology, a
former Oz company
which went into liquidation over a year ago.

Now Trevor has said a 20 watt Rotel or NAD would do.
We can assume that means 20 watts into 4 ohms, which is only 2.2 amps.

I'd actually be surprised if anyone needed more than 50 watts, just to be sure.

But before anyone here can talk any sense, lets see the measurements
of the RSIIIA speakers.



Patrick Turner.

Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way to go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.




But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to provide
the performances You need, a tubed amp shall use a "large" number of low
internal resistance power valves, like series-pass triodes (6S33, 6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.



I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.



A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS easier
designing a suitable OPT to 2 Ohms. This load can be further reduced using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause they all
need some 10K as anode load, and this makes the design of the OPT a real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very low
impedance which works at about 250V B+, it could be a good one but You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60 approx.

Patrick Turner.



Ciao

Fabio

"west" ha scritto nel messaggio
...
First of all I wish to thank all who gave me such wonderful and prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the company
not to publish them for marketing considerations. He is to snail mail me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2 ohm
current capabilities is the more important criteria. That's all he would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to Patrick,
Fabio, John, et al.
west

After some chatting I guess our Aussie fellow has got the right idea. My
concern was to find a power tubes setup which could allow a very low
reflected load, in order to ease as much as possible the design (and
construction) of a suitable output transformer, say a 600 Ohm to 3 Ohms.
This can be surely done with good sounding audio tubes like the 6550, which
is still in production, doesn't cost an arm and a leg and does not have
dangerous caps. But I need to support my ideas (be them correct or not), so
I'd go for a multiple PPP using the EL36, which in fact has a good
reputation for linearity, a low impedance, has an ordinary Octal base
instead of a Magnoval.. briefly it is a better choice than EL500 or EL509.
Advantages with respect to KT88s or eq? Lower voltage, lower power - it is
easier to set 10-12 of them in parallel PP in order to reach the
abovementioned 600R load without ending up with a 300W unit, maybe using a
lower voltage operating point than the one suggested by Pat (thus increasing
the fraction of Class A operation). I'd say 250V a-k, bias -25V and 60 mA
per tube.

Ciao

Fabio

"Patrick Turner" ha scritto nel messaggio
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.




But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to provide
the performances You need, a tubed amp shall use a "large" number of low
internal resistance power valves, like series-pass triodes (6S33, 6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.



I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.



A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause they
all
need some 10K as anode load, and this makes the design of the OPT a real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very low
impedance which works at about 250V B+, it could be a good one but You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.



Ciao

Fabio

"west" ha scritto nel messaggio
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say
for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2
ohm
current capabilities is the more important criteria. That's all he
would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.
west

Fabio Berutti wrote:

After some chatting I guess our Aussie fellow has got the right idea. My
concern was to find a power tubes setup which could allow a very low
reflected load, in order to ease as much as possible the design (and
construction) of a suitable output transformer, say a 600 Ohm to 3 Ohms.

Ah, well I guess I will have to parle with my learned Cognescenti Italiano
for a few moments longer.

6AS7G are not best used without an OPT.

But each triode within 6AS7 has Ra = 280 ohms, so for the 2 sections Ra = 140
ohms.
With a quad of 6AS7G in PP we would have an Ra-a = 140 ohms.

My rough calculation is that each tube is set up with fixed bias and Ia = 70 mA,

Ea = 150v so Pad for each whole tube is 10.5 watts, which is comfortable for
these tubes.

The RLa-a would have to be 600 ohms for the quad, and
class A1 PP swing is about 140 vrms to give 32 watts, of which about 8w is class
A.

I would bet you at least a fine bottle of Italy's best red wine that
such an amp would give a superb performance compared to
the same tubes used without any OPT.

The Ro without any NFB with a 600 : 3 ohm OPT would be
( 3 / 600 ) x Ra-a = 0.7 ohms, plus the winding resistance so total = 0.8 ohms,
still rather high and giving a damping factor with a match to 3 ohms of only
3.75,
a figure I'd like to see tripled, so it would appear at least
10 db of global NDB would have to be applied, and then Ro would about fall to
0.3 approx.
The 10 db of NFB would also tidy up some of the AB thd, and methinks the amp
would be quite fair, but still would not at least measure as well
as a quad of KT88, or KT90, the latter being the better trioded tubes by far.



This can be surely done with good sounding audio tubes like the 6550, which
is still in production, doesn't cost an arm and a leg and does not have
dangerous caps. But I need to support my ideas (be them correct or not), so
I'd go for a multiple PPP using the EL36, which in fact has a good
reputation for linearity, a low impedance, has an ordinary Octal base
instead of a Magnoval..

Alas they are no longer made, and they have a top cap.

But, should anyone of you find a cache of 6FW5, you will have a superb
tube which I measured as being exactly equal to EL36/6CM5, but
which has no top cap, but does have a standard octal base, but slightly
different pin out
to EL34, 6550 etc.

Of all the TV line output tetrodes, the EL36/6CM5 ( or 6FW5 ) is the only
type I think is worth trioding, ( or ultralining with 50% taps ).
6CD6 is another one around, but less common.
I am not much of a fan of 6DQ6.



briefly it is a better choice than EL500 or EL509.

I think EL36/6CM5 is better too. The max current in a 6CM5 is 0.5 amps; same as
larger tubes; they gave it a good cathode like its bigger audio cousins.

In triode, EL36 has Ra = 600 ohms only, 1/2 EL34, so
you get a wider Vswing even though Ea = only +375v, which is about
the maximum for triode that one should use.


Advantages with respect to KT88s or eq? Lower voltage, lower power - it is
easier to set 10-12 of them in parallel PP in order to reach the
abovementioned 600R load without ending up with a 300W unit, maybe using a
lower voltage operating point than the one suggested by Pat (thus increasing
the fraction of Class A operation). I'd say 250V a-k, bias -25V and 60 mA
per tube.

With a dozen KT88/6550 in triode and with RLa-a = 600 ohms,
RL a-a for a pair is equivalent to 3,600 ohms, about the lowest
RL one should ever use in any PP config of KT88/6550.
Only 22 watts AB1 is possible for a pair in triode with 3.6k and Ea = 400v,
or 20 watts including OPT losses, so with a dozen you get 6 x 20 = 120 watts in
triodeAB1.

Far better is to raise Ea to 500v, and use RLa-a of 6.66k a-a.
This will give 30 watts, or around 180 watts with a dozen in triode.

The % of class A is better as well as efficiency, and thd will be lower.

The load for 12 = 1.1k and if the secondary is 3 ohms the
Ro = ( 3 / 1,100 ) x Ra-a = 0.0027 x ( 2,200 / 6 ) = 1.0 ohms.

Should we wish for class A power with Ea = +500v and triodes,
RL a-a could be 12k per pair, or 2k for the 12 triodes,
and po = 20 watts but very near all class A,
even if Pd is conservatively low.
Then the Ro for 12 tubes is ( 3 / 2,000 ) x ( 2,200 / 6 ) = 0.55 ohms,
or around 0.6 ohms with an OPT with 10% winding losses.

This gives DF = 5, with 3 ohms, and so to improve the DF
we would still darn well need to apply some NFB, or set up the tubes
with an evan higher RLa-a of say 20k a-a.

With loads of 6 ohms connected to an amp set for 3 ohms, the
DF doubles, and is fine.

Regards,

Patrick Turner.




Ciao

Fabio

"Patrick Turner" ha scritto nel messaggio
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.




But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to provide
the performances You need, a tubed amp shall use a "large" number of low
internal resistance power valves, like series-pass triodes (6S33, 6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.



I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.



A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause they
all
need some 10K as anode load, and this makes the design of the OPT a real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very low
impedance which works at about 250V B+, it could be a good one but You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.



Ciao

Fabio

"west" ha scritto nel messaggio
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say
for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2
ohm
current capabilities is the more important criteria. That's all he
would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.
west

Mr. Turner,

It seems that at the good end the most "sensible" unit, combining normal
tubes with no exposed hi-V caps, low output impedance, high current
capability, a correct amount of NFB and so on, looks very similar to Yours.
This is what I call "evolutive convergence": in front of a given problem,
and given the same constraints, the majority of a group of well-trained
engineers should get to similar results.
Nevertheless I still like to design my own amps for my own entertainment,
just to make them different from anything else.. and if they're worse, well,
I'll just have a glass of my grandpa's red wine... it really helps lowering
odd harmonics. I have lots of fun imagining "strange" solutions using
Unobtainium tubes connected in unnatural ways (BTW, do polarities revert
"down under"?)
But shortly, I think that "any" good amp (professionally engineered) shoud
routinely take into account "difficult" loads; therefore, I suppose that
even Your smaller "85" (it's the 85 the one with "only" 4 tubes per channel,
isn't it?) should get the work done just fine with the Infinity III or with
any other LS; some more toobs are needed only if West happens to live in the
NASA Saturn Five Assembly Building .. maybe this is why he bought the
Infinities...

Have a nice day

Fabio

PS: I agree, the 6AS7 is a fairly good tube when used with a suitable OPT.
I made a small amp for a friend using the 2 halves in PP and a 2K5 load and
it sounds really brilliant with his old 94 dB bookshelf LS.

"Patrick Turner" ha scritto nel messaggio
...


Fabio Berutti wrote:

After some chatting I guess our Aussie fellow has got the right idea. My
concern was to find a power tubes setup which could allow a very low
reflected load, in order to ease as much as possible the design (and
construction) of a suitable output transformer, say a 600 Ohm to 3 Ohms.

Ah, well I guess I will have to parle with my learned Cognescenti Italiano
for a few moments longer.

6AS7G are not best used without an OPT.

But each triode within 6AS7 has Ra = 280 ohms, so for the 2 sections Ra =
140
ohms.
With a quad of 6AS7G in PP we would have an Ra-a = 140 ohms.

My rough calculation is that each tube is set up with fixed bias and Ia =
70 mA,

Ea = 150v so Pad for each whole tube is 10.5 watts, which is comfortable
for
these tubes.

The RLa-a would have to be 600 ohms for the quad, and
class A1 PP swing is about 140 vrms to give 32 watts, of which about 8w is
class
A.

I would bet you at least a fine bottle of Italy's best red wine that
such an amp would give a superb performance compared to
the same tubes used without any OPT.

The Ro without any NFB with a 600 : 3 ohm OPT would be
( 3 / 600 ) x Ra-a = 0.7 ohms, plus the winding resistance so total = 0.8
ohms,
still rather high and giving a damping factor with a match to 3 ohms of
only
3.75,
a figure I'd like to see tripled, so it would appear at least
10 db of global NDB would have to be applied, and then Ro would about fall
to
0.3 approx.
The 10 db of NFB would also tidy up some of the AB thd, and methinks the
amp
would be quite fair, but still would not at least measure as well
as a quad of KT88, or KT90, the latter being the better trioded tubes by
far.



