flipper wrote:

What are the disadvantages to using an output load different than
specified to get a different primary impedance?

For example. Say you want a 5k primary to 8 ohm but have a 2k5 to 4
ohm. Impedance wise it matches to just use an 8 ohm speaker on the 2k5
to get 5K, so what is(are) the reason(s) not to?

The transformer is specified for a specific RLa-a to secondary because the
information includes
that the primary inductance and winding losses will suit the RLa-a.
So a tranny with 2.5k primary will have 1/2 the inductance of the one with
5ka-a.
LF Stability with NFB could be an issue.

The 2.5k : 4 tranny may have 0.7 times the primary turns of the 5k tranny
because inductance
varies with the square of P turns. Therefore the 2.5k tranny may have
lower winding losses
when used for 5k : 8 .
The above is only valid for where the trannies are rated for the same
power.

In practice the 2k5 : 4 may be fine.

The ARC VT100 has load matches form 1.6k : 4 and 8 ohms.
When 8 ohms is used on the 4 ohm tap, the load the 4 x 6550 tubes sees is
3.2k, so the tubes are loaded
with 6.4k per pair which is the better load for more class A and less
distortion and lower Rout.
8 ohm speakers should only ever be used with the 4 ohm outlet in this amp.

I measured the power with 8 ohms on the 4 ohm tap to be 80 watts which is
still plenty.




And vices versa. Say you want 2k5 to 4 but have a 5k to 8. Why not
just use the 5k with a 4 ohm load, for 2.5k primary?

Winding losses will be twice as high when a 5k : 8 OPT is subsituted for
2.5k : 4.
Also leakage inductance can make HF losses worse.

Don't do this.

Use the right load for the tubes.

Patrick Turner



Or is it ok?

flipper wrote:

On Tue, 20 Jun 2006 14:28:02 GMT, Patrick Turner
wrote:



flipper wrote:

What are the disadvantages to using an output load different than
specified to get a different primary impedance?

For example. Say you want a 5k primary to 8 ohm but have a 2k5 to 4
ohm. Impedance wise it matches to just use an 8 ohm speaker on the 2k5
to get 5K, so what is(are) the reason(s) not to?

The transformer is specified for a specific RLa-a to secondary because the
information includes
that the primary inductance and winding losses will suit the RLa-a.

Ah. Yes, I suspected something like that.

Is there some rather 'typical' guideline they may have gone by for low
end radio style OPTs that I could use to rough guess what the
inductance and winding resistance of an unknown OPT, other than just
2K5/3,2 (plus maybe 2 watt, or 3 watt), might be?

So a tranny with 2.5k primary will have 1/2 the inductance of the one with
5ka-a.

I presume, from your comments below, that they'd lower the primary
winding for the 2K5, rather than adjust the secondary, to reduce the
winding resistance. Cheaper than thicker wire.

LF Stability with NFB could be an issue.

Because there's less inductance to the 2K5?

The 2.5k : 4 tranny may have 0.7 times the primary turns of the 5k tranny
because inductance
varies with the square of P turns. Therefore the 2.5k tranny may have
lower winding losses
when used for 5k : 8 .
The above is only valid for where the trannies are rated for the same
power.

In practice the 2k5 : 4 may be fine.

Hmm. OK. Wouldn't the low end suffer from the lower inductance,
though?

The ARC VT100 has load matches form 1.6k : 4 and 8 ohms.
When 8 ohms is used on the 4 ohm tap, the load the 4 x 6550 tubes sees is
3.2k, so the tubes are loaded
with 6.4k per pair which is the better load for more class A and less
distortion and lower Rout.
8 ohm speakers should only ever be used with the 4 ohm outlet in this amp.

I measured the power with 8 ohms on the 4 ohm tap to be 80 watts which is
still plenty.


And vices versa. Say you want 2k5 to 4 but have a 5k to 8. Why not
just use the 5k with a 4 ohm load, for 2.5k primary?

Winding losses will be twice as high when a 5k : 8 OPT is subsituted for
2.5k : 4.
Also leakage inductance can make HF losses worse.

Those both make sense to me.

Seems to me that the low end might be boosted from the higher
inductance, though. Or am I off base?

Don't do this.

Use the right load for the tubes.

I don't follow. 2K5 would be the 'right load'.


Patrick Turner

Thanks. This is the kind of information I was looking for.

The P inductance of good OPT is such that its reactance is equal to the RLa-a
at
less than 14 Hz and at only 5Vrms a-a of signal.

An original Williamson OPT had 600H max at full power but only 100H at 5Vrms.
The reactance of 100H is equal 10kohms at 16Hz, and driven by triodes the
-3dB point at very low levels is about 4Hz.
Fsat at full power is around 10Hz; 16watts, 10ka-a

During the years between 1947 and now, amp makers have routinely ignored
most of what Williamson stipulated would make a good OPT and have equiped their
amps with
25H instead of 100H, and not worried thet Fsat is maybe 30Hz or above at full
power.

Makers take the cheap option.

