• 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.

Tube amp OPT for Headphone Amp? (impedance matching question)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Transformer impedance ratios always kind of mystified me. Actually transformers mystify me. But I was thinking about them and came up with a question...

Let's take a typical push-pull output transformer designed for 6V6 or EL84 output tubes, which has an 8k ohm primary and a 16 ohm tap on its secondary.

Say I wanted to make a headphone amp, and I want to use a 6SN7 as the output triode pair (push-pull).

Say I have a pair of headphones with a quoted impedance of 50 ohms.

Now, if the step-down ratio for my output transformer is 8k pri to 16R secondary, that equals 8000/16 = 500.

Taking my 50 ohm headphones as the load, 50 ohms X 500 = 25k.

Does that mean if I put those 50 ohm headphones on the 16 ohm secondary tap, the primary of the OPT would "look" like a 25k ohm load?

If yes, then wouldn't 25k be a reasonable primary load for a 6SN7 triode? (Rp for the 6SN7 is about 10k ohms, so the load will be 2.5X the Rp of the tube).

So say I happen to have a pair of 6BQ5/EL84 OPT's... Could I use them as OPT's for a headphone amp with 6SN7 triodes as the output "finals"? Power out would be very low, I know, but this is for headphones...

-=|=-
 
Now, if the step-down ratio for my output transformer is 8k pri to 16R secondary, that equals 8000/16 = 500.

500 is the impedance ratio; the step-down (winding) ratio is the root of 500 so it is 22,36.

Taking my 50 ohm headphones as the load, 50 ohms X 500 = 25k.

Does that mean if I put those 50 ohm headphones on the 16 ohm secondary tap, the primary of the OPT would "look" like a 25k ohm load?

Right.

If yes, then wouldn't 25k be a reasonable primary load for a 6SN7 triode? (Rp for the 6SN7 is about 10k ohms, so the load will be 2.5X the Rp of the tube).

So say I happen to have a pair of 6BQ5/EL84 OPT's... Could I use them as OPT's for a headphone amp with 6SN7 triodes as the output "finals"? Power out would be very low, I know, but this is for headphones...

Yes you can, but the problem will likely be that your headphone will sound bass shy.
The EL84 OPT's will have enough primary inductance for a pair of EL84's which, triode connected, have some 2k of Rp.
For PP 6SN7's with their Rp an order of about 4 higher there will not be enough primary induction, so there will be roll-off pretty much above 20 Hz.
But give it a try and listen.
 
pieter t said:
Yes you can, but the problem will likely be that your headphone will sound bass shy.
The EL84 OPT's will have enough primary inductance for a pair of EL84's which, triode connected, have some 2k of Rp.
For PP 6SN7's with their Rp an order of about 4 higher there will not be enough primary induction, so there will be roll-off pretty much above 20 Hz.
8k primary probably means pentode EL84, which has much higher anode impedance. The issue is the ratio of effective load to inductance. The effective load is the anode impedance in parallel with the transformed load impedance. As the triode has a lowish anode impedance it should be OK. The effective load will be 10k||25k=7.1k so quite near what the transformer was designed for.

The rolloff, wherever it occurs, will be below the corner frequency not above it.
 
I don't agree.
Presume the transformer has a primary inductance of 60 to 80 Hy's which it should have at least, that will simply not be enough for a 6SN7.
In my experience 6SN7's and comparable tubes for Rp (26, 76) are about the limit of what you can load by inductance, be it choke or transformer. A 6J5 (half a 6SN7) needs at least 100 Hy's not to loose bandwidth above 20 Hz.
Maybe rongon will let us know how it performs.
 
Although it's true that transformers' parasitic reactances do limit the range of impedances that they can reflect well, in this case (as DF96 carefully points out) rongon's use is quite similar to the likely original purpose. A good transformer for push-pull EL84's will have about 1000 Henry primary inductance:

Z565-48 17.5(35W) Watt Triode USA Dyna Clone Output Transformer for Dynaco MADE IN USA

Thanks to the OP for the inspiration to do something similar with an old pair of Eico ST70 OPT's and EL84's trioded. Time to get out the Greenlee punches!