This can be surely done with good sounding audio tubes like the 6550,
which
is still in production, doesn't cost an arm and a leg and does not have
dangerous caps. But I need to support my ideas (be them correct or not),
so
I'd go for a multiple PPP using the EL36, which in fact has a good
reputation for linearity, a low impedance, has an ordinary Octal base
instead of a Magnoval..

Alas they are no longer made, and they have a top cap.

But, should anyone of you find a cache of 6FW5, you will have a superb
tube which I measured as being exactly equal to EL36/6CM5, but
which has no top cap, but does have a standard octal base, but slightly
different pin out
to EL34, 6550 etc.

Of all the TV line output tetrodes, the EL36/6CM5 ( or 6FW5 ) is the only
type I think is worth trioding, ( or ultralining with 50% taps ).
6CD6 is another one around, but less common.
I am not much of a fan of 6DQ6.



briefly it is a better choice than EL500 or EL509.

I think EL36/6CM5 is better too. The max current in a 6CM5 is 0.5 amps;
same as
larger tubes; they gave it a good cathode like its bigger audio cousins.

In triode, EL36 has Ra = 600 ohms only, 1/2 EL34, so
you get a wider Vswing even though Ea = only +375v, which is about
the maximum for triode that one should use.


Advantages with respect to KT88s or eq? Lower voltage, lower power - it
is
easier to set 10-12 of them in parallel PP in order to reach the
abovementioned 600R load without ending up with a 300W unit, maybe using
a
lower voltage operating point than the one suggested by Pat (thus
increasing
the fraction of Class A operation). I'd say 250V a-k, bias -25V and 60
mA
per tube.

With a dozen KT88/6550 in triode and with RLa-a = 600 ohms,
RL a-a for a pair is equivalent to 3,600 ohms, about the lowest
RL one should ever use in any PP config of KT88/6550.
Only 22 watts AB1 is possible for a pair in triode with 3.6k and Ea =
400v,
or 20 watts including OPT losses, so with a dozen you get 6 x 20 = 120
watts in
triodeAB1.

Far better is to raise Ea to 500v, and use RLa-a of 6.66k a-a.
This will give 30 watts, or around 180 watts with a dozen in triode.

The % of class A is better as well as efficiency, and thd will be lower.

The load for 12 = 1.1k and if the secondary is 3 ohms the
Ro = ( 3 / 1,100 ) x Ra-a = 0.0027 x ( 2,200 / 6 ) = 1.0 ohms.

Should we wish for class A power with Ea = +500v and triodes,
RL a-a could be 12k per pair, or 2k for the 12 triodes,
and po = 20 watts but very near all class A,
even if Pd is conservatively low.
Then the Ro for 12 tubes is ( 3 / 2,000 ) x ( 2,200 / 6 ) = 0.55 ohms,
or around 0.6 ohms with an OPT with 10% winding losses.

This gives DF = 5, with 3 ohms, and so to improve the DF
we would still darn well need to apply some NFB, or set up the tubes
with an evan higher RLa-a of say 20k a-a.

With loads of 6 ohms connected to an amp set for 3 ohms, the
DF doubles, and is fine.

Regards,

Patrick Turner.

There has been a long discussion on OPT load, in this thread. Since the
problem is "how to feed a 2 ohms LS", I translated it into "what is the
power tubes setup that would ease the transformer design", but, being no
transformer technician, I probably got too far looking for a 600 ohms to 3
ohms matching.
Now my question to the braver RATs is: how far can a transforming ratio go,
before it becomes a mess to wind anything suitable?

Ciao

Fabio

"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number of low
internal resistance power valves, like series-pass triodes (6S33, 6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause they
all
need some 10K as anode load, and this makes the design of the OPT a real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very low
impedance which works at about 250V B+, it could be a good one but You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.

Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat 2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp #2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

"west" wrote in message
. ..
"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way
to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number of
low
internal resistance power valves, like series-pass triodes (6S33,
6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn
ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause
they
all
need some 10K as anode load, and this makes the design of the OPT a
real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to
Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for
safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very
low
impedance which works at about 250V B+, it could be a good one but
You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the
amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.
Patrick,
Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat
2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity
RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp #2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

Why are we focusing only on 2 ohms? This is the impedance for a specific
frequency. One must look at the entire impedance curve for a particular
speaker as Professor P. stated. I'm starting to think that this curve is
more valuable than knowing a speakers SPL(db)/2.83V/M. Fabio, do you think
you have the skills to dissect a single input speaker system and surgically
separate the woofers, mids and tweeter to make a bi or even tri amp system
when there was none to begin with? Seems like a lot of math involved but no
more than 9th grade algebra. It's been a while since I figured polar to
rectangular notation are its reverse. With impedance you have to have an
angle with it. Zero degrees of course is pure resistive,
negative-capacitive, and positive- inductive. I think it can be done. Then
you have an amp(s) custom tailored to your needs. However as they say in
Perugia ... "Tra il dire e il fare, c'e`di mezzo il mare...talk is cheap."
Ciao,
west

IMHO a good amp is a good one when it is good for everything: classic and
rock, Audiostatics and Infinity. This means it has to provide a generous
power (but I suppose 100-150W are more than enough), a good bandwidth, a
good damping factor etc, and it has to do it reliably, continuously, without
costing an arm and a leg and without forcing You to strengthen the floor to
carry its weight. Bi- or tri-amping? Maybe. If You already have many amps
You can make an active crossover, then connect one amp to each single
loudspeaker and save the money needed for a new big one, but it's abit
over-complicated for my taste. I think that You should get a plain, simple,
good 100W tubed amp (say a Citation 4 if You can borrow one, just not to
disturb Patrick everytime) and hear how it works in Your house, with Your
speakers and Your taste. Measurements do not necessarily tell all the
truth. Then go on accordingly, there's no substitute for experience.

Ciao

Fabio

BTW: have You been in Perugia during Umbria Jazz?


"west" ha scritto nel messaggio
. ..

"west" wrote in message
. ..
"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way
to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number of
low
internal resistance power valves, like series-pass triodes (6S33,
6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn
ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause
they
all
need some 10K as anode load, and this makes the design of the OPT a
real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to
Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for
safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very
low
impedance which works at about 250V B+, it could be a good one but
You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the
amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.
Patrick,
Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build
them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat
2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity
RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp #2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

Why are we focusing only on 2 ohms? This is the impedance for a specific
frequency. One must look at the entire impedance curve for a particular
speaker as Professor P. stated. I'm starting to think that this curve is
more valuable than knowing a speakers SPL(db)/2.83V/M. Fabio, do you think
you have the skills to dissect a single input speaker system and
surgically
separate the woofers, mids and tweeter to make a bi or even tri amp system
when there was none to begin with? Seems like a lot of math involved but
no
more than 9th grade algebra. It's been a while since I figured polar to
rectangular notation are its reverse. With impedance you have to have an
angle with it. Zero degrees of course is pure resistive,
negative-capacitive, and positive- inductive. I think it can be done. Then
you have an amp(s) custom tailored to your needs. However as they say in
Perugia ... "Tra il dire e il fare, c'e`di mezzo il mare...talk is cheap."
Ciao,
west

west wrote:

"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number of low
internal resistance power valves, like series-pass triodes (6S33, 6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause they
all
need some 10K as anode load, and this makes the design of the OPT a real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very low
impedance which works at about 250V B+, it could be a good one but You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.

Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat 2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp #2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

I'd probably offer them each the same amp
able to make 100 watts into either 6 ohms or 3.8 ohms.
due to at least two load matches possible.

If a dude has speakers that have a LF aor HF band which dips to 2 ohms, my
amps won't mind it, because some loading where RL = 1.2 the rated
3.8 or 6 ohm outlets is OK, just as it won't hurt
anything to use twice the load of the outlet.

But If the speaker averages 2 ohms across the whole band,
I would give the customer an OPT which
can be matched to suit, so when his mates come around
and connect their gee whiz SS up, they smoke, but
not my tubes.

I'd still wanna know the Z curve before I build any amps.

The 300 watters are the exception.

They have """oddles""" of headroom.

And can drive any load period.

The OPT can be configured for 5 ohms, or 2.2 ohms.

If one only wants 100 watts, and one has 6 ohm speakers,
the 2.2 ohm connection is used and 8 of the 12 output tubes can be removed.
The 4 left will cope fine.

Patrick Turner.

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say
for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2
ohm
current capabilities is the more important criteria. That's all he
would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of
prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best, cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was not
all
that inefficient. It was about average. Tubes need not apply. Transistors
are the best option. Whilst not the IIIb, here is the impedance curve of
the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg

--
Trevor Wilson
www.rageaudio.com.au

I cannot see any evidence of you having given West any good advice.

**Then you're not paying attention to logic and common-sense. I have
extensive experisnce will many of the old Infinity speakers. I know the
RSIIIb very well indeed. They are particularly poorly suited to use with
tube amps. In fact, they are particularly unsuited to use with *any*
amplifier not capable of delivering prodigious current and providing good
damping, at LF. Tube amps, whilst having some desirable charteristics, are
not real good at these things.


The impedance curve is most misleading.

**No, it is not. It is exactly what you see.

There is no identification of exactly what speaker it is for,
and is it impedance?, or phase shift?

**I thought I explained that. It is impedance. LMS displays both impedance
and pahse shift, or either one. I deleted the pahse shift information in the
interests of clarity.


If it is impedance, lowest Z = 0.8 ohms at 25Hz,
which seems utterly impossible,
since the DCR would surely make Z minimum at least = DCR, no?

**Undoubtedly.


You went on to say that a 20 watt Rotel or NAD amp would be OK
and that to use tubes would be unwise.

This is utter BS.

**No, Patrick, it is an inescapable fact. I specified those two types of
amps, since in the case of Rotel, no current limiters are used. Older style
NADs also lack current limiters.


Even if the curves were the right ones for Infinity RSIIIA, and
i don't see any proof that they are,

**I stated, very clearly, that the curve was for the RSIIIb's lineal
descendent, the Kappa 9. Did you not read that part?

then any 50 watt tube
amp with a 4 ohm outlet match would
be OK, since Z averages 4 ohms from 80Hz to 6 kHz.

**I never suggested that a decent tube amp would not work. It is just that
the sonic compromise would be very severe.

Z averages 6 ohms between 120Hz and 1 kHz, the band which has the most
signal.
Tubes are more than capable of coping with load mismatches
where the amount of energy is low, ie, below 80 Hz and above 6 kHz.
Many tube amps have their maximum power output
for a load below the nominated load for a given load match.
So that where one uses the 4 ohm outlet, often maximum power at 3 ohms
at clipping is 30% above the maximum 4 ohm power at clipping.
Below 3 ohms the max power may well roll off, but it
certainly does not mean that there would not be ample
power, voltage and current to cope with the low load
at extremes of bandwidth.

A pair of KT88 is ideally set up to
make 35w to 8 ohms, lots of class A, 55 wattsAB to 4 ohms, including about
15
watts
of class A, and perhaps 65 watts into less than 3 ohms, with not much
class A.

The 20 watt level is between about 1 ohm and 12 ohms.

