[Ernst]

Hello All,

I'm about to kick off the subject project. The plan includes Hammond
1650W output transformers. These are rated at 1900 ohms plate-to-
plate. The manufacturer recommends several output tube types, in
varying complements. Considering recommended data for pairs of tubes
in AB push-pull, it seems that only 6146 or 5933 (807 variant) types
would come close to the appropriate plate load when used in multiple
pairs.

I understand that a couple of the rules of thumb for output
transformers a

To have enough inductance, and therefore load at the lowest frequency
of operation for the application.

That the plate load is a function of the reflected speaker load.

That said, I would like to use EL34/6CA7, KT-series, or TT21/TT22
tubes because of their favorable THD characteristics.

Am I off-base here, or should I stick to the numbers?

Regards,

Ernst

[Ernst]

On Feb 18, 4:05 pm, "Bret Ludwig" wrote:

You are ahead of the game but 250 watts is going to be a lot of
tubes, and to be honest most Hammond transformers are not really
audiophile grade.

The 807, 6146, or some sweep tubes are fine if operated in regulated
screen pure pentode or with a screen winding for effective ultralinear
operation at a DC offset from the plates.

Run the tubes conservatively according to book numbers and they will
give long life and performance.

Bret, I appreciate the insight. Regarding the Hammond transformers,
in what manner do they fall short? I had looked at Plitron toroidals,
which are spendy and would require pairs to approach the desired power
level. Handwound units utilizing Amidon cores were considered (as if
I have the patience for that exercise). What brand do you suggest?

Ernst

[Phil Allison]

"Ernst"

I'm about to kick off the subject project. The plan includes Hammond
1650W output transformers. These are rated at 1900 ohms plate-to-
plate. The manufacturer recommends several output tube types, in
varying complements. Considering recommended data for pairs of tubes
in AB push-pull, it seems that only 6146 or 5933 (807 variant) types
would come close to the appropriate plate load when used in multiple
pairs.

I understand that a couple of the rules of thumb for output
transformers a

To have enough inductance, and therefore load at the lowest frequency
of operation for the application.

That the plate load is a function of the reflected speaker load.

That said, I would like to use EL34/6CA7, KT-series, or TT21/TT22
tubes because of their favorable THD characteristics.

Am I off-base here, or should I stick to the numbers?



** For 250 watts rms output with lowish THD requires a minimum of 6 x 6550
/ KT88 tubes in class AB to handle the dissipation at and near near full
power.

If the OP tranny has a 1900 ohms primary, then the rms primary voltage has
to be 689 volts or 1950 volts p-p, neglecting losses.

This implies an (unregulated) DC supply of about 650 volts for the plates
and a separate, well filtered, supply of about 400 volts for the screens.

Better have a look at the schem for the Ampeg model " SVT " bass amp.

Completely nuts for home hi-fi.



........ Phil

[Mark Aitchison]

Have you considered transformerless? Yes, I know it would involve a
l-o-t of valves, but you're probably going to have to use a lot anyway.

OTL 250W into 8 ohms using EL34's? You could do that for a mere $1500.
Per channel. Plus some for power supply, drivers, hardware... :-)


Mark A

[Ernst]

On Feb 19, 2:35 am, "Phil Allison" wrote:
"Ernst"





I'm about to kick off the subject project. The plan includes Hammond
1650W output transformers. These are rated at 1900 ohms plate-to-
plate. The manufacturer recommends several output tube types, in
varying complements. Considering recommended data for pairs of tubes
in AB push-pull, it seems that only 6146 or 5933 (807 variant) types
would come close to the appropriate plate load when used in multiple
pairs.

I understand that a couple of the rules of thumb for output
transformers a

To have enough inductance, and therefore load at the lowest frequency
of operation for the application.

That the plate load is a function of the reflected speaker load.

That said, I would like to use EL34/6CA7, KT-series, or TT21/TT22
tubes because of their favorable THD characteristics.

Am I off-base here, or should I stick to the numbers?

** For 250 watts rms output with lowish THD requires a minimum of 6 x 6550
/ KT88 tubes in class AB to handle the dissipation at and near near full
power.

If the OP tranny has a 1900 ohms primary, then the rms primary voltage has
to be 689 volts or 1950 volts p-p, neglecting losses.

This implies an (unregulated) DC supply of about 650 volts for the plates
and a separate, well filtered, supply of about 400 volts for the screens.

Better have a look at the schem for the Ampeg model " SVT " bass amp.

Completely nuts for home hi-fi.