Thanks,
Chris
 
Chris,

My two cents that the specification of the Z565 having 1000H at 60Hz is a typo; I am sure it must be 100H.
Primary inductance values in the hundreds of Hy's can only be seen in specifications of Plitron output transformers, where a 250 or so AC voltage (50Hz) is used to measure inductance (please note that primary inductance depends on frequency and amplitude).
Normally inductance values are measured with 1 to 5 VAC or so at 100Hz, 1kHz and sometimes 10Khz with the typical L(CR) meters.
 
Actually, y'all are probably right. It does sound pretty durned high. Inductance will increase with applied AC voltage, but ....

I've got some old style (SCA35) transformers and one really old A410. I'll try to measure some and report back, but we'll all have lost interest before I get it done. Arf!

Thanks,
Chris
 
A friend of mine is suddenly interested in headphones, and now so am I. Thanks everybody for your previous answers. Good stuff. Here's some more stuff on this...


Yes you can, but the problem will likely be that your headphone will sound bass shy.
The EL84 OPT's will have enough primary inductance for a pair of EL84's which, triode connected, have some 2k of Rp.
For PP 6SN7's with their Rp an order of about 4 higher there will not be enough primary induction, so there will be roll-off pretty much above 20 Hz.
But give it a try and listen.


OK, I'm thoroughly confused.

Let's assume the OPT primary L is only 80 Henries. So the output triode I choose needs to have rp of 2500 ohms or less so the bass goes low enough. Let's say I use a 5687 (I have lots of 'em).

So the bandwidth of the amp will be more a function of the OPT primary L than the OPT primary Z? Should I look at it like output power and distortion are related to the primary Z to rp ratio, while bass response is limited by the primary L to rp ratio?

(For max power, Zprimary should be 2*rp ?)
(For less distortion, Zprimary should 4*rp?)
(For good bass response, output triode's rp should be good and low?)

I remember asking about pri L require for good bandwidth in an SE OPT for a given output tube rp. The answer was basically 'as many Henries as you can get.' 60H is a bare minimum for triodes of the 5687, 6N30P class. More is better.

_________________________________________________________________


What about the secondary? The headphones I'd use have a rated impedance of 55 ohms. I have another pair that are 40 ohms. The headphones I really want are 300 ohms.

So let's say I use a PP pair of 5687 triodes (one envelope for each OPT).

rp = 2500 ohms
Ip = 15mA

OPT pri Z = 10k (plate to plate)
OPT pri L = 80H
OPT sec Z = 16 ohms (the 16 ohm tap)


What happens when I connect the 55 ohm headphones to the 16 ohm sec tap?
What happens when I connect the 300 ohm headphones to this?
How do those compare to connecting 16 ohm headphones to the 16 ohm tap?


I thought connecting a higher load (55 ohms) to a 16 ohm secondary causes the primary inductance to appear lower to the load. So I'd lose bass response with the 55 ohm headphones, and lose more bass response with the 300 ohm headphones. Right?

I'd lose power going into the higher loads too, correct? (Not much of a problem with headphones.)

OK, that's a lot of questions. I should get off my rear and build this thing, so I can start answering these questions.

Thanks y'all!

-=|=-
 
Rongon,

Be careful with: this transformer has a primary impedance of, for example, 10k.
Transformers as such do not have an "impedance", because it depends on what you put in front of and behind the transformer.
The most important thing to know in this respect is the winding ratio of the transformer.
Let's say we have a SE transformer with a primary to secondary winding ratio of 25:1.
When we connect an 8 ohm load to the secondary (the usual loudspeaker) the primary impedance is the reflected 8 x 25² = 5k.
But connecting a 4 ohm load to the same transformer, the primary impedance is 2k5.
So we cannot speak of transformer impedances as long as loads are not fixed.

When you speak of your PP transformer as having prim. Z of 10k p-p, and sec. Z of 16 ohms, all we know is that the winding ratio is, hey, the same 25:1.
When 10k prim. Z is specified we must believe that the induction of the primary will "fit" the source impedance in order to have enough LF bandwidth.