And this would be when using the 8 ohm outlet.
If there was a 4 ohm outlet, you'd get all the above powers at loads 1/2
what I
said,
so 65 watts absolute max into 1.5 ohms, which is 6.6 amps rms.

Now if you were to take advantage of a high current capable amp using SS,
and you were abole to use 75 peak amp capability, we are talking
53 A rms, which is 79.5 vrms into 1.5 ohms which is 4,213 watts.
You'd bleed from the ears.
Huh? But I thought we could do 75 peak amps?
Unless the rails of the amp were up around +/- 125v, forget
ever seeing anything like 75 amps, and what's more,
you won't EVER EVER need such current ablity.


Unless anyone can get the exact and actual impedance
curve the factory prepared all those years ago and for the exact model
number,
there is only one proper way to find out how hard it is to drive Infinity
RSIIIA

**Just a reminder, Patrick, we're discussing the RSIIIb, not the RSIIIa. Not
that there is much difference. There is, however, a significant difference
between the RSIII and the RSIIIa.


is to measure the speaker impedance properly,
and do the measurments on peak voltage and peak
current for your loudest
listening using a borrowed powerful SS amp, of least 200W.

The impedance curve which Trevor has guided us to looks
quite wrong.

**Then YOU present a curve YOU have mneasured for the Kappa 9. I'll wait.


But if it were correct, there is a worry that if a tube amp has Ro = 0.5
ohms,
then where Z 3 ohms the response would fall and a dB of response
would go missing above 2 kHz, because Z averages 2 ohms above 2 kHz.
This would be the case with some tube amps using their "8 ohm" outlet.
But the 4 ohm outlet will have Ro = half the 8 ohm outlet Ro,
or Ro = 0.25 ohms, so negligible losses will occur at HF.

People should always remember Trevor Wilson strongly favours
solid state amps, which he trades, especially those made by ME Technology,
a
former Oz company
which went into liquidation over a year ago.

Now Trevor has said a 20 watt Rotel or NAD would do.
We can assume that means 20 watts into 4 ohms, which is only 2.2 amps.

I'd actually be surprised if anyone needed more than 50 watts, just to be
sure.

But before anyone here can talk any sense, lets see the measurements
of the RSIIIA speakers.

**Fair enough. I'll see what I can manage. They turn up from time to time.


--
Trevor Wilson
www.rageaudio.com.au

Fabio Berutti wrote:

Mr. Turner,

It seems that at the good end the most "sensible" unit, combining normal
tubes with no exposed hi-V caps, low output impedance, high current
capability, a correct amount of NFB and so on, looks very similar to Yours.
This is what I call "evolutive convergence": in front of a given problem,
and given the same constraints, the majority of a group of well-trained
engineers should get to similar results.
Nevertheless I still like to design my own amps for my own entertainment,
just to make them different from anything else.. and if they're worse, well,
I'll just have a glass of my grandpa's red wine... it really helps lowering
odd harmonics. I have lots of fun imagining "strange" solutions using
Unobtainium tubes connected in unnatural ways (BTW, do polarities revert
"down under"?)

If the music was recorded in the nothern hemisphere, and the CD not
correctly orientated during mastering, then we have to stand on our heads to
listen,
since otherwise the music comes out backwards, which unfazes a lotta folks.....


But shortly, I think that "any" good amp (professionally engineered) shoud
routinely take into account "difficult" loads; therefore, I suppose that
even Your smaller "85" (it's the 85 the one with "only" 4 tubes per channel,
isn't it?) should get the work done just fine with the Infinity III or with
any other LS; some more toobs are needed only if West happens to live in the
NASA Saturn Five Assembly Building .. maybe this is why he bought the
Infinities...

85 watts from 4 x 6550 would suit 90% of ppl.
CJ uses 4 to get 140 watts.
They can't resist the extra watt numbers because dudes think watts,
not fidelity.

90% of dudes use maybe 2 watts maximum average,
or their wives will leave them.




Patrick Turner



Have a nice day

Fabio

PS: I agree, the 6AS7 is a fairly good tube when used with a suitable OPT.
I made a small amp for a friend using the 2 halves in PP and a 2K5 load and
it sounds really brilliant with his old 94 dB bookshelf LS.

"Patrick Turner" ha scritto nel messaggio
...


Fabio Berutti wrote:

After some chatting I guess our Aussie fellow has got the right idea. My
concern was to find a power tubes setup which could allow a very low
reflected load, in order to ease as much as possible the design (and
construction) of a suitable output transformer, say a 600 Ohm to 3 Ohms.

Ah, well I guess I will have to parle with my learned Cognescenti Italiano
for a few moments longer.

6AS7G are not best used without an OPT.

But each triode within 6AS7 has Ra = 280 ohms, so for the 2 sections Ra =
140
ohms.
With a quad of 6AS7G in PP we would have an Ra-a = 140 ohms.

My rough calculation is that each tube is set up with fixed bias and Ia =
70 mA,

Ea = 150v so Pad for each whole tube is 10.5 watts, which is comfortable
for
these tubes.

The RLa-a would have to be 600 ohms for the quad, and
class A1 PP swing is about 140 vrms to give 32 watts, of which about 8w is
class
A.

I would bet you at least a fine bottle of Italy's best red wine that
such an amp would give a superb performance compared to
the same tubes used without any OPT.

The Ro without any NFB with a 600 : 3 ohm OPT would be
( 3 / 600 ) x Ra-a = 0.7 ohms, plus the winding resistance so total = 0.8
ohms,
still rather high and giving a damping factor with a match to 3 ohms of
only
3.75,
a figure I'd like to see tripled, so it would appear at least
10 db of global NDB would have to be applied, and then Ro would about fall
to
0.3 approx.
The 10 db of NFB would also tidy up some of the AB thd, and methinks the
amp
would be quite fair, but still would not at least measure as well
as a quad of KT88, or KT90, the latter being the better trioded tubes by
far.



This can be surely done with good sounding audio tubes like the 6550,
which
is still in production, doesn't cost an arm and a leg and does not have
dangerous caps. But I need to support my ideas (be them correct or not),
so
I'd go for a multiple PPP using the EL36, which in fact has a good
reputation for linearity, a low impedance, has an ordinary Octal base
instead of a Magnoval..

Alas they are no longer made, and they have a top cap.

But, should anyone of you find a cache of 6FW5, you will have a superb
tube which I measured as being exactly equal to EL36/6CM5, but
which has no top cap, but does have a standard octal base, but slightly
different pin out
to EL34, 6550 etc.

Of all the TV line output tetrodes, the EL36/6CM5 ( or 6FW5 ) is the only
type I think is worth trioding, ( or ultralining with 50% taps ).
6CD6 is another one around, but less common.
I am not much of a fan of 6DQ6.



briefly it is a better choice than EL500 or EL509.

I think EL36/6CM5 is better too. The max current in a 6CM5 is 0.5 amps;
same as
larger tubes; they gave it a good cathode like its bigger audio cousins.

In triode, EL36 has Ra = 600 ohms only, 1/2 EL34, so
you get a wider Vswing even though Ea = only +375v, which is about
the maximum for triode that one should use.


Advantages with respect to KT88s or eq? Lower voltage, lower power - it
is
easier to set 10-12 of them in parallel PP in order to reach the
abovementioned 600R load without ending up with a 300W unit, maybe using
a
lower voltage operating point than the one suggested by Pat (thus
increasing
the fraction of Class A operation). I'd say 250V a-k, bias -25V and 60
mA
per tube.

With a dozen KT88/6550 in triode and with RLa-a = 600 ohms,
RL a-a for a pair is equivalent to 3,600 ohms, about the lowest
RL one should ever use in any PP config of KT88/6550.
Only 22 watts AB1 is possible for a pair in triode with 3.6k and Ea =
400v,
or 20 watts including OPT losses, so with a dozen you get 6 x 20 = 120
watts in
triodeAB1.

Far better is to raise Ea to 500v, and use RLa-a of 6.66k a-a.
This will give 30 watts, or around 180 watts with a dozen in triode.

The % of class A is better as well as efficiency, and thd will be lower.

The load for 12 = 1.1k and if the secondary is 3 ohms the
Ro = ( 3 / 1,100 ) x Ra-a = 0.0027 x ( 2,200 / 6 ) = 1.0 ohms.

Should we wish for class A power with Ea = +500v and triodes,
RL a-a could be 12k per pair, or 2k for the 12 triodes,
and po = 20 watts but very near all class A,
even if Pd is conservatively low.
Then the Ro for 12 tubes is ( 3 / 2,000 ) x ( 2,200 / 6 ) = 0.55 ohms,
or around 0.6 ohms with an OPT with 10% winding losses.

This gives DF = 5, with 3 ohms, and so to improve the DF
we would still darn well need to apply some NFB, or set up the tubes
with an evan higher RLa-a of say 20k a-a.

With loads of 6 ohms connected to an amp set for 3 ohms, the
DF doubles, and is fine.

Regards,

Patrick Turner.

Fabio Berutti wrote:

There has been a long discussion on OPT load, in this thread. Since the
problem is "how to feed a 2 ohms LS", I translated it into "what is the
power tubes setup that would ease the transformer design", but, being no
transformer technician, I probably got too far looking for a 600 ohms to 3
ohms matching.
Now my question to the braver RATs is: how far can a transforming ratio go,
before it becomes a mess to wind anything suitable?

One can use one turn of foil to make a secondary winding
instead of say 48 turns used for say 8 ohms.
That's a match from say 5k to 0.0035 ohms.
But losses and leakage would make that difficult,
but to get to 1 ohm from 5k a-a is simple.
instead of 48 turns used for 8 ohms, 3 x 16 turn windings of the same wire
in the same layer would give a match to 0.889 ohms, if each winding was in
parallel.

Patrick Turner.




Ciao

Fabio

"Patrick Turner" wrote in message
...


west wrote:

"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible
way to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number of
low
internal resistance power valves, like series-pass triodes (6S33,
6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of
NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8
ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further
reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn
ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause
they
all
need some 10K as anode load, and this makes the design of the OPT a
real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to
Your
peculiar needs, would be to use TV output tubes, say EL509 or
36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for
safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very
low
impedance which works at about 250V B+, it could be a good one but
You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the
amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.

Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build
them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat
2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity
RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp
#2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

I'd probably offer them each the same amp
able to make 100 watts into either 6 ohms or 3.8 ohms.
due to at least two load matches possible.

If a dude has speakers that have a LF aor HF band which dips to 2 ohms, my
amps won't mind it, because some loading where RL = 1.2 the rated
3.8 or 6 ohm outlets is OK, just as it won't hurt
anything to use twice the load of the outlet.

But If the speaker averages 2 ohms across the whole band,
I would give the customer an OPT which
can be matched to suit, so when his mates come around
and connect their gee whiz SS up, they smoke, but
not my tubes.

I'd still wanna know the Z curve before I build any amps.

The 300 watters are the exception.

They have """oddles""" of headroom.

And can drive any load period.

The OPT can be configured for 5 ohms, or 2.2 ohms.

If one only wants 100 watts, and one has 6 ohm speakers,
the 2.2 ohm connection is used and 8 of the 12 output tubes can be
removed.
The 4 left will cope fine.

