Originally Posted by hitman47
OK, thank you for that informative answer, I was guessing it was something among those lines. I guess the real question is if the lower voltage impairs the sound quality in any way? Because otherwise, I'll just leave it like this, but if I can still improve the quality of the amp I'd be willing to order a 220V 24V wallwart. And about the input, if I wire it up to the left terminal, there's an effect if I flip the switch. I'll have to try it with two inputs as soon as I get that grubDAC done (accidentally broke the PCB with that crap soldering iron and solder; lifted a pad, I'm gonna have to order a new PCB first). Also, touching the pot only produces a noise when there's no input connected, otherwise everything's dead quiet now that I have the slightly lower voltage set.
It's quite possible you are connecting to the wrong set of inputs. I forget which is which, but if the music switches off when you throw the switch, then you've got it correctly connected. It won't make much difference if you connect two sources, the sources will switch when you throw the switch and it won't matter.
As far as lifting a trace, that shouldn't require you to order a new PCB. Simply solder a wire to each end of the trace where it's still attached. I've got boards with little resistor leads soldered instead of traces all over, because I did so much testing on them. IOW, use a jumper to connect the ends of the traces that are still attached to the PCB. It's really not too difficult.
As for losing performance with lower voltage, here's what happens: the supply voltage (set at V+) dictates the voltage swing that the tubes are able to use. This dictates the dynamic range of the amp. Now, the voltage swing is dependent on the impedance of the headphones you use. High impedance headphones like Sennheiser HD580/600/650 will want to use that entire voltage swing (plus or minus 13.5V at 27VDC). Lower impedance phones will not. Lower gain tubes will not, either. So, say for instance you are using Grados at 32ohms and 12FK6 tubes (low gain) and you set the voltage at 22VDC (being conservative to make sure you have linear regulation). You would set the tube bias at 11V (not 13.5) and that would give you plus or minus 11V to swing. The 12FK6 tubes would probably never swing that much with their gain and the Grados would never know the difference, anyway.
Unfortunately, the plates on the tubes perform better at higher voltages. In fact, we "unofficially" recommend setting the bias at 15V on a MOSFET-MAX with 12FM6's, simply because that tube sounds better at that higher bias. If you have 27V, that still gives you 12V on the plus side to swing, equivalent to many amps with a 24VDC power supply. However, if you tried that at your 22VDC, you'd only have a top swing of 7V, which is worse than a portable amp with 2x9V batteries (18V). IOW, the dynamic range of the amp would be less than a portable amp with two 9V batteries.
Now again, with low impedance phones you'd never notice this, IMHO. With higher impedance phones, you would. Bottom line, to get the best performance possible out of the amp, it's better to have the higher voltage than not. Only you can decide whether it's worth the extra investment.
EDIT: I forgot to mention that MOSFETs compound the issue of low power supply voltage. Unlike BJT's, MOSFETs actually have a turn-on voltage threshold. This is the "Gate Threshold Voltage" I believe, and is a full 2.0V for the Z24/9Z34 MOSFETs. So, you'd have to subtract that amount from the voltage swing available to the buffer once the signal left the tubes (clipping might occur, iow).
Edited by tomb - 9/30/11 at 7:12am