Valve (tube) speaker transformer or smoothing choke idea

Ivan wrote:Hi Harry,
"Two 24V "christmas-tree" transformers ($5 ea) are a LOT cheaper than a retro valve output transformer ($45)."
I am confused now. How do you intend to use these 230/24 V (or 12 V) transformers? When used as an output transformer, it has no winding suitable for DC compensation. When used as a DC choke, the core will probably saturate and the inductance will drop.

VBR from Ivan

Hi Ivan,
A 230V to 12V (12-0-12 with one half used) is about the correct truns ratio to match an EL84 to 16Ω speaker. The transfomer does have to be modified in that the laminations are removed and re-inserted in the same direction, with the mag break, but there is no winding involved. The coils are already there. But this is only ok for comms quality, not really for HiFi. But it is a quick'n dirty solution.

Ivan wrote:Hi Harry,
BTW, I am pretty unsure how to make a compensated transformer with a tap for the 2nd grid of a pentode.

VBR from Ivan

I am not so sure either, but I will look at the impedances of the anode and G2, then make a tap accordingly, see how that works.

Ivan wrote:Hi Harry,
"Also, to get a good linearity, I believe one would not need to cancel 100% of the DC, just 50% would make a big difference."
It is important to keep the core flux in the linear part of the maiden magnetizing curve of the core. If the flux gets to the region, where the curve bends or even further, it causes not only distortion. The inductance and power transfer drop suddenly. All depends on the material and cross-section of the core.

VBR from Ivan

Yes, Ivan, those points are really relevant, but I do not think that the transfer function will have a hard limit.
When I have got a basic amp working I intend to do some experiments with this and see what can be done, how far I can push it.



I have wanted to do more experiments with tubes for a very long time, and I am hoping to get a basic working chassis I can tinker with, before the good antenna weather arrives :-)

John_1981 wrote:It may be cheating, but could you fit a transistor based active load in the anode circuit, then capacitively couple into the audio transformer to remove the DC component? I have seen this approach with transistor based class A amplifiers and think it would be a small price to pay to avoid potential core saturation?

High voltage PNP transistor, pair of diodes or an LED as a reference voltage and a couple of biassing resistors should do the trick.

Hi John,
That sounds like a clever approach. Just for experience sake it would be worth a try to see how it compares with a "straight" load.
Thank you very much for sharing the information. It is certainly worth investigating.


However, if I am to replicate a "Narshall" amplifier, the purists would step back in horror. At the moment I am having long chats with a guy at work who was in a rock group in his hay-days, and we both want the same amplifier, with the only liberty taken of swapping EL34's for EL80's. My friend can work on the two cabinets and I can work on the two amplifiers.


One of the objects of the exercise is that the unit should look similar, with the fibre-glass board with components mounted horisontally in the middle, three transformer/chokes and four tubes.


Very best regards from Harry - SM0VPO

It may be cheating, but could you fit a transistor based active load in the anode circuit, then capacitively couple into the audio transformer to remove the DC component? I have seen this approach with transistor based class A amplifiers and think it would be a small price to pay to avoid potential core saturation?

High voltage PNP transistor, pair of diodes or an LED as a reference voltage and a couple of biassing resistors should do the trick.

Hi Harry,
"The opposite end of the transformer winding (-250V) would be coupled to ground through a resistor to balance the current."
The resistor will be power hungry, it will probably eat as much power as the end tube. In a network powered device it may be still acceptable.

"In addition to the resistor, a choke in series to allow the DC to pass, but block the AC component. In this way it would not affect the AC characteristics a lot but could balance out the current."
Yes, but the choke itself must be protected against the core saturation as well.

"Two 24V "christmas-tree" transformers ($5 ea) are a LOT cheaper than a retro valve output transformer ($45)."
I am confused now. How do you intend to use these 230/24 V (or 12 V) transformers? When used as an output transformer, it has no winding suitable for DC compensation. When used as a DC choke, the core will probably saturate and the inductance will drop.
Audio output transformers should be designed and wound in a special manner to minimize the influence of internal stray capacitances. AC power transformers do not deal with this problem. So the frequency response in trebles may be poor.

BTW, I am pretty unsure how to make a compensated transformer with a tap for the 2nd grid of a pentode.