Patrick Turner.

That's what I said a while back, Patrick. If you can not physically remove
the tubes in your monster amp, but bank them with a switch, perhaps a 4-8-12
switch. I don't think anyone has done this and I'll take only 20% royalties
if you patent it! I have to talk to you about your Godzilla Amp or maybe the
Son of Godzilla. My budget will probably make me end up with Godzilla's
mother-in-law.
west

Fabio Berutti wrote:

IMHO a good amp is a good one when it is good for everything: classic and
rock, Audiostatics and Infinity. This means it has to provide a generous
power (but I suppose 100-150W are more than enough), a good bandwidth, a
good damping factor etc, and it has to do it reliably, continuously, without
costing an arm and a leg and without forcing You to strengthen the floor to
carry its weight. Bi- or tri-amping? Maybe. If You already have many amps
You can make an active crossover, then connect one amp to each single
loudspeaker and save the money needed for a new big one, but it's abit
over-complicated for my taste. I think that You should get a plain, simple,
good 100W tubed amp (say a Citation 4 if You can borrow one, just not to
disturb Patrick everytime) and hear how it works in Your house, with Your
speakers and Your taste. Measurements do not necessarily tell all the
truth. Then go on accordingly, there's no substitute for experience.

Well said.

A guy I know tried biamping with SE for mids and treble
and large PP for the bass, using the passive elements of the Xover
and filtering bass from the lower powered mid/treb amps, but then he went
back to one good amp for the lot.

Another has also biamped, with horns for mids and treble and SET
and MC60 McI for bass 15".
He aso has a lot of bothers because the horns were found to be very uneven in
response,
and he has no treble balance control.
He is using an SS active crossover, and all is rather complex, and its hard for
him to easily change speakers
or amps unless he measures it all carefully, and he has no gear or training for
that....

I tried bi-amping, then abandoned the idea.

And one day I might try to get to Italy.

Patrick Turner.



Ciao

Fabio

BTW: have You been in Perugia during Umbria Jazz?

"west" ha scritto nel messaggio
. ..

"west" wrote in message
. ..
"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible way
to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number of
low
internal resistance power valves, like series-pass triodes (6S33,
6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8 ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn
ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause
they
all
need some 10K as anode load, and this makes the design of the OPT a
real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to
Your
peculiar needs, would be to use TV output tubes, say EL509 or 36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for
safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very
low
impedance which works at about 250V B+, it could be a good one but
You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the
amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.
Patrick,
Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build
them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat
2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity
RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp #2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

Why are we focusing only on 2 ohms? This is the impedance for a specific
frequency. One must look at the entire impedance curve for a particular
speaker as Professor P. stated. I'm starting to think that this curve is
more valuable than knowing a speakers SPL(db)/2.83V/M. Fabio, do you think
you have the skills to dissect a single input speaker system and
surgically
separate the woofers, mids and tweeter to make a bi or even tri amp system
when there was none to begin with? Seems like a lot of math involved but
no
more than 9th grade algebra. It's been a while since I figured polar to
rectangular notation are its reverse. With impedance you have to have an
angle with it. Zero degrees of course is pure resistive,
negative-capacitive, and positive- inductive. I think it can be done. Then
you have an amp(s) custom tailored to your needs. However as they say in
Perugia ... "Tra il dire e il fare, c'e`di mezzo il mare...talk is cheap."
Ciao,
west

"Patrick Turner" ha scritto nel messaggio
...


Fabio Berutti wrote:

IMHO a good amp is a good one when it is good for everything: classic and
rock, Audiostatics and Infinity. This means it has to provide a generous
power (but I suppose 100-150W are more than enough), a good bandwidth, a
good damping factor etc, and it has to do it reliably, continuously,
without
costing an arm and a leg and without forcing You to strengthen the floor
to
carry its weight. Bi- or tri-amping? Maybe. If You already have many
amps
You can make an active crossover, then connect one amp to each single
loudspeaker and save the money needed for a new big one, but it's abit
over-complicated for my taste. I think that You should get a plain,
simple,
good 100W tubed amp (say a Citation 4 if You can borrow one, just not to
disturb Patrick everytime) and hear how it works in Your house, with Your
speakers and Your taste. Measurements do not necessarily tell all the
truth. Then go on accordingly, there's no substitute for experience.

Well said.

A guy I know tried biamping with SE for mids and treble
and large PP for the bass, using the passive elements of the Xover
and filtering bass from the lower powered mid/treb amps, but then he went
back to one good amp for the lot.

Another has also biamped, with horns for mids and treble and SET
and MC60 McI for bass 15".
He aso has a lot of bothers because the horns were found to be very uneven
in
response,
and he has no treble balance control.
He is using an SS active crossover, and all is rather complex, and its
hard for
him to easily change speakers
or amps unless he measures it all carefully, and he has no gear or
training for
that....

I tried bi-amping, then abandoned the idea.

And one day I might try to get to Italy.


If You don't mind listening to a humble couple of 6L6GC (the 811 is not
running, so far) I'll treat You a beer.

Ciao

Fabio


Patrick Turner.



Ciao

Fabio

BTW: have You been in Perugia during Umbria Jazz?

"west" ha scritto nel messaggio
. ..

"west" wrote in message
. ..
"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible
way
to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4
ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number
of
low
internal resistance power valves, like series-pass triodes (6S33,
6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of
NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8
ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes
LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further
reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn
ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice,
'cause
they
all
need some 10K as anode load, and this makes the design of the OPT
a
real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due
to
Your
peculiar needs, would be to use TV output tubes, say EL509 or
36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should
be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for
safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a
very
low
impedance which works at about 250V B+, it could be a good one but
You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes
the
amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the
output
impedance and will take care of the minor non-linearities of TV
scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.
Patrick,
Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build
them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat
2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity
RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp
#2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

Why are we focusing only on 2 ohms? This is the impedance for a
specific
frequency. One must look at the entire impedance curve for a particular
speaker as Professor P. stated. I'm starting to think that this curve
is
more valuable than knowing a speakers SPL(db)/2.83V/M. Fabio, do you
think
you have the skills to dissect a single input speaker system and
surgically
separate the woofers, mids and tweeter to make a bi or even tri amp
system
when there was none to begin with? Seems like a lot of math involved
but
no
more than 9th grade algebra. It's been a while since I figured polar to
rectangular notation are its reverse. With impedance you have to have
an
angle with it. Zero degrees of course is pure resistive,
negative-capacitive, and positive- inductive. I think it can be done.
Then
you have an amp(s) custom tailored to your needs. However as they say
in
Perugia ... "Tra il dire e il fare, c'e`di mezzo il mare...talk is
cheap."
Ciao,
west

Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's say
for
electrostatics, I would need a large current amp that can put out high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2
ohm
current capabilities is the more important criteria. That's all he
would
say
besides there are absolutely no parts available for these 20 year old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of
prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best, cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was not
all
that inefficient. It was about average. Tubes need not apply. Transistors
are the best option. Whilst not the IIIb, here is the impedance curve of
the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg

--
Trevor Wilson
www.rageaudio.com.au

I cannot see any evidence of you having given West any good advice.

**Then you're not paying attention to logic and common-sense. I have
extensive experisnce will many of the old Infinity speakers. I know the
RSIIIb very well indeed. They are particularly poorly suited to use with
tube amps. In fact, they are particularly unsuited to use with *any*
amplifier not capable of delivering prodigious current and providing good
damping, at LF. Tube amps, whilst having some desirable charteristics, are
not real good at these things.

But you are almost completely ignorant about tube amp design methods to obtain
good damping, plenty of power, huge current ability, and all the things you
think are impossible with tubes.
When was the last time you built a tube amp?
1960?
Of course I know you gave up farnarcling around with tube design or any SS
designing in about 1976
when Peter Stein of ME technologies brought out his range of Oz SS amps
using all bjt design.

This whole issue of being able to provide sufficient current is
not a big deal. a dozen KT88/6550 when matched properly to a load with an OPT
will provide all the sufficient current that could ever be wanted.

Before last christmas I designed and built an SS amp for a client just to bring
myself
up to date about what is possible, and what is not, with SS amps.
with 10 output bjts each with 15 amp peak current ability, so 75 peak amps is
possible and it was capable
of about 400 watts *continuous* into 2 ohms.
400w into 2 ohms is only 28 vrms, or 14 amps rms.

My 300 watt tube amps could easily be configured for a load to easily provide 14
Arms,
simply because they are capable of 400 watts into 2 ohms.
I am using a pile of local CFB in the output stage,
( and I doubt you have the slightest idea what that really means, )
so the power before getting anywhere near clipping is clean.

Just don't try to say tube amps can't do it Trevor,
They can, and they do.



And who would ever need this high current?

I can't think of a situtation, because regardless of load value,
even with speakers of 86 dB/W/M, 400 watts x two channels gives an SPL
of 121 dB, which would have the police around to make an arrest.


The impedance curve is most misleading.

**No, it is not. It is exactly what you see.

I don't believe you, and neither should anyone else.

There is no identifying info on the document.

No person has signed it after checking several samples.

Only a present day measurement would satisfy me.




There is no identification of exactly what speaker it is for,
and is it impedance?, or phase shift?

**I thought I explained that. It is impedance. LMS displays both impedance
and pahse shift, or either one. I deleted the pahse shift information in the
interests of clarity.

And you confused us all with a partially deleted document.

Please try to be slightly more convincing than you are.






If it is impedance, lowest Z = 0.8 ohms at 25Hz,
which seems utterly impossible,
since the DCR would surely make Z minimum at least = DCR, no?

**Undoubtedly.

What is the measured DCR for the Infinitys?




You went on to say that a 20 watt Rotel or NAD amp would be OK
and that to use tubes would be unwise.

This is utter BS.

**No, Patrick, it is an inescapable fact. I specified those two types of
amps, since in the case of Rotel, no current limiters are used. Older style
NADs also lack current limiters.

Well I thought after taking my intellectual cudgel to your arguments and claims
that maybe you got it wrong, and meant a 200 watt Rotel or Nad.
But you are saying a 20 watt amp would do.

a 20 watt SS amp is usually one rated to make 20 watts into 8 ohms,
so that's 12.6 vrms output, so the rails must be about +/- 20v,
and so with no current limiting, maybe a swing of 10vrms into 2 ohms is
possible,
so I = 5Arms only.

That's not a lot of current ability, but hey, maybe its ok.
A tube can *easily* do 5 amps.






Even if the curves were the right ones for Infinity RSIIIA, and
i don't see any proof that they are,

**I stated, very clearly, that the curve was for the RSIIIb's lineal
descendent, the Kappa 9. Did you not read that part?

Maybe not even close to the RSIIIb.

Who knows?



then any 50 watt tube
amp with a 4 ohm outlet match would
be OK, since Z averages 4 ohms from 80Hz to 6 kHz.

**I never suggested that a decent tube amp would not work. It is just that
the sonic compromise would be very severe.

You cannot have it both ways.

