FAQ for the headphone amplifier

Although these questions aren't exactly frequent, they have been asked at least once :-)

What power supply voltage are you using?

Personally, I'm using +/-15V regulated, but that's mostly because I have a pair of old 2kW Sennheiser headphones and I really do need that kind of headroom for transients. A friend of mine is using it with +/-10V and it can work on even lower voltages although I can't really recommend doing that, the output isn't exactly rail-to-rail. The power really should be regulated, though, as the power supply rejection isn't terribly impressive. That's not surprising, of course (it is a very uncomplicated design).

What is the approximate current draw (two channels)?

Well, my amplifier dissipates about 4W per channel (if I recall correctly), for an astonishing efficiency of about 1% ;-) Now, I know it's a bit high, but it is class A after all. The reason is that the bias current in the output stage must remain high to deal with low impedance headphones, and the supply voltage must be high to deal with high impedance ones. Speaking of low impedance headphones, I see now that R13 could be increased a bit. The standard (which I didn't know of at the time) is 100 Ohm. Naturally, you can lower the dissipation a bit by using lower supply voltages.

What is the proper setting of R17?

Ahh, the secret of R17. I really should have explained that somewhere, shouldn't I? Well, it's like this: Because the exact value of the current in the the input stage is fairly uncritical, I use it to make sure that the voltage at the gate of M5 (with zero input) will be such that the voltage on the output will be at zero. If this is adjusted incorrectly (fear not, it's real easy) it will still not show up on the output since it will be corrected by the DC servo, _but_, that will in turn unbalance the input stage which will then lead to higher distortion. I'm going from memory a bit here, but I think that the best method depends on how good a voltmeter you have. If you have a rather good one you can assemble the whole thing and then tweak R17 back and forth until the output of U1 is steady as near zero volts as you can get it. If you have a $8 voltmeter purchased at Rat Shack, it's probably easier to assemble everything except for R8 and twiddle R17 until the output of the whole amplifier is as near zero as possible. Then add R8, and off you go. U1 doesn't have to be an OP-07, if you have trouble finding a cheap supply of those. Just make sure it doesn't have terribly high input offset voltage, the speed and distortion really doesn't matter. Oh, and make sure you do all this at the intended supply voltage and after the amplifier has heated up to it's operating temperature. (It doesn't make a _very_ big difference, but since you care to build your own amplifier you probably care to adjust it as well as possible, right? :-)

M1, M2, M3 and M4 seem to be difficult to find components where I live (San Diego, CA, USA). Could you suggest replacements or possibly some specifications so that I could cross-reference them? Out here I can get NTE, Fairchild, and possibly some Harris, IR, or SGS Thompson.

Well, actually I think all the ones I used are from Harris. The electronics club I was in got a shitload of stuff when Harris switched their Swedish distributor and the old distributor wanted to get rid of their components... The ideal replacement for the IRFD1Z1 would be something like a PHN210 (two N-channel FETs in one package), but that's from Philips so I guess you're out of luck there. You could get something like a IRFD110 or IRFD120, I seem to recall that they're almost identical to the IRFD1Z1 and they were the ones I designed in but later changed to IRFD1Z1 since I couldn't find them again in the MOSFET drawer. They're from Harris originally. Fairchild at least has the datasheet now, so they should have taken over the production of it. As for the RFP14N05, if you have no trouble getting the IRF9520 you should have no trouble getting the IRF520, so just use that instead. It's even better, really. If you can't find these, let me know which distributor you intend to use, I'll check them out to see if I can find something that will both work well and have a reasonable price.