"Also, to get a good linearity, I believe one would not need to cancel 100% of the DC, just 50% would make a big difference."
It is important to keep the core flux in the linear part of the maiden magnetizing curve of the core. If the flux gets to the region, where the curve bends or even further, it causes not only distortion. The inductance and power transfer drop suddenly. All depends on the material and cross-section of the core.

VBR from Ivan

Hi Ivan, thank you very much for a really fast response.

The idea I had in mind was to have a 250-0-250V to 24V audio transformer with a 250V winding in the anode of the tube (+250V).

The opposite end of the transformer winding (-250V) would be coupled to ground througha resistor to balance the current.

In addition to the resistor, a choke in series to allow the DC to pass, but block the AC component. In this way it would not affect the AC characteristics a lot but could balance out the current. Two 24V "christmas-tree" transformers ($5 ea) are a LOT cheaper than a retro valve output transformer ($45). In this way the extra winding would not short out much of the AF signal, if anything.

I like your idea of using a decoupled coil. This could also be a good idea.

Also, to get a good linearity, I believe one would not need to cancel 100% of the DC, just 50% would make a big difference.

When I build the "Narshall" amplifier I would like to do a load of experiments before I do much with my old acoustic guitar. I am certain that "modern thinking" with "old technology" can give some really significant improvements over old established designs.

Very best regards from Harry - SM0VPO

Admin wrote:Hi all,
...Would it be possible to add a second winding and add DC to it, such that it cancels much of the standing magnetic field created in the first coil?...

A push-pull speaker transformer carries a standing DC from two tubes, but each 1/2 of the winding has the standing current with the opposite polarity. In this way one does not need to worry so much about a magnetic breaker between the E and I laminations (the 0.1mm thick piece of paper).

But a single-ended speaker transformer in class-A only carries the standing DC from one tube. Ok, the second coil may increase the load a little and reduce the output level to the speaker. I think it could be a good experiment for acceptable loss and saturation.

If this works then I could use a "bog-standard" mains transformer (230V AC to 24V AC = about 8000Ω to 16Ω) without modification when used as a speaker transformer. I don't think it would need a total cancellation, just enough to prevent the core from saturation (instead of over -dimensioning the core size). The only problem I can see with this is the frequency response if the transformer is designed for 50Hz will have a low inductance at 20Hz.

Thoughts anyone?

BR Harry - SM0VPO

Hi Harry,
it is an interesting idea, not stupid at all.
IMHO the compensation coil must have at least the same inductance as the main (primary) winding. If the compensation winding had much less turns and you wanted to increase its magnetic field by increasing the DC durrent, this winding would short a great deal of the AC component.
My first idea is to make an output transformer with two identical windings (bifilar would be the best), connected in the opposite direction - see the picture. One of the windings would be bridged by a big enough capacitor. The DC component of the anode current would flow through both windings, compensating the DC magnetic field in the core. The AC component of the anode current would flow mostly through the capacitor and one winding, magnetizing the core and passing the power to the secondary.
A standard mains transformer is not suitable for this. A mains transformer with two 110V windings would be nice!
The frequency response of tube amplifiers below 50 Hz is often limited in any case.

VBR from Ivan

Hi all,
This may sound like a stupid idea, but one of the problems with transformers and chokes that carry DC is that of there is a standing magnetic field. This can cause the iron core to become saturated.

Take the case of a 5 Henry PSU choke carrying 100mA of current.
Would it be possible to add a second winding and add DC to it, such that it cancels much of the standing magnetic field created in the first coil?

A push-pull speaker transformer carries a standing DC from two tubes, but each 1/2 of the winding has the standing current with the opposite polarity. In this way one does not need to worry so much about a magnetic breaker between the E and I laminations (the 0.1mm thick piece of paper).

But a single-ended speaker transformer in class-A only carries the standing DC from one tube. Ok, the second coil may increase the load a little and reduce the output level to the speaker. I think it could be a good experiment for acceptable loss and saturation.

If this works then I could use a "bog-standard" mains transformer (230V AC to 24V AC = about 8000Ω to 16Ω) without modification when used as a speaker transformer. I don't think it would need a total cancellation, just enough to prevent the core from saturation (instead of over -dimensioning the core size). The only problem I can see with this is the frequency response if the transformer is designed for 50Hz will have a low inductance at 20Hz.

Thoughts anyone?

BR Harry - SM0VPO