What output Z for headphone amp?
 

I am trying to design the output transformer for a push-pull class A headphone amplifier but I am having a tough time deciding what to do about the secondary configuration. For maximum power transfer the output and load impedances should be matched, but this gets a bit complicated when the load impedance ranges from 32 to 300 ohms. I would really like to limit the maximum number of secondary windings to two, so I figure that if each were 32R, then I could use a single 4pdt switch to select between parallel (32R) and series (128R) for both channels. Having more than two secondaries would require an expensive multipole rotary switch to implement, and I rather not go there.

But what effects will result from driving 300R cans, or even 600R studio cans, from a 128R impedance. I mean, besides the obvious effect of lower power output and higher damping factor (which I don't think is really important vis-a-vis headphones). What if I just went with a single 32R winding? Also, what about selecting the impedance via taps, instead of trying to rearrange parallel/series connections of the secondaries? This would certainly be much easier to accomplish, but what are the drawbacks?

Me with tube audio: :drool:

Hi,

Never had any trouble driving cans with the 8 Ohm tap of a regular OPT.

Since you plan on a PP headphone amp you'll quite likely have a lot of power to spare anyway.

Cheers,;)

100mW into 32R is the design spec. Were it changed to 100mW into 8R, the resulting power capability into 300R cans would be a less than impressive 2.67mW.

:Ouch:

Hi,

The important factor is to know the average effeciency of the headphones you want to hook up to it.

Didn't do any maths but your powerconversion seems O.K for transistors not for a tubed OPT coupled device...I could be wrong though.

Cheers,;)

Can you design for about 7 volts RMS behind a 30Ω source impedance? (Physical resistor, device plate impedance, or power supply regulation.) That will drive about any phones made to rated max power.

Ummm, this is the TUBES section, right? If you are coming out of triode tubes, match the plate resistance to 30 ohms, check that the no-load output voltage is 7VRMS or more. With most tubes, that is just not a problem.

You need 7VRMS or more for some hi-Z phones.

You need 100mA RMS or more for some lo-Z phones. (This is why transformerless headphone amps are rarely impressive in lo-Z cans.)

You never need both 7V and 100mA at the same time. A resistive model is (I was shocked) an excellent model to the galaxy of headphone impedances and power ratings. (May have something to do with the IEC mob, but also historical trends.)

That's assuming you actually want to supply the full rated 50mW-250mW that phones can stand without smoke. Many-many people do seem to be happy with just a few milliWatts. One tuner-tube in totem-pole push-pull with no transformer seems to make many people happy, even with lo-Z cans if they are not too insensitive or if listening levels are not high.

I did an extensive study once and posted (2-page thread) it on HeadWize. And the same idea in another form.

Hi,

Discalimer: I've never "designed" a transformer in my life! (Specified a few though;) )

However, I would not be looking at maximum power transfer as an important factor.
Whether you gain of lose a few mW is unimportant. The most important thing IMO is that the variation of impedance is not reflected back into the driving circuit, where it could make operation unpredictable.
Far better to go for the low Z source and Hi Z load all the way through.

Cheers,

Hi,

That's how it should be done in my little book too...which is why I feel a standard 8 Ohm or 16 Ohm tap should be able to drive 90% of all headphones successfully.

A couple of milwatts can get really loud with some cans already...

Cheers,;)

Yes, I know. The Grado SR-60's have a sensitivity of 98dB SPL for 1mW input and a 32 ohm impedance; the Sennheiser HD600 have a sensitivity of 97dB SPL at 1mW with a 300R impedance (rising to 600R at 100Hz). These two headphones require similar amounts of power to sound equally loud, but very different voltage/current combinations to get there.


My understanding of the tube OPT coupled stage is thus: the maximum amount of relatively undistorted voltage swing is produced when the anodes are loaded by the proper impedance. If 50Vrms is available at the primary, for 100mW total output a current of 2mA will flow (Z=25k). This turns into ~1.79Vrms and 100mA when transformed to a 32R impedance. The maximum undistorted signal at the secondary, then, is limited to 1.79Vrms. As far as I know, connecting 300R cans to the 32R winding will not automatically enable the output tubes to produce a higher voltage swing at a lower current.

If my understanding is correct, then having only an 8R winding available would be a disaster, n'est-ce pas?

Jeffreyj,

Not how I see it, at least for triodes.
The maximum undistorted voltage swing will occur when the load is near infinite.
In practice, arrange for the impedance to be as high as possible to give the lowest distortion. Of course you'll need some power, so a compromise will be needed.

Cheers,

Hi,

Pas vraiment...so, no it should be fine.

Look at it this way, the PP triodes should see a nice high load, as D'Haen points out, but the lower secundary impedance should not bother it or the headphones at all.

It should and will drive any impedance higher than its own without any problems...that's the beauty about tubes; you're most often fine when the load is equal or higher than the secundary impedance.

With semis, that's a totally different story, akin to the opposite.

Give it a try, find a half decent tubeamp and hook up a headphone to its' LS terminals using some loading resistors; 120R 5W IIRC and have a listen...

Should be fine, IME.

Cheers,;)