Zen v4 headphone amp

[Rob M]

I'm trying to configure a Zen v4 as a headphone amp , using Nelson's PCBs but with Tortello-inspired parts substitutions. It's mostly pretty straightforward: IRF610's for Q1 and Q2, lower regulated supply voltage to 24V, lower bias current to around 250ma by changing R0 and R1, and adjust R14, R15, and R16 appropriately.

I'm not sure what to do with the regulator, though. For Q5, should I stick with the original part? Can I substitute IRF610 there as well? Fewer different parts makes life a little bit easier, but should I be worried about it's lower transconductance? And, does changing the voltage and possibly Q5 mean that I'll have to change how much feedback should be coming through R19 and C11? I don't think I understand what that modulation is doing. Is the idea just to minimize the fluctuations in the voltage, or are we trying to introduce fluctuations, like in the active current source?

Also, is there anything else I should be careful about in scaling down the Zen v4 to headphone dimensions?

[Verbal]

Hi,
Great to see another headphone Zen in the making, you won't regret it...
Since this is a class A amp, the IRF610's lower transconductance and therefore higher power supply output impedance shouldn't really matter.
R19/C11 modulate the supply voltage according to the output, which gives you a higher peak supply voltage and therefore more headroom. To drive cans you don't really need that, so maybe you better leave them out altogether.

[Rob M]

Originally posted by Verbal
Hi,
Great to see another headphone Zen in the making, you won't regret it...

Thanks, glad to hear it. I wasn't sure it would work very well. In my simulations, the distortion goes up in a hurry as the bias current goes down: THD at 1KHz with 2A bias is 0.05%, but with 300mA it's 0.5% and with 150mA it's almost 2%. I don't have a lot of confidence in simulated distortion figures, but it's still not a good sign. I guess I'll just have to build it and find out for myself! .

Since this is a class A amp, the IRF610's lower transconductance and therefore higher power supply output impedance shouldn't really matter.
R19/C11 modulate the supply voltage according to the output, which gives you a higher peak supply voltage and therefore more headroom. To drive cans you don't really need that, so maybe you better leave them out altogether.

Okay, thanks. Squeezing out a little more power is definitely not an issue -- the tinnitus I've got already is plenty.

[Verbal]

"THD at 1KHz with 2A bias is 0.05%, but with 300mA it's 0.5% and with 150mA it's almost 2%. I don't have a lot of confidence in simulated distortion figures, but it's still not a good sign"

Under which conditions? In any case, Marcello's measurements (see at Headwize) look far better, even at earsplitting output levels.
So don't worry, go ahead, and let us know how it turns out...

[Rob M]

Originally posted by Verbal
Under which conditions? In any case, Marcello's measurements (see at Headwize) look far better, even at earsplitting output levels.

Um, yeah, now that you mention it, that's at a little over 1Vrms output into 32 ohms. At lower levels, the distortion goes down quite a bit, and in fact my simulated numbers aren't all that far from what Marcello reports.

So don't worry, go ahead, and let us know how it turns out...

Will do! I've got the PCBs now, and the rest of the parts are on their way. Thanks for the advice -- no matter how it turns out, it'll be better than the last headphone amp I tried.

[Rob M]

So far so good! It's all together and working now, and it sounds great. The next step is to build a Jan Meier-style crossfeed for it and puit the whole thing into a case.

I've got a transformer question now, though. My first attempt at the power supply used a single el cheapo 50VA transformer for both channels. Technically, that should have been enough, since the amp only draws around 500ma @ 30VDC per channel, but the results were not very good. No big surprise there, I guess.

Here's the puzzle: giving each channel its own transformer improved things, as I expected, but it also had an unexpected effect: it cured a slight hum problem that I hadn't been able to solve through grounding.

Can anyone explain why an inadequate power supply transformer would produce a hum on the output? Or is it more likely that the hum was due to a wiring problem which got solved as a by-product of rewiring the power supply for a second transformer?

[joensd]

So far so good! It's all together and working now, and it sounds great. The next step is to build a Jan Meier-style crossfeed for it and puit the whole thing into a case.

What exactly are you using for power supply, Rob?
Marcello mentions a turn on thump without a slow turn on PSU or a timer PSU (like in the Headwize article).


TIA
Jens

[Rob M]

Originally posted by joensd

What exactly are you using for power supply, Rob?

A scaled-down version of the one in the Zen v4 article, with 6A bridges and 4400uF per channel.

Marcello mentions a turn on thump without a slow turn on PSU or a timer PSU (like in the Headwize article).

Yeah, it takes quite a while (~10 seconds) for the DC offset to go away, so some kind of muting circuit would be a good idea. I'm using this:

http://www.epanorama.net/circuits/ampdelay.html

but I'm sure there are more elegant solutions.

[Verbal]

Originally posted by Rob M
I'm using this:

http://www.epanorama.net/circuits/ampdelay.html

but I'm sure there are more elegant solutions.

I like the way Marcello does it better - connect the relay so that it shorts the output to ground with no power, and then gets pulled open by the delay circuit, avoids some unnecessary contacts in the signal path. Depending on what type of relay you have, it might be just a matter of rewiring your circuit...

[Rob M]

Originally posted by Verbal

connect the relay so that it shorts the output to ground with no power, and then gets pulled open by the delay circuit, avoids some unnecessary contacts in the signal path.

I was wondering about that... isn't shorting the output going to be hard on the transistors?