You say a tube amp will work, and then say it won't, sonically.
Make up your mind.

Plenty here would suggest that SS amps, especially a ****ant 20 watt Rotel or
Nad
would have more trouble with these allegedly troublesome speakers
than a tube amp would.




Z averages 6 ohms between 120Hz and 1 kHz, the band which has the most
signal.
Tubes are more than capable of coping with load mismatches
where the amount of energy is low, ie, below 80 Hz and above 6 kHz.
Many tube amps have their maximum power output
for a load below the nominated load for a given load match.
So that where one uses the 4 ohm outlet, often maximum power at 3 ohms
at clipping is 30% above the maximum 4 ohm power at clipping.
Below 3 ohms the max power may well roll off, but it
certainly does not mean that there would not be ample
power, voltage and current to cope with the low load
at extremes of bandwidth.

A pair of KT88 is ideally set up to
make 35w to 8 ohms, lots of class A, 55 wattsAB to 4 ohms, including about
15
watts
of class A, and perhaps 65 watts into less than 3 ohms, with not much
class A.

The 20 watt level is between about 1 ohm and 12 ohms.

And this would be when using the 8 ohm outlet.
If there was a 4 ohm outlet, you'd get all the above powers at loads 1/2
what I
said,
so 65 watts absolute max into 1.5 ohms, which is 6.6 amps rms.

Now if you were to take advantage of a high current capable amp using SS,
and you were abole to use 75 peak amp capability, we are talking
53 A rms, which is 79.5 vrms into 1.5 ohms which is 4,213 watts.
You'd bleed from the ears.
Huh? But I thought we could do 75 peak amps?
Unless the rails of the amp were up around +/- 125v, forget
ever seeing anything like 75 amps, and what's more,
you won't EVER EVER need such current ablity.


Unless anyone can get the exact and actual impedance
curve the factory prepared all those years ago and for the exact model
number,
there is only one proper way to find out how hard it is to drive Infinity
RSIIIA

**Just a reminder, Patrick, we're discussing the RSIIIb, not the RSIIIa. Not
that there is much difference. There is, however, a significant difference
between the RSIII and the RSIIIa.

To prevent every one including you and me being bogged down with much discussion
about nothing,
I have strongly recommended West carry out correctly done
impedance measurements of his speakers. Not just one, but both, in case one is
different to the other,
which has been known to happen.

Until we know a real curve, then we are all **** and wind.




is to measure the speaker impedance properly,
and do the measurments on peak voltage and peak
current for your loudest
listening using a borrowed powerful SS amp, of least 200W.

The impedance curve which Trevor has guided us to looks
quite wrong.

**Then YOU present a curve YOU have mneasured for the Kappa 9. I'll wait.


Happy waiting. Better get a sofa, I'll be awhile.



But if it were correct, there is a worry that if a tube amp has Ro = 0.5
ohms,
then where Z 3 ohms the response would fall and a dB of response
would go missing above 2 kHz, because Z averages 2 ohms above 2 kHz.
This would be the case with some tube amps using their "8 ohm" outlet.
But the 4 ohm outlet will have Ro = half the 8 ohm outlet Ro,
or Ro = 0.25 ohms, so negligible losses will occur at HF.

People should always remember Trevor Wilson strongly favours
solid state amps, which he trades, especially those made by ME Technology,
a
former Oz company
which went into liquidation over a year ago.

Now Trevor has said a 20 watt Rotel or NAD would do.
We can assume that means 20 watts into 4 ohms, which is only 2.2 amps.

I'd actually be surprised if anyone needed more than 50 watts, just to be
sure.

But before anyone here can talk any sense, lets see the measurements
of the RSIIIA speakers.

**Fair enough. I'll see what I can manage. They turn up from time to time.

I don't mind the discussions about tubes and finding the right amp to drive a
given pair,
but it all should make sense, and not be just old farts braying meaninglessly
into spaces.

If you want current ability, and voltage ability, and ability to drive 2 ohms
to ear shattering levels, buy a 300 watt amp happy to drive 2 ohms all day.

If the speaker Z rises to 8 ohms, there will still be enough voltage to manage
plenty of power.

I am trying to prevent a man buying a 300 watt amp needlessly.

I could easily wind a wide bandwidth transformers usable with any
amp to have an 8 ohm input, and taps off for
2,3,4,5,6,7,8,9,10 ohm speakers.
So this would allow considerable impedance tailoring
without the losses incurred when using resistors in series with midranges and
treble
units to obtain the right attenuations, and the source impedance drops with
a secondary of lower Z than the primary input.

Patrick Turner.





--
Trevor Wilson
www.rageaudio.com.au

west wrote:

"Patrick Turner" wrote in message
...


west wrote:

"Patrick Turner" wrote in message
...


Fabio Berutti wrote:

OK, now I got the point. It is true that SS is the more sensible
way to
go
to get many amps on 2 Ohms.

Not necessarily.

To make sure a tube amp will cope with a load
of ***an average of 2 ohms***, then we'd have a 2 ohm outlet, or an
available
OPT secondary configuration to allow 2 ohms to be used to
still give the right anode loads.
Many tube amps can be set to ideally match an average load of 4 ohms,
so a dip of Z to 2 ohms away from the band where Z is mainly above
4 ohms will be OK.

But this is a tubed group and the sand state of
matter is considered only to make beaches, not amps. In order to
provide
the performances You need, a tubed amp shall use a "large" number of
low
internal resistance power valves, like series-pass triodes (6S33,
6S41,
6AS7...) or TV scan pentodes or trioded 6550/KT88s.

Low Ra tubes are not required.

Amps configured in beam tetrode of pentode are quite able to work
with 2 ohms.

You just need the right OPT impedance ratio and the usual 20 dB of
NFB.


I'm not in favour of
OTLs, particularly if You need 2 Ohms capability.

Nor am I unless ppl use 4 times as many tubes as one would with 8
ohms.

A triple push-pull of
6550s can work into a primary reflected load of 1K, which makes LOTS
easier
designing a suitable OPT to 2 Ohms. This load can be further
reduced
using
more (or different) tubes.

Yes, but a pair of KT88 could be used in class AB with 10k a-a,
but with a 2 ohm secondary load with a 5,000:1 Z ratio = 70:1 turn
ratio.
In triode they'd give 30 watts of mostly class A and be capable
of a maximum of 17 amps at the output, if Ia max was 500 mA.
Without any Global NFB the amp would measure Ro = about 0.45 ohms,
so the usual 12 dB of global NFB could reduce that to 0.15 ohms,
quite adequate, don't you think?

KT90 would produce more power, more current.


All "giant" tubes like GM-70 are definitely not Your choice, 'cause
they
all
need some 10K as anode load, and this makes the design of the OPT a
real
mess without any advantage.

Definately no good for OTL, unless we want to drive ESL direct.....


My only modification with respect to Patrick's designs, just due to
Your
peculiar needs, would be to use TV output tubes, say EL509 or
36LW6,
because of their lower impedance and larger current capability.

EL509 in pentode has a very high ra compared to the loads it should be
used
with,
and is not really any better than KT88, EL34 etc,
and still one has to use NFB like EAR509 does, and that uses 44 dB.
EL509 has Pda = 35 watts, and very low class A ability.
KT90 with Pad = 55 watts and KT88/6550 with Pda = 42 watts
are all better audio tubes than the EL/PL509 etc.


Unfortunately they all have top caps, which are not the best for
safety.
The EL500 (Russian 6P36) is a small (15W) and cheap tube with a very
low
impedance which works at about 250V B+, it could be a good one but
You'd
need to pack some 16 of them per channel...
By the way, working at lower voltages and higher currents makes the
amp
closer to 100% Class A.
A sensible amount of feedback will help reducing further the output
impedance and will take care of the minor non-linearities of TV scan
pentodes.

I like 6CM5/EL36 in triode.
They can take Pda = 18 watts at idle and two make
28 watts max in class AB1 triode using an OPT of course.
Forget OTL, unless you like class B amps and lots of NFB.

So an octet of triode 6CM5 in triode can easily make 100 watts,
Ea = 375v, Ia = 40 mA each, RLa-a = about 1.7k, Eg1 fixed bias = -60
approx.

Patrick Turner.

Just trying to put some of this in a practical nutshell, Patrick.:
You have 2 customers that need around 100 watts and want you to build
them
an amp. Cust #1 says I need to swing a lot of voltage for my Acoustat
2+2s.
Cust #2 (me) says I need a lot of current at 2 ohms for my Infinity
RSIIIbs.
Both amps are about 100watts, generous amounts of class A.
What will you most likely have for your output design in amp #1 & amp
#2,
tubes, topology OPT, etc? (Hope I'm not putting you on the spot, Pat
although I know you're getting a workout). :-(
west

I'd probably offer them each the same amp
able to make 100 watts into either 6 ohms or 3.8 ohms.
due to at least two load matches possible.

If a dude has speakers that have a LF aor HF band which dips to 2 ohms, my
amps won't mind it, because some loading where RL = 1.2 the rated
3.8 or 6 ohm outlets is OK, just as it won't hurt
anything to use twice the load of the outlet.

But If the speaker averages 2 ohms across the whole band,
I would give the customer an OPT which
can be matched to suit, so when his mates come around
and connect their gee whiz SS up, they smoke, but
not my tubes.

I'd still wanna know the Z curve before I build any amps.

The 300 watters are the exception.

They have """oddles""" of headroom.

And can drive any load period.

The OPT can be configured for 5 ohms, or 2.2 ohms.

If one only wants 100 watts, and one has 6 ohm speakers,
the 2.2 ohm connection is used and 8 of the 12 output tubes can be
removed.
The 4 left will cope fine.

Patrick Turner.

That's what I said a while back, Patrick. If you can not physically remove
the tubes in your monster amp, but bank them with a switch, perhaps a 4-8-12
switch. I don't think anyone has done this and I'll take only 20% royalties
if you patent it! I have to talk to you about your Godzilla Amp or maybe the
Son of Godzilla. My budget will probably make me end up with Godzilla's
mother-in-law.
west

ARC and Jadis beat me to it.

Lennard in Oz also makes monster amps.

And there is a design from Wireless World from 1957 for 400 watts from
10 x KT88, all sitting on a panel of 3/4" plywood.
Not quite as pretty as ARC reference 600, but it did the business,
and had a simple Williamson circuit, but with an additional 12AU7 in the drive
amps used to make two CF buffers to drive the 5 paralleled grid bias R.

Its all been done before, simply and cheaply.

Its much cheaper to make a 400 watt amp than make ten 40 watt amps
when priced in dollars per watt.


One can have a switch to simply turn off the heaters for
the 3 quads of tubes, so you can have 4, 8 or 12 outputs.
With no heaters, the tubes sit there and are inactive, with only tiny
static electrode capacitances affecting the circuit.

I don't want to really do this. Why buy 300 watt capable amps when 50 may do?

And of course people could alter the B+ of my amps to allow
6V6 to be used; the load match can be set for 1.1k : 2.2 ohms,
which translates to 2.2k to 4.4ohms, all doable with a dozen el cheapo 6V6,
and still able to make 120 watts.