When you connect a 55 ohm load, the source will see a prim. Z of 34k4 a.s.o.
With this higher load you will actually gain some power, as the source "sees" a much higher impedance so gain of the source will approach mu.
Bass response will depend more on source impedance, the lower Rp tubes showing better LF bandwidth.

Hope this helps a bit.
 
Rongon,

Be careful with: this transformer has a primary impedance of, for example, 10k.
Transformers as such do not have an "impedance", because it depends on what you put in front of and behind the transformer.
The most important thing to know in this respect is the winding ratio of the transformer...

Hi, yes that helped.

So, once again, I have to remember that transformers are step up/down devices. My 10k primary is only 10k compared to its 4, 8 and 16 ohm secondaries. So if I connect my 55 ohm headphones to the 16 ohm secondary tap, and seeing that 55/16 = 3.44, then that 10k primary Z will now appear to be 34,400 ohms.

I saw that and thought that a 6SN7 (rp = 10k ohms) should work fine with that. But then it was pointed out to me that the primary inductance of the OPT was likely to be no more than 100 Hy, and that could present problems when used with a tube with rp of about 10k. So, now it's back to 5687 (rp = 2.5k or so). Too bad, because I have a few 6SN7's I was hoping to use. But I have lots of 5687's around, might as well use them. I could even parallel them for PPP use, get rp down well under 1500 ohms. Or parallel the 6SN7's?

I also found this thread...

http://www.diyaudio.com/forums/headphone-systems/18655-what-output-z-headphone-amp.html

...Wherein it was stated that if you design an amp that can deliver about 4 watts into an 8 ohm load, that should deliver enough power to even a 300 ohm pair of headphones.

The reasons why are a little difficult for me to grasp, but I'll read it again and try to understand.

-

And another thought... What about damping factor? How does the pri/sec ratio affect the output resistance of the final amp? If the primary Z is way higher than rp (let's say 5*rp or even 10*rp), does that help lower the output resistance of the final amp?

-
 
Last edited:
And another thought... What about damping factor? How does the pri/sec ratio affect the output resistance of the final amp? If the primary Z is way higher than rp (let's say 5*rp or even 10*rp), does that help lower the output resistance of the final amp?

Yes. Just divide the source impedance (for example the 700 ohm Rp of a 300B tube) by the winding ratio squared.
So in our example of the 25:1 transformer the output impedance is 700 / 625 = a little over 1 ohm (damping factor of 8 at 8 ohm); in this case we must add the DCR of the secondary, which should be some 0.2 ohms for a quality transformer.
 
Hey, thanks for the responses. I think the general design is shaping up.

Basically make a 2 to 4 watt per channel stereo push-pull tube amp using standard OPTs of about 7k to 10k p-to-p primary Z.

The output tubes could be paralleled 5687's, maybe triode-strapped 6K6, 6V6, EL84, whatever's handy. 45's?

Is there any kind of resistor network that should go across the OPT secondary windings to protect against overloading the headphone drivers?

--
 
Moderator
Joined 2002
Paid Member
Is there any kind of resistor network that should go across the OPT secondary windings to protect against overloading the headphone drivers?

Subscribed to get the answer to this question.

Background - I have this amplifier: http://www.diyaudio.com/forums/twin...i-tube-amplifer-us-120-only-silver-color.html

I'm now planning on a pair of cans (either the AD-700 or the M-50 from Audio-Technica). The amp is kind of unused as its very low power output left it suitable for only a very limited set of situations. Sounds very good, and I want to explore its suitability as a headphone amp. Should be well within its range of capabilities.
 
Moderator
Joined 2002
Paid Member
^^Got the answer myself. Stupid of me to not have checked, 100 ohm across the secondary to protect the amp from open. And another 40-50 ohm in series with the headphone, to protect against high SPL. The ground is not floating, so can be used on the sleeve connector (common).
 
Cool, thanks.

Wouldn't 50 ohms in series with the headphone be added to the headphone impedance, so seen as an additional load by the output tubes and OPT?

100 ohms across a 32 ohm load would change that to a 24 ohm load, would it not?

I'm probably missing something fundamental, but I'm curious.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.