If one uses 5ka-a on a Williamson, more than enough inductance exists,
and if one uses 20ka-a, barely enough exists for stability reasons alone.
But from a response viewpoint, the amount of L is ample.

Once you understand basic LCR theory and can analyse a simple experimental LCR
arrangement which represents the output tubes and the OPT in its simplest model
then you
won't have to ask so many questions that have been asked 100 times here before.

Patrick Turner.

flipper wrote:

On Wed, 21 Jun 2006 07:29:32 GMT, Patrick Turner
wrote:



flipper wrote:

On Tue, 20 Jun 2006 14:28:02 GMT, Patrick Turner
wrote:



flipper wrote:

What are the disadvantages to using an output load different than
specified to get a different primary impedance?

For example. Say you want a 5k primary to 8 ohm but have a 2k5 to 4
ohm. Impedance wise it matches to just use an 8 ohm speaker on the 2k5
to get 5K, so what is(are) the reason(s) not to?

The transformer is specified for a specific RLa-a to secondary because the
information includes
that the primary inductance and winding losses will suit the RLa-a.

Ah. Yes, I suspected something like that.

Is there some rather 'typical' guideline they may have gone by for low
end radio style OPTs that I could use to rough guess what the
inductance and winding resistance of an unknown OPT, other than just
2K5/3,2 (plus maybe 2 watt, or 3 watt), might be?

So a tranny with 2.5k primary will have 1/2 the inductance of the one with
5ka-a.

I presume, from your comments below, that they'd lower the primary
winding for the 2K5, rather than adjust the secondary, to reduce the
winding resistance. Cheaper than thicker wire.

LF Stability with NFB could be an issue.

Because there's less inductance to the 2K5?

The 2.5k : 4 tranny may have 0.7 times the primary turns of the 5k tranny
because inductance
varies with the square of P turns. Therefore the 2.5k tranny may have
lower winding losses
when used for 5k : 8 .
The above is only valid for where the trannies are rated for the same
power.

In practice the 2k5 : 4 may be fine.

Hmm. OK. Wouldn't the low end suffer from the lower inductance,
though?

The ARC VT100 has load matches form 1.6k : 4 and 8 ohms.
When 8 ohms is used on the 4 ohm tap, the load the 4 x 6550 tubes sees is
3.2k, so the tubes are loaded
with 6.4k per pair which is the better load for more class A and less
distortion and lower Rout.
8 ohm speakers should only ever be used with the 4 ohm outlet in this amp.

I measured the power with 8 ohms on the 4 ohm tap to be 80 watts which is
still plenty.


And vices versa. Say you want 2k5 to 4 but have a 5k to 8. Why not
just use the 5k with a 4 ohm load, for 2.5k primary?

Winding losses will be twice as high when a 5k : 8 OPT is subsituted for
2.5k : 4.
Also leakage inductance can make HF losses worse.

Those both make sense to me.

Seems to me that the low end might be boosted from the higher
inductance, though. Or am I off base?

Don't do this.

Use the right load for the tubes.

I don't follow. 2K5 would be the 'right load'.


Patrick Turner

Thanks. This is the kind of information I was looking for.

The P inductance of good OPT is such that its reactance is equal to the RLa-a
at
less than 14 Hz and at only 5Vrms a-a of signal.

An original Williamson OPT had 600H max at full power but only 100H at 5Vrms.
The reactance of 100H is equal 10kohms at 16Hz, and driven by triodes the
-3dB point at very low levels is about 4Hz.
Fsat at full power is around 10Hz; 16watts, 10ka-a

During the years between 1947 and now, amp makers have routinely ignored
most of what Williamson stipulated would make a good OPT and have equiped their
amps with
25H instead of 100H, and not worried thet Fsat is maybe 30Hz or above at full
power.

Makers take the cheap option.

If one uses 5ka-a on a Williamson, more than enough inductance exists,
and if one uses 20ka-a, barely enough exists for stability reasons alone.
But from a response viewpoint, the amount of L is ample.

Once you understand basic LCR theory and can analyse a simple experimental LCR
arrangement which represents the output tubes and the OPT in its simplest model
then you
won't have to ask so many questions that have been asked 100 times here before.

Patrick Turner.

I doubt makers of the typical table radio were seeking Williamson
level 'hi-fi' but thanks anyway. Sorry to have bothered you.

Makers of table radios were very numerous and competiton was fierce so they all
compromised
quality as much as they could get away with to keep costs of production low and
allow their shareholders,
mainly rich *******s, to enjoy a large house and car.

But the LCR theory applies to whatever they did make.

Patrick Turner.

flipper wrote:

On Thu, 22 Jun 2006 08:20:25 GMT, Patrick Turner
wrote:



flipper wrote:

On Wed, 21 Jun 2006 07:29:32 GMT, Patrick Turner
wrote:



flipper wrote:

On Tue, 20 Jun 2006 14:28:02 GMT, Patrick Turner
wrote:



flipper wrote:

What are the disadvantages to using an output load different than
specified to get a different primary impedance?