But the load impedance changing isn't easy.

There are 12 secondary windings able to be connected to
a load match to 5 ohms or 2.2 hms to 1.1k primary load.

The secs consist of 6 x 48 turns and 6 x 24 turns, so
you can have 6 x 72 turns all parallel to get 5 ohms,
or 9 x 48 turn windings in parallel for 2.2 ohms and the same low winding losses

and same bandwidth.

A soldering iron is required to make the change. I have not figured a way to
have
a plate which can be rotated and screwed tightly down in two positions to allow
easy change of load matches.

But who would want to?

The load match to 5 ohms was my choice because modern speakers often have a 6
ohm
Z with dips to 3 ohms.
My amps all produce more power as the RL is reduced to near 1/2 the rated
load match.

With 350 watts into 3 ohms available when using the 5 ohm load match,
( but never likely to ever be used, )
the load could dip to 2 ohms in places and not be much trouble,
since maybe 100 watts is still possible into 2 ohms,
and I bet such speakers as you have don't need that power at any
impedance level.

Its all very well for me to say the max is 350 watts into 3 ohms.
But in fact its means there is at least 100 watts from 2 ohms to 16,
so wide Z changes won't matter much if the main band averages 4 ohms+ as i
suspect they do
with your Infinities.

But we need to measure them to be sure.

Patrick Turner.

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail
mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's
say
for
electrostatics, I would need a large current amp that can put out
high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2
ohm
current capabilities is the more important criteria. That's all he
would
say
besides there are absolutely no parts available for these 20 year
old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of
prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best,
cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was not
all
that inefficient. It was about average. Tubes need not apply.
Transistors
are the best option. Whilst not the IIIb, here is the impedance curve
of
the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg

--
Trevor Wilson
www.rageaudio.com.au

I cannot see any evidence of you having given West any good advice.

**Then you're not paying attention to logic and common-sense. I have
extensive experisnce will many of the old Infinity speakers. I know the
RSIIIb very well indeed. They are particularly poorly suited to use with
tube amps. In fact, they are particularly unsuited to use with *any*
amplifier not capable of delivering prodigious current and providing good
damping, at LF. Tube amps, whilst having some desirable charteristics,
are
not real good at these things.

But you are almost completely ignorant about tube amp design methods to
obtain
good damping, plenty of power, huge current ability, and all the things
you
think are impossible with tubes.

**Those things are not impossible. They're just bloody hard and very
expensive to achieve. Tubes are good at some things. High current and high
damping factors are just not two of them.

When was the last time you built a tube amp?

**Ca. 1968. Though I was still building tube preamps up until around 1975.

1960?

**Nah. I was 7 years old in 1960.

Of course I know you gave up farnarcling around with tube design or any SS
designing in about 1976

**Around 1978, actually.

when Peter Stein of ME technologies brought out his range of Oz SS amps
using all bjt design.

**Indeed. Revolutionary, at the time. I purchased one of his first amps
awhile ago. It surprises people, even now, with it's exceptional sound
quality. Don't forget: Part of Peter's rationale for building amps was to
examine why people still enjoyed the sound of tube amps. He established the
reasons why tube amps performed so well (in some areas) and built BJT
designs to emulate the good qualities of tube amps and retained the good
qualities of SS amps.


This whole issue of being able to provide sufficient current is
not a big deal. a dozen KT88/6550 when matched properly to a load with an
OPT
will provide all the sufficient current that could ever be wanted.

**And the cost would be? And the output impedance would be?


Before last christmas I designed and built an SS amp for a client just to
bring
myself
up to date about what is possible, and what is not, with SS amps.
with 10 output bjts each with 15 amp peak current ability, so 75 peak amps
is
possible and it was capable
of about 400 watts *continuous* into 2 ohms.
400w into 2 ohms is only 28 vrms, or 14 amps rms.

My 300 watt tube amps could easily be configured for a load to easily
provide 14
Arms,
simply because they are capable of 400 watts into 2 ohms.
I am using a pile of local CFB in the output stage,
( and I doubt you have the slightest idea what that really means, )
so the power before getting anywhere near clipping is clean.

Just don't try to say tube amps can't do it Trevor,
They can, and they do.

**And the cost difference is?




And who would ever need this high current?

**Asked and answered.


I can't think of a situtation, because regardless of load value,
even with speakers of 86 dB/W/M, 400 watts x two channels gives an SPL
of 121 dB, which would have the police around to make an arrest.


The impedance curve is most misleading.

**No, it is not. It is exactly what you see.

I don't believe you, and neither should anyone else.

There is no identifying info on the document.

No person has signed it after checking several samples.

Only a present day measurement would satisfy me.

**I don't give a **** what would satisfy you, Patrick. If you wish to
measure the Kappa 9 and prove me wrong, then feel free to do so. Until then,
you would appear to be full of **** and wind.





There is no identification of exactly what speaker it is for,
and is it impedance?, or phase shift?

**I thought I explained that. It is impedance. LMS displays both
impedance
and pahse shift, or either one. I deleted the pahse shift information in
the
interests of clarity.

And you confused us all with a partially deleted document.

**No Patrick. Only YOU were confused. Everyone else managed to work it out.


Please try to be slightly more convincing than you are.






If it is impedance, lowest Z = 0.8 ohms at 25Hz,
which seems utterly impossible,
since the DCR would surely make Z minimum at least = DCR, no?

**Undoubtedly.

What is the measured DCR for the Infinitys?

**Dunno. Don't care. I know of no music which extends down to DC. I measure
speakers over the range of human hearing and sometimes beyond.





You went on to say that a 20 watt Rotel or NAD amp would be OK
and that to use tubes would be unwise.

This is utter BS.

**No, Patrick, it is an inescapable fact. I specified those two types of
amps, since in the case of Rotel, no current limiters are used. Older
style
NADs also lack current limiters.

Well I thought after taking my intellectual cudgel to your arguments and
claims
that maybe you got it wrong, and meant a 200 watt Rotel or Nad.
But you are saying a 20 watt amp would do.

**Indeed, it will do just fine.


a 20 watt SS amp is usually one rated to make 20 watts into 8 ohms,
so that's 12.6 vrms output, so the rails must be about +/- 20v,
and so with no current limiting, maybe a swing of 10vrms into 2 ohms is
possible,
so I = 5Arms only.

That's not a lot of current ability, but hey, maybe its ok.
A tube can *easily* do 5 amps.

**At what cost?







Even if the curves were the right ones for Infinity RSIIIA, and
i don't see any proof that they are,

**I stated, very clearly, that the curve was for the RSIIIb's lineal
descendent, the Kappa 9. Did you not read that part?

Maybe not even close to the RSIIIb.

Who knows?



then any 50 watt tube
amp with a 4 ohm outlet match would
be OK, since Z averages 4 ohms from 80Hz to 6 kHz.

**I never suggested that a decent tube amp would not work. It is just
that
the sonic compromise would be very severe.

You cannot have it both ways.

You say a tube amp will work, and then say it won't, sonically.
Make up your mind.

**Stop being stupid (if that is possible).


Plenty here would suggest that SS amps, especially a ****ant 20 watt Rotel
or
Nad
would have more trouble with these allegedly troublesome speakers
than a tube amp would.

**They can suggest all they wish. AFTER they've actually tried it. I have. I
actually know what I am tlaking about.





Z averages 6 ohms between 120Hz and 1 kHz, the band which has the most
signal.
Tubes are more than capable of coping with load mismatches
where the amount of energy is low, ie, below 80 Hz and above 6 kHz.
Many tube amps have their maximum power output
for a load below the nominated load for a given load match.
So that where one uses the 4 ohm outlet, often maximum power at 3 ohms
at clipping is 30% above the maximum 4 ohm power at clipping.
Below 3 ohms the max power may well roll off, but it
certainly does not mean that there would not be ample
power, voltage and current to cope with the low load
at extremes of bandwidth.

A pair of KT88 is ideally set up to
make 35w to 8 ohms, lots of class A, 55 wattsAB to 4 ohms, including
about
15
watts
of class A, and perhaps 65 watts into less than 3 ohms, with not much
class A.

The 20 watt level is between about 1 ohm and 12 ohms.

And this would be when using the 8 ohm outlet.
If there was a 4 ohm outlet, you'd get all the above powers at loads
1/2
what I
said,
so 65 watts absolute max into 1.5 ohms, which is 6.6 amps rms.

Now if you were to take advantage of a high current capable amp using
SS,
and you were abole to use 75 peak amp capability, we are talking
53 A rms, which is 79.5 vrms into 1.5 ohms which is 4,213 watts.
You'd bleed from the ears.
Huh? But I thought we could do 75 peak amps?
Unless the rails of the amp were up around +/- 125v, forget
ever seeing anything like 75 amps, and what's more,
you won't EVER EVER need such current ablity.


Unless anyone can get the exact and actual impedance
curve the factory prepared all those years ago and for the exact model
number,
there is only one proper way to find out how hard it is to drive
Infinity
RSIIIA

**Just a reminder, Patrick, we're discussing the RSIIIb, not the RSIIIa.
Not
that there is much difference. There is, however, a significant
difference
between the RSIII and the RSIIIa.

To prevent every one including you and me being bogged down with much
discussion
about nothing,
I have strongly recommended West carry out correctly done
impedance measurements of his speakers. Not just one, but both, in case
one is
different to the other,
which has been known to happen.

Until we know a real curve, then we are all **** and wind.




is to measure the speaker impedance properly,
and do the measurments on peak voltage and peak
current for your loudest
listening using a borrowed powerful SS amp, of least 200W.

The impedance curve which Trevor has guided us to looks
quite wrong.

**Then YOU present a curve YOU have mneasured for the Kappa 9. I'll wait.


Happy waiting. Better get a sofa, I'll be awhile.



But if it were correct, there is a worry that if a tube amp has Ro =
0.5
ohms,
then where Z 3 ohms the response would fall and a dB of response
would go missing above 2 kHz, because Z averages 2 ohms above 2 kHz.
This would be the case with some tube amps using their "8 ohm" outlet.
But the 4 ohm outlet will have Ro = half the 8 ohm outlet Ro,
or Ro = 0.25 ohms, so negligible losses will occur at HF.

People should always remember Trevor Wilson strongly favours
solid state amps, which he trades, especially those made by ME
Technology,
a
former Oz company
which went into liquidation over a year ago.

Now Trevor has said a 20 watt Rotel or NAD would do.
We can assume that means 20 watts into 4 ohms, which is only 2.2 amps.

I'd actually be surprised if anyone needed more than 50 watts, just to
be
sure.

But before anyone here can talk any sense, lets see the measurements
of the RSIIIA speakers.

**Fair enough. I'll see what I can manage. They turn up from time to
time.

I don't mind the discussions about tubes and finding the right amp to
drive a
given pair,
but it all should make sense, and not be just old farts braying
meaninglessly
into spaces.

If you want current ability, and voltage ability, and ability to drive 2
ohms
to ear shattering levels, buy a 300 watt amp happy to drive 2 ohms all
day.

