• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

What is a decent transformer company available in the US and where to buy them?

My experience with tubes is not as extensive as the others here. Compared to Lundahl AM, vintage Klangfilm, Stancors and Hashimoto, the Edcor's can certainly stand on their own. The price difference is a no-brainer, if you like to try different designs and keep building (the hobby gets very expensive, if you know what I mean).

The Lundahl AMs take a long time to sound nice. Break-in or whatever.

I'm running Altec 9849 (414+802 on Azura) and 604e. Maybe other speakers will sound different.
 
I purchased the cheap 5H Edcors mentioned in the article for use in Tubelab’s SE amplifier using 45s. I’m adding a 200Hz second order high pass to the input of the amplifier, along with a pass-through so I can send the full-range signal to an active midbass cabinet.
 
Define “poor”? Do you mean inadequate?

Please see George’s distortion measurements:

Budget Output Transformers | Tubelab

I’ve never seen double digit distortion with any of these small transformers with “poor” inductance.

The ratio between inductive reactance XL and equivalent resistance Req IS the measure of how much inductance you need at a given low frequency.

XL=2*pi*f*Lp where Lp is the primary inductance

Req = (Rload + RDC) //Rplate

To get low distortion XL >> Req (the much bigger simbol >> means at least 6-8 times, more is better)

This is relatively easy down to about 40Hz with triodes. At 30Hz it is only possible with low plate resistance tubes like 300B, 2A3 etc...with other tubes with Rp in the 1K-2K range it can be still fairly good.


The 5H people talk about in that blog for a 5K transformer is poor for 30Hz or less except for tubes that have 200R or, possibly, lower plate resistance. Sometimes transformers have better inductance than declared sometimes not. This is obviously related to actual inductance, not catalogue numbers....
 
Last edited:
No, it is your opinion that the writers assumptions are flawed and that he doesn’t go far enough in his explanation. However, It was written by an engineer who makes a solid case that chasing the last octave of low frequency performance can lead to unnecessary compromises.

The article has to be taken with the concept of severalty: he makes multiple and distinct arguments for using lower inductance transformers.

Chris, if you really want “All good fortune” then answer the OPs question (which I don’t see that you have).

🙄

It's not an opinion. The primary inductance appears in shunt. The higher the better. Take any good book, look at the model for low frequency of the transformer. In order to be "transparent" is has to be as high as possible.
It has nothing to do with transient response. High inductance is NECESSARY to minimize excitation current and minimize distortion (which involves the tube as well).

Frequency response is really a useless parameter to look at. If want to cut response to avoid issues better to do it in another part of the amp where there is no such problem. In other words, limiting the low frequency response with the output transformer is not a good idea.
 
Last edited:
Budget Output Transformers | Tubelab

I’ve never seen double digit distortion with any of these small transformers with “poor” inductance.

No? That test shows precisely what I say. Take the UBT as it has been mentioned here too. The amplifier cannot deliver full power at 30Hz despite it is sold as a 15W transformer. That because of its inductance and/or saturation.
Put 5W through at 30Hz and see what you get....
Besides response is 3 dB down at 19Hz (the power supply also plays a role here) which means you have not given up the first octave at all.

All this for a 3K OPT with 17H (nominal) primary inductance which is not low at all, generally speaking.

Transformer - One Electron, Output, 3000 Ω to 16/8/4 Ω | Amplified Parts

Certainly not like a 5K transformer with 5H as in that blog you posted earlier. That would be double digits at higher frequency. Granted.
 
Last edited:
It's all in the specs.
The Edcor XSE15-8-5k with 5H of primary inductance: manufacturer spec 70Hz - ...
George measured core saturation with this OPT at 5W at low frequencies:
Low frequency effects at 5 watts: Visible saturation can be seen on the scope for any frequency below 74 Hz at 50 mA, 61Hz at 80mA.
So don't use this OPT full range, but use it like Brinkman does.
The Cascade Tubes article is a joke; luckily he ends with his theory being a thesis.
 
Last edited:
You know how you're clever and everything.. can you provide a real working example please.
Thanks.

First this method is just an estimate of what you need, real life is always a bit more complicated but it does provide a safe starting point anyway.

A real working case, what do you mean? Do you want to see how calculate for a real amplifier before buying the actual transformer?
 
To be fair, I dont think there is much that you can teach me. Happy to learn, but, cmon this is all first year science at any university, or less.
My question has roots back to what you wrote before about load impedance, resistance and inductive reactance wrt frequency.
Knock yourself out, please.
 
Last edited:
The picture can be considered a "universal" behaviour of distortion as function of total induction for different values of XL/Req.
Ra is approximated with load resistance reflected to the primary for simplicity and DC resistance of the windings is not taken into account. However things are not much different and, in case the DC resistance is excessive, things get worse...
 

Attachments

  • THDvsXLsuReq.JPG
    THDvsXLsuReq.JPG
    90.2 KB · Views: 210
Of course it is related to a core made of FeSi grain oriented laminations with a gap and DC induction is optimized for max headroom (i.e. DC induction is equal to or just above 50% of the total relative to the working conditions and output power level)
 
Last edited:
What is Req?
Back to post #27.

Again? It's written in the picture.....

Take a transformer with 25:1 turn ratio, primary DC resistance 125R, secondary DC resistance 0.2R. For 8R resistive load the reflected primary impedance is 5K.
Then pick your tube and say you want to use a 300B at 400V/75mA where its plate resistance is 750R.

Then Ra in that formula, also including the DC resistance of both primary and secondary, would be:

Ra= 5K + 125R + (25*25*0.2)= 5250R

Ri = 750R

then Req =5250 // 750 = 656.25 R

If ignore the DC resistance of the windings to go faster Req = 652.17 R....very little difference as one only uses this formula for an estimate.

Then if you want stay on the curve where a ratio of 8 between XL and Req and decides to have it at 30Hz:

(2*pi*30*L)/656.25=8

From there one can calculate L = 27.85 H
 
There was a link to blog where there is a "theory" about better behaviour of small transformers with low inductance. They should be better because transient response is better (but that is not the case, unfortunately) so this would be a good choice, giving up low frequency response. The primary inductance appears in shunt and so the higher the better and there is no such impact on transient response at all.
In practice one has to find the best compromise among all parameters as usual....
 
No? That test shows precisely what I say. Take the UBT as it has been mentioned here too. The amplifier cannot deliver full power at 30Hz despite it is sold as a 15W transformer. That because of its inductance and/or saturation.
Put 5W through at 30Hz and see what you get....
Besides response is 3 dB down at 19Hz (the power supply also plays a role here) which means you have not given up the first octave at all.

All this for a 3K OPT with 17H (nominal) primary inductance which is not low at all, generally speaking.

Transformer - One Electron, Output, 3000 Ω to 16/8/4 Ω | Amplified Parts

Certainly not like a 5K transformer with 5H as in that blog you posted earlier. That would be double digits at higher frequency. Granted.

No, you clearly didn’t pay attention and failed to read the Tubelab test. The Edcor XSE15-8-5K has a frequency extension of 70-18k. It’s really best suited for a guitar amp.

The Edcor products with 40hz extension are an excellent option. That’s what the poster asked for and I answered it.

Raise your hand if you’ve built and tested with Edcor, One Electron and Electra-print.

I’m raising mine, are you???