....... Phil


Thanks Phil

[Ernst]

OTL 250W into 8 ohms using EL34's? You could do that for a mere $1500.
Per channel. Plus some for power supply, drivers, hardware... :-)

Well, only $1500 per side? Maybe I should consider making this a home
theater effort instead!!

grin

Ernst

[Patrick Turner]

Ernst wrote:

Hello All,

I'm about to kick off the subject project. The plan includes Hammond
1650W output transformers. These are rated at 1900 ohms plate-to-
plate. The manufacturer recommends several output tube types, in
varying complements. Considering recommended data for pairs of tubes
in AB push-pull, it seems that only 6146 or 5933 (807 variant) types
would come close to the appropriate plate load when used in multiple
pairs.

I understand that a couple of the rules of thumb for output
transformers a

To have enough inductance, and therefore load at the lowest frequency
of operation for the application.

That the plate load is a function of the reflected speaker load.

That said, I would like to use EL34/6CA7, KT-series, or TT21/TT22
tubes because of their favorable THD characteristics.

Am I off-base here, or should I stick to the numbers?

Regards,

Ernst

Feel free to use the information provided for the 300W amps shown at

http://www.turneraudio.com.au/300monobloc.html

There are following pages with full schematics at the same website.

The 300W amps are based around cathode feedback OPTs,
but you could leave all that out and just use plain UL and fixed bias
with 12 bias pots if you are brave.
I used plain UL with partial cathode bias at first in the first 300W amp
models but
in my later model i have settled for the CFB of the output stage, like
Quad-II on several bottles
or steroid pills.

The Quad-II basic output circuit is the best way to use beam tetrodes or
pentodes IMHO,
but few companies make an off-the-shelf OPT with provision for CFB, and
most certainly not
at 250W ability.

For a special, go perhaps to Sowter, I am too busy for the next 6 mths.

UL is next best, triode is more better for some who don't mind the
inefficiency,
and plain beam tetrode or pentode can be good if the class A % is kept
high and lots more
global NFB is used.

Happy soldering.

Patrick Turner.

[Patrick Turner]

Ernst wrote:

On Feb 18, 4:05 pm, "Bret Ludwig" wrote:

You are ahead of the game but 250 watts is going to be a lot of
tubes, and to be honest most Hammond transformers are not really
audiophile grade.

The 807, 6146, or some sweep tubes are fine if operated in regulated
screen pure pentode or with a screen winding for effective ultralinear
operation at a DC offset from the plates.

Run the tubes conservatively according to book numbers and they will
give long life and performance.

Bret, I appreciate the insight. Regarding the Hammond transformers,
in what manner do they fall short? I had looked at Plitron toroidals,
which are spendy and would require pairs to approach the desired power
level. Handwound units utilizing Amidon cores were considered (as if
I have the patience for that exercise). What brand do you suggest?

Ernst

If all you want is un-refined power, a quad of TT21 will give up to 280
watts AB2.

If you want 20dB less thd/imd, a dozen such tubes or KT88 are required.

Patrick Turner.

[Patrick Turner]

Mark Aitchison wrote:

Have you considered transformerless? Yes, I know it would involve a
l-o-t of valves, but you're probably going to have to use a lot anyway.

OTL 250W into 8 ohms using EL34's? You could do that for a mere $1500.
Per channel. Plus some for power supply, drivers, hardware... :-)

Mark A

It depends on the class A content and load matching required.

For a 5k load for each EL34 in triode class A takes 625 tubes for OTL if
the load is 8 ohms.

But then you could get 3,000 watts, at about 5 watts per tube.
You'd need 12,000 watts of input power.

A single pair of 845 0r 211 should do 250 watts AB fairly easy, but not
OTL.

Patrick Turner.

[Mark Aitchison]

Bret Ludwig wrote:
OTL is nuts at this _sustained_ power level.

By the way, in case it wasn't obvious in my post, I do think OTL valve
amps over about 100W are extremely impractical - obviously the cost
being the big factor. The heat - well, that might be an advantage in
some cases (e.g. if you want a combined home theatre and sauna)!

But notice there are some highish power OTL amps out there, e.g.
Transcendent's "The Beast" with 160 watts RMS into 8 ohms using EL509's,
at about $2400 *without the tubes* for two monoblocks. Using tubes not
designed for high quality audio, or using audio tubes under conditions
sub-optimal for linearity (as I was using for my calculation of under
$1500 for one 250W channel of EL34's), is a compromise that makes me
think that tubes are being used where transistors would do much better -
yet the tubes here probably aren't giving the classic "tube" sound - so
high-power OTL must be a case of using tubes just for the ability to say
"it uses tubes" rather than for any audible or technical quality.