If the speaker Z rises to 8 ohms, there will still be enough voltage to
manage
plenty of power.

I am trying to prevent a man buying a 300 watt amp needlessly.

**Indeed. A 20 Watt Rotel will work a treat with the RSIIIb. No current
limiting, you see. They're relatively inexpensive and fine sounding
amplifiers.


I could easily wind a wide bandwidth transformers usable with any
amp to have an 8 ohm input, and taps off for
2,3,4,5,6,7,8,9,10 ohm speakers.

**I'm sure you can. Tell me: What do you do when the impedance falls to
(say) 1 Ohm and rises to (say) 9 Ohms. Are you gonna swap those taps around
to suit?


--
Trevor Wilson
www.rageaudio.com.au

Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail
mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp, let's
say
for
electrostatics, I would need a large current amp that can put out
high
current into a 2 ohm load. The wattage is 75 to 200 watts, but the 2
ohm
current capabilities is the more important criteria. That's all he
would
say
besides there are absolutely no parts available for these 20 year
old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of
prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best,
cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was not
all
that inefficient. It was about average. Tubes need not apply.
Transistors
are the best option. Whilst not the IIIb, here is the impedance curve
of
the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg

--
Trevor Wilson
www.rageaudio.com.au

I cannot see any evidence of you having given West any good advice.

**Then you're not paying attention to logic and common-sense. I have
extensive experisnce will many of the old Infinity speakers. I know the
RSIIIb very well indeed. They are particularly poorly suited to use with
tube amps. In fact, they are particularly unsuited to use with *any*
amplifier not capable of delivering prodigious current and providing good
damping, at LF. Tube amps, whilst having some desirable charteristics,
are
not real good at these things.

But you are almost completely ignorant about tube amp design methods to
obtain
good damping, plenty of power, huge current ability, and all the things
you
think are impossible with tubes.

**Those things are not impossible. They're just bloody hard and very
expensive to achieve. Tubes are good at some things. High current and high
damping factors are just not two of them.

Unfortunately, arguing with you is about the least rewarding activities
I could think of persuing, since you just don't
understand tube use.

People here don't mind how hard it is to build somethiung with tubes.

And we don't care about the expense.

Your insinuation is that we should not try,
or that we waste our time, or that our persuit is in some way illegitimate,
like building a wooden yacht, when it is possible to build a fibreglass stink
boat
for far less money.

I'm just letting you know our activities won't be changed by a word you say,
and that high current and high damping factors are quite high enough
using tubes for the output devices.





When was the last time you built a tube amp?

**Ca. 1968. Though I was still building tube preamps up until around 1975.

1960?

**Nah. I was 7 years old in 1960.

Of course I know you gave up farnarcling around with tube design or any SS
designing in about 1976

**Around 1978, actually.

when Peter Stein of ME technologies brought out his range of Oz SS amps
using all bjt design.

**Indeed. Revolutionary, at the time. I purchased one of his first amps
awhile ago. It surprises people, even now, with it's exceptional sound
quality. Don't forget: Part of Peter's rationale for building amps was to
examine why people still enjoyed the sound of tube amps. He established the
reasons why tube amps performed so well (in some areas) and built BJT
designs to emulate the good qualities of tube amps and retained the good
qualities of SS amps.

After 30 years have passed, not a huge number in the two main audio clubs in Oz
would opt to by ME gear for their sound lounges.





This whole issue of being able to provide sufficient current is
not a big deal. a dozen KT88/6550 when matched properly to a load with an
OPT
will provide all the sufficient current that could ever be wanted.

**And the cost would be? And the output impedance would be?

Cost has got nothing to do with performance.

How much does a Mark Levinson or a Krell cost?

You are trying to argue SS amps are the best on a tubes group.

Its like trying to sell pork in a synagogue.




Before last christmas I designed and built an SS amp for a client just to
bring
myself
up to date about what is possible, and what is not, with SS amps.
with 10 output bjts each with 15 amp peak current ability, so 75 peak amps
is
possible and it was capable
of about 400 watts *continuous* into 2 ohms.
400w into 2 ohms is only 28 vrms, or 14 amps rms.

My 300 watt tube amps could easily be configured for a load to easily
provide 14
Arms,
simply because they are capable of 400 watts into 2 ohms.
I am using a pile of local CFB in the output stage,
( and I doubt you have the slightest idea what that really means, )
so the power before getting anywhere near clipping is clean.

Just don't try to say tube amps can't do it Trevor,
They can, and they do.

**And the cost difference is?

Cost is not a consideration.






And who would ever need this high current?

**Asked and answered.


I can't think of a situtation, because regardless of load value,
even with speakers of 86 dB/W/M, 400 watts x two channels gives an SPL
of 121 dB, which would have the police around to make an arrest.


The impedance curve is most misleading.

**No, it is not. It is exactly what you see.

I don't believe you, and neither should anyone else.

There is no identifying info on the document.

No person has signed it after checking several samples.

Only a present day measurement would satisfy me.

**I don't give a **** what would satisfy you, Patrick. If you wish to
measure the Kappa 9 and prove me wrong, then feel free to do so. Until then,
you would appear to be full of **** and wind.





There is no identification of exactly what speaker it is for,
and is it impedance?, or phase shift?

**I thought I explained that. It is impedance. LMS displays both
impedance
and pahse shift, or either one. I deleted the pahse shift information in
the
interests of clarity.

And you confused us all with a partially deleted document.

**No Patrick. Only YOU were confused. Everyone else managed to work it out.

Nope, I challenged the validity of the impedance graph.
Everyone is now doubtful about what you have provided us.



Please try to be slightly more convincing than you are.






If it is impedance, lowest Z = 0.8 ohms at 25Hz,
which seems utterly impossible,
since the DCR would surely make Z minimum at least = DCR, no?

**Undoubtedly.

What is the measured DCR for the Infinitys?

**Dunno. Don't care. I know of no music which extends down to DC. I measure
speakers over the range of human hearing and sometimes beyond.

You should care about the DCR.

The low bass Z of the speakers you referred us to, whatever they were,
went down to 0.8 ohms at some bass F, which is strange since
most speakers have resonant impedance oeaks at LF, and hence a curve
quite unlike the one you tried to trick us with.

The rest of your reply brings nothing new, useful or informative to
the subject.

I await Mr West's measurements of his speakers.

Patrick Turner.

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I
did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker
was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail
mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp,
let's
say
for
electrostatics, I would need a large current amp that can put out
high
current into a 2 ohm load. The wattage is 75 to 200 watts, but
the 2
ohm
current capabilities is the more important criteria. That's all
he
would
say
besides there are absolutely no parts available for these 20 year
old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of
prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best,
cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was
not
all
that inefficient. It was about average. Tubes need not apply.
Transistors
are the best option. Whilst not the IIIb, here is the impedance
curve
of
the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg

--
Trevor Wilson
www.rageaudio.com.au

I cannot see any evidence of you having given West any good advice.

**Then you're not paying attention to logic and common-sense. I have
extensive experisnce will many of the old Infinity speakers. I know
the
RSIIIb very well indeed. They are particularly poorly suited to use
with
tube amps. In fact, they are particularly unsuited to use with *any*
amplifier not capable of delivering prodigious current and providing
good
damping, at LF. Tube amps, whilst having some desirable charteristics,
are
not real good at these things.

But you are almost completely ignorant about tube amp design methods to
obtain
good damping, plenty of power, huge current ability, and all the things
you
think are impossible with tubes.

**Those things are not impossible. They're just bloody hard and very
expensive to achieve. Tubes are good at some things. High current and
high
damping factors are just not two of them.

Unfortunately, arguing with you is about the least rewarding activities
I could think of persuing, since you just don't
understand tube use.

**Wrong Patrick. Instead of arguing ACTUAL facts, you prefer to slur. If you
want to argue actual facts, I am ready.


People here don't mind how hard it is to build somethiung with tubes.

**You speak for EVERYONE? Curious.


And we don't care about the expense.

**You speak for EVERYONE? Curious. I don't beleive you. Anyone attempting to
power a pair of almost 20 year old speakers, is likely to be operating under
some kind of budgetary restraints.


Your insinuation is that we should not try,
or that we waste our time, or that our persuit is in some way
illegitimate,
like building a wooden yacht, when it is possible to build a fibreglass
stink
boat
for far less money.

**And again, you spin my words off into some kind of delusion of your own.
PAY CLOSE ATTENTION:

The best, cheapest way of getting satisfactory performance from the Infinity
RSIIIb speaker, is to use a good, SS amp, which has high current capability.
It is certainly possible to get satisfactory performance from a tube amp
(with that particular speaker system), but it would cost FAR more. Is that
so difficult for you to understand?


I'm just letting you know our activities won't be changed by a word you
say,
and that high current and high damping factors are quite high enough
using tubes for the output devices.

**You speak for EVERYONE? Curious.






When was the last time you built a tube amp?

**Ca. 1968. Though I was still building tube preamps up until around
1975.

1960?

**Nah. I was 7 years old in 1960.

Of course I know you gave up farnarcling around with tube design or any
SS
designing in about 1976

**Around 1978, actually.

when Peter Stein of ME technologies brought out his range of Oz SS amps
using all bjt design.

**Indeed. Revolutionary, at the time. I purchased one of his first amps
awhile ago. It surprises people, even now, with it's exceptional sound
quality. Don't forget: Part of Peter's rationale for building amps was to
examine why people still enjoyed the sound of tube amps. He established
the
reasons why tube amps performed so well (in some areas) and built BJT
designs to emulate the good qualities of tube amps and retained the good
qualities of SS amps.

After 30 years have passed, not a huge number in the two main audio clubs
in Oz
would opt to by ME gear for their sound lounges.

**So?






This whole issue of being able to provide sufficient current is
not a big deal. a dozen KT88/6550 when matched properly to a load with
an
OPT
will provide all the sufficient current that could ever be wanted.

**And the cost would be? And the output impedance would be?

Cost has got nothing to do with performance.

**Cost has EVERYTHING to do with performance. Just look at the comparative
popularity of a Porsche 968 and a Subaru WRX. Both have similar levels of
point to point performance, but the Subaru sells in far larger numbers.


How much does a Mark Levinson or a Krell cost?

**A lot.


You are trying to argue SS amps are the best on a tubes group.

**No. I am assisting the OP in his quest to power a pair of ancient Infinity
RSIIIb speakers. For HIS, SPECIFIC appliaction, a good, high current SS amp,
is the best, least expensive option.


Its like trying to sell pork in a synagogue.

**Perhaps. However, I prefer to beleive that not everyone who posts here is
incapable of logical, rational thought.





Before last christmas I designed and built an SS amp for a client just
to
bring
myself
up to date about what is possible, and what is not, with SS amps.
with 10 output bjts each with 15 amp peak current ability, so 75 peak
amps
is
possible and it was capable
of about 400 watts *continuous* into 2 ohms.
400w into 2 ohms is only 28 vrms, or 14 amps rms.