For example. Say you want a 5k primary to 8 ohm but have a 2k5 to 4
ohm. Impedance wise it matches to just use an 8 ohm speaker on the 2k5
to get 5K, so what is(are) the reason(s) not to?

The transformer is specified for a specific RLa-a to secondary because the
information includes
that the primary inductance and winding losses will suit the RLa-a.

Ah. Yes, I suspected something like that.

Is there some rather 'typical' guideline they may have gone by for low
end radio style OPTs that I could use to rough guess what the
inductance and winding resistance of an unknown OPT, other than just
2K5/3,2 (plus maybe 2 watt, or 3 watt), might be?

So a tranny with 2.5k primary will have 1/2 the inductance of the one with
5ka-a.

I presume, from your comments below, that they'd lower the primary
winding for the 2K5, rather than adjust the secondary, to reduce the
winding resistance. Cheaper than thicker wire.

LF Stability with NFB could be an issue.

Because there's less inductance to the 2K5?

The 2.5k : 4 tranny may have 0.7 times the primary turns of the 5k tranny
because inductance
varies with the square of P turns. Therefore the 2.5k tranny may have
lower winding losses
when used for 5k : 8 .
The above is only valid for where the trannies are rated for the same
power.

In practice the 2k5 : 4 may be fine.

Hmm. OK. Wouldn't the low end suffer from the lower inductance,
though?

The ARC VT100 has load matches form 1.6k : 4 and 8 ohms.
When 8 ohms is used on the 4 ohm tap, the load the 4 x 6550 tubes sees is
3.2k, so the tubes are loaded
with 6.4k per pair which is the better load for more class A and less
distortion and lower Rout.
8 ohm speakers should only ever be used with the 4 ohm outlet in this amp.

I measured the power with 8 ohms on the 4 ohm tap to be 80 watts which is
still plenty.


And vices versa. Say you want 2k5 to 4 but have a 5k to 8. Why not
just use the 5k with a 4 ohm load, for 2.5k primary?

Winding losses will be twice as high when a 5k : 8 OPT is subsituted for
2.5k : 4.
Also leakage inductance can make HF losses worse.

Those both make sense to me.

Seems to me that the low end might be boosted from the higher
inductance, though. Or am I off base?

Don't do this.

Use the right load for the tubes.

I don't follow. 2K5 would be the 'right load'.


Patrick Turner

Thanks. This is the kind of information I was looking for.

The P inductance of good OPT is such that its reactance is equal to the RLa-a
at
less than 14 Hz and at only 5Vrms a-a of signal.

An original Williamson OPT had 600H max at full power but only 100H at 5Vrms.
The reactance of 100H is equal 10kohms at 16Hz, and driven by triodes the
-3dB point at very low levels is about 4Hz.
Fsat at full power is around 10Hz; 16watts, 10ka-a

During the years between 1947 and now, amp makers have routinely ignored
most of what Williamson stipulated would make a good OPT and have equiped their
amps with
25H instead of 100H, and not worried thet Fsat is maybe 30Hz or above at full
power.

Makers take the cheap option.

If one uses 5ka-a on a Williamson, more than enough inductance exists,
and if one uses 20ka-a, barely enough exists for stability reasons alone.
But from a response viewpoint, the amount of L is ample.

Once you understand basic LCR theory and can analyse a simple experimental LCR
arrangement which represents the output tubes and the OPT in its simplest model
then you
won't have to ask so many questions that have been asked 100 times here before.

Patrick Turner.

I doubt makers of the typical table radio were seeking Williamson
level 'hi-fi' but thanks anyway. Sorry to have bothered you.

Makers of table radios were very numerous and competiton was fierce so they all
compromised
quality as much as they could get away with to keep costs of production low and
allow their shareholders,
mainly rich *******s, to enjoy a large house and car.

Yeah. I know. It's a crime they didn't make radios so expensive no one
could afford them and the lack of sales would have at least put the
"rich *******s" in tenants too so all would be swank with the world.

But the LCR theory applies to whatever they did make.

I already understand LCR theory, it's what their targets were I don't
know, but thanks again.


Maybe you do know a bit but OPTs are a big mystery to a lotta folks.
But they are just a mass of L,C&R.

A 5k to 8 OPT has a Z ratio of 625 :1 and a
transformer of 625 ohms to 1 ohm couldn't be used in an amp wanting
a 5k primary if you had 8 ohms used as the load on the 625 : 1 tranny
because convention assumes the inductance would be much lower for a 625 ohm primary.

As soon as you know the primary impedance match for an OPT you should know there will
be a certain
inductance which allows a response to somewhere below 100Hz.
Just how much Lp there is determines the hi in the fi with regard to bass.

The leakage inductance also relates itself to primary inductance for a given geometry.
and turns used. So if Lp is doubled when the turns are increased around a given core
by a factore of 1.41, then so is the LL. Si increasing P turns lowers the LF cut off
and
lowers the HF cut off; it moves the whole band span down;

Patrick Turner.


Patrick Turner.