Still, if I was going to go mad and make a high power OTL amplifier as
some sort of dare, I think it might be feasible to use Ian Hegglun's
idea for square-law amplifiers, stick to class A even, and do the whole
thing for about $600 per channel if the right tubes are used. I have a
bunch of PL84's, so I might make a scaled-down version to see if it works.

But, back to sanity, the idea of a good 200-400W tube amp project is
very worthwhile, and something that has been largely neglected because
of somewhat strange fashions in audiophile circles.

I designed, but never built, a tube amp in this power range - with
features such as ramping down a power supply voltage when not needed.
If there is anybody willing to give it a go, let me know.

Mark Aitchison,
Christchurch, New Zealand.

[Ernst]

On Feb 19, 5:03 pm, "Bret Ludwig" wrote:

I didn't know Amidon made a tapewound core, which is the only kind
suitable at audio frequencies. I would not consider toroids for
several reasons most of which relate to DC saturation.

Very interesting. I had considered the (Amidon) cores that were rated
for line frequency power conversion, in the quest for low-frequency
response. They provided a very comprehensive tech sheet that included
every parameter under the sun that seemed to support the prospect.
Now granted, the bread and butter of this outfit is amateur radio.
That said, I had a protracted discussion with their engineering
section regarding core material, suitability to audio applications,
and (of course) power handling capabilty. The latter was centered
around if the cores they offered at the time would accomodate enough
of the gauge of wire for the anticipated current and inductance
requirements to suit a couple of hundred watts of 6146W flavored
horsepower.

The truth is, it was looking favorable. The plans were interupted by a
deployment to the Indian Ocean.

Anyway, the general theory at the time was that SE topology was
subject to DC saturation, but push-pull, due to the nature of center-
feed was not. I recall Toshiba offerings with permenant magnets,
gaps, and all sorts of tricks to counter act the saturation effect as
applied to SE.

Does any of that make sense?

Ernst

[Patrick Turner]

Ernst wrote:

On Feb 19, 5:03 pm, "Bret Ludwig" wrote:

I didn't know Amidon made a tapewound core, which is the only kind
suitable at audio frequencies. I would not consider toroids for
several reasons most of which relate to DC saturation.

Very interesting. I had considered the (Amidon) cores that were rated
for line frequency power conversion, in the quest for low-frequency
response. They provided a very comprehensive tech sheet that included
every parameter under the sun that seemed to support the prospect.
Now granted, the bread and butter of this outfit is amateur radio.
That said, I had a protracted discussion with their engineering
section regarding core material, suitability to audio applications,
and (of course) power handling capabilty. The latter was centered
around if the cores they offered at the time would accomodate enough
of the gauge of wire for the anticipated current and inductance
requirements to suit a couple of hundred watts of 6146W flavored
horsepower.

The truth is, it was looking favorable. The plans were interupted by a
deployment to the Indian Ocean.

Anyway, the general theory at the time was that SE topology was
subject to DC saturation, but push-pull, due to the nature of center-
feed was not. I recall Toshiba offerings with permenant magnets,
gaps, and all sorts of tricks to counter act the saturation effect as
applied to SE.

Does any of that make sense?

Well, Ernst, you challege our powers to see the sense you are
endeavouring to convey.

PP OPT are very prone to saturation if there is a dc imbalance in each
1/2 of the PP circuit.

SE OPT are gapped, and have nothing to balance, so DC is either more or
it is less, and
+/- 20% won't upset the sound or cause saturation with signals well
below clipping.


OPT transformer cores are only there to provide the coil with sufficient
inductance
to give a high impedance magnetic component at LF.

In a good OPT or either PP or SE type, you can remove the iron core
entirely
and have excellent response from about 5 kHz upwards, since the transfer
depends on the winding geometry
and turns and not the core material.
The core material has a declining permeability as F rises, and by 10kHz,
is very low
with ordinary Si Fe laminations. Other materials, such as tape wound
amorphous cores
have more permeability at the higher F but it isn't needed; its only
needed at lower F,
unless the OPT is made for HF only, and the turns are reduced, and there
is a dependance on the core permeability
for a required amount of inductance at the higher F.

Lundahll make OPT with amorphous cores, so go there to find out about.

Patrick Turner.



Ernst

[Ernst]

On Feb 22, 4:54 am, Patrick Turner wrote:

Well, Ernst, you challege our powers to see the sense you are
endeavouring to convey.

I do appreciate your rising to the occassion, not to mention the
pointers.

-Ernst