My 300 watt tube amps could easily be configured for a load to easily
provide 14
Arms,
simply because they are capable of 400 watts into 2 ohms.
I am using a pile of local CFB in the output stage,
( and I doubt you have the slightest idea what that really means, )
so the power before getting anywhere near clipping is clean.

Just don't try to say tube amps can't do it Trevor,
They can, and they do.

**And the cost difference is?

Cost is not a consideration.

**Cost is ALWAYS a consideration. ALWAYS.







And who would ever need this high current?

**Asked and answered.


I can't think of a situtation, because regardless of load value,
even with speakers of 86 dB/W/M, 400 watts x two channels gives an SPL
of 121 dB, which would have the police around to make an arrest.


The impedance curve is most misleading.

**No, it is not. It is exactly what you see.

I don't believe you, and neither should anyone else.

There is no identifying info on the document.

No person has signed it after checking several samples.

Only a present day measurement would satisfy me.

**I don't give a **** what would satisfy you, Patrick. If you wish to
measure the Kappa 9 and prove me wrong, then feel free to do so. Until
then,
you would appear to be full of **** and wind.





There is no identification of exactly what speaker it is for,
and is it impedance?, or phase shift?

**I thought I explained that. It is impedance. LMS displays both
impedance
and pahse shift, or either one. I deleted the pahse shift information
in
the
interests of clarity.

And you confused us all with a partially deleted document.

**No Patrick. Only YOU were confused. Everyone else managed to work it
out.

Nope, I challenged the validity of the impedance graph.

**Good. Then YOU provide your alternate graph. I'll wait.

Everyone is now doubtful about what you have provided us.

**You speak for EVERYONE? Curious.




Please try to be slightly more convincing than you are.






If it is impedance, lowest Z = 0.8 ohms at 25Hz,
which seems utterly impossible,
since the DCR would surely make Z minimum at least = DCR, no?

**Undoubtedly.

What is the measured DCR for the Infinitys?

**Dunno. Don't care. I know of no music which extends down to DC. I
measure
speakers over the range of human hearing and sometimes beyond.

You should care about the DCR.

**Why? How much music goes down to DC?


The low bass Z of the speakers you referred us to, whatever they were,
went down to 0.8 ohms at some bass F, which is strange since
most speakers have resonant impedance oeaks at LF, and hence a curve
quite unlike the one you tried to trick us with.

**I tried to trick no one. I measured a pair of speakers. I published the
plot. Nothing more. Moreover, if you had ANY knowledge of the Kappa 9 (and
the RSIIIb) you would understand that the plot I presented, is, indeed, very
possible.

I suggest that before you launch further attacks, you do some basic research
on the speakers in question. Measuring the impedance would be a good start.


--
Trevor Wilson
www.rageaudio.com.au

Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote lots of stuff I have no intention to reply to.

No need for me to reply to any of the content below.

Patrick Turner.

:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"Patrick Turner" wrote in message
...


Trevor Wilson wrote:

"west" wrote in message
...
First of all I wish to thank all who gave me such wonderful and
prodigious
advice on selecting an amp for my inefficient Infinity RS IIIBs.
Although I haven't finished the calculations as advised yet, I
did
speak
to
an Infinity tech. He said all the specs on the aforesaid speaker
was
published, the efficiency was not because it probably behoove the
company
not to publish them for marketing considerations. He is to snail
mail
me
some tech specs, whatever that means.
The nice gentleman did say unlike a large voltage swing amp,
let's
say
for
electrostatics, I would need a large current amp that can put out
high
current into a 2 ohm load. The wattage is 75 to 200 watts, but
the 2
ohm
current capabilities is the more important criteria. That's all
he
would
say
besides there are absolutely no parts available for these 20 year
old
speakers.
I hope this info may help the pursuit. Thanks again especially to
Patrick,
Fabio, John, et al.

**The man is correct. The RSIIIa and IIIb require amplifiers of
prodigious
current ability. A 20 Watt NAD or Rotel would be one of the best,
cheap
options for these guys. BTW: In terms of dB/2.83V/M the RSIIIb was
not
all
that inefficient. It was about average. Tubes need not apply.
Transistors
are the best option. Whilst not the IIIb, here is the impedance
curve
of
the
lineal descendent of the IIIb:

www.rageaudio.com.au/kapp9.jpg

--
Trevor Wilson
www.rageaudio.com.au

I cannot see any evidence of you having given West any good advice.

**Then you're not paying attention to logic and common-sense. I have
extensive experisnce will many of the old Infinity speakers. I know
the
RSIIIb very well indeed. They are particularly poorly suited to use
with
tube amps. In fact, they are particularly unsuited to use with *any*
amplifier not capable of delivering prodigious current and providing
good
damping, at LF. Tube amps, whilst having some desirable charteristics,
are
not real good at these things.

But you are almost completely ignorant about tube amp design methods to
obtain
good damping, plenty of power, huge current ability, and all the things
you
think are impossible with tubes.

**Those things are not impossible. They're just bloody hard and very
expensive to achieve. Tubes are good at some things. High current and
high
damping factors are just not two of them.

Unfortunately, arguing with you is about the least rewarding activities
I could think of persuing, since you just don't
understand tube use.

**Wrong Patrick. Instead of arguing ACTUAL facts, you prefer to slur. If you
want to argue actual facts, I am ready.


People here don't mind how hard it is to build somethiung with tubes.

**You speak for EVERYONE? Curious.


And we don't care about the expense.

**You speak for EVERYONE? Curious. I don't beleive you. Anyone attempting to
power a pair of almost 20 year old speakers, is likely to be operating under
some kind of budgetary restraints.


Your insinuation is that we should not try,
or that we waste our time, or that our persuit is in some way
illegitimate,
like building a wooden yacht, when it is possible to build a fibreglass
stink
boat
for far less money.

**And again, you spin my words off into some kind of delusion of your own.
PAY CLOSE ATTENTION:

The best, cheapest way of getting satisfactory performance from the Infinity
RSIIIb speaker, is to use a good, SS amp, which has high current capability.
It is certainly possible to get satisfactory performance from a tube amp
(with that particular speaker system), but it would cost FAR more. Is that
so difficult for you to understand?


I'm just letting you know our activities won't be changed by a word you
say,
and that high current and high damping factors are quite high enough
using tubes for the output devices.

**You speak for EVERYONE? Curious.






When was the last time you built a tube amp?

**Ca. 1968. Though I was still building tube preamps up until around
1975.

1960?

**Nah. I was 7 years old in 1960.

Of course I know you gave up farnarcling around with tube design or any
SS
designing in about 1976

**Around 1978, actually.

when Peter Stein of ME technologies brought out his range of Oz SS amps
using all bjt design.

**Indeed. Revolutionary, at the time. I purchased one of his first amps
awhile ago. It surprises people, even now, with it's exceptional sound
quality. Don't forget: Part of Peter's rationale for building amps was to
examine why people still enjoyed the sound of tube amps. He established
the
reasons why tube amps performed so well (in some areas) and built BJT
designs to emulate the good qualities of tube amps and retained the good
qualities of SS amps.

After 30 years have passed, not a huge number in the two main audio clubs
in Oz
would opt to by ME gear for their sound lounges.

**So?






This whole issue of being able to provide sufficient current is
not a big deal. a dozen KT88/6550 when matched properly to a load with
an
OPT
will provide all the sufficient current that could ever be wanted.

**And the cost would be? And the output impedance would be?

Cost has got nothing to do with performance.

**Cost has EVERYTHING to do with performance. Just look at the comparative
popularity of a Porsche 968 and a Subaru WRX. Both have similar levels of
point to point performance, but the Subaru sells in far larger numbers.


How much does a Mark Levinson or a Krell cost?

**A lot.


You are trying to argue SS amps are the best on a tubes group.

**No. I am assisting the OP in his quest to power a pair of ancient Infinity
RSIIIb speakers. For HIS, SPECIFIC appliaction, a good, high current SS amp,
is the best, least expensive option.


Its like trying to sell pork in a synagogue.

**Perhaps. However, I prefer to beleive that not everyone who posts here is
incapable of logical, rational thought.





Before last christmas I designed and built an SS amp for a client just
to
bring
myself
up to date about what is possible, and what is not, with SS amps.
with 10 output bjts each with 15 amp peak current ability, so 75 peak
amps
is
possible and it was capable
of about 400 watts *continuous* into 2 ohms.
400w into 2 ohms is only 28 vrms, or 14 amps rms.

My 300 watt tube amps could easily be configured for a load to easily
provide 14
Arms,
simply because they are capable of 400 watts into 2 ohms.
I am using a pile of local CFB in the output stage,
( and I doubt you have the slightest idea what that really means, )
so the power before getting anywhere near clipping is clean.

Just don't try to say tube amps can't do it Trevor,
They can, and they do.

**And the cost difference is?

Cost is not a consideration.

**Cost is ALWAYS a consideration. ALWAYS.







And who would ever need this high current?

**Asked and answered.


I can't think of a situtation, because regardless of load value,
even with speakers of 86 dB/W/M, 400 watts x two channels gives an SPL
of 121 dB, which would have the police around to make an arrest.


The impedance curve is most misleading.

**No, it is not. It is exactly what you see.

I don't believe you, and neither should anyone else.

There is no identifying info on the document.

No person has signed it after checking several samples.

Only a present day measurement would satisfy me.

**I don't give a **** what would satisfy you, Patrick. If you wish to
measure the Kappa 9 and prove me wrong, then feel free to do so. Until
then,
you would appear to be full of **** and wind.





There is no identification of exactly what speaker it is for,
and is it impedance?, or phase shift?

**I thought I explained that. It is impedance. LMS displays both
impedance
and pahse shift, or either one. I deleted the pahse shift information
in
the
interests of clarity.

And you confused us all with a partially deleted document.

**No Patrick. Only YOU were confused. Everyone else managed to work it
out.

Nope, I challenged the validity of the impedance graph.

**Good. Then YOU provide your alternate graph. I'll wait.

Everyone is now doubtful about what you have provided us.

**You speak for EVERYONE? Curious.




Please try to be slightly more convincing than you are.






If it is impedance, lowest Z = 0.8 ohms at 25Hz,
which seems utterly impossible,
since the DCR would surely make Z minimum at least = DCR, no?

**Undoubtedly.

What is the measured DCR for the Infinitys?

**Dunno. Don't care. I know of no music which extends down to DC. I
measure
speakers over the range of human hearing and sometimes beyond.

You should care about the DCR.

**Why? How much music goes down to DC?


The low bass Z of the speakers you referred us to, whatever they were,
went down to 0.8 ohms at some bass F, which is strange since
most speakers have resonant impedance oeaks at LF, and hence a curve
quite unlike the one you tried to trick us with.

**I tried to trick no one. I measured a pair of speakers. I published the
plot. Nothing more. Moreover, if you had ANY knowledge of the Kappa 9 (and
the RSIIIb) you would understand that the plot I presented, is, indeed, very
possible.

I suggest that before you launch further attacks, you do some basic research
on the speakers in question. Measuring the impedance would be a good start.

--
Trevor Wilson
www.rageaudio.com.au