AMB Laboratories DIY Audio

[Mullet]

It gets more bizarre... after checking all parts, resistors, caps, etc. around U7 and reflowing everything including U7 now I'm getting:

Sleeve to Tip: .360Vac
Sleeve to Ring: .723Vac

So now the right channel is louder or right centered. I've confirmed this with a few pairs of headphones. Looks like I *might* take the shotgun approach and redo the whole Gamma-2 section of the DAC. Another option is the pull all the caps around U7 and order them (I melted 'em) and another U7 first. I'd hate to redo U7 again to have to buy everything in the long run anyways. What a pisser!

[amb]

Well, obviously what you were doing made a change, which seems to imply it was indeed a soldering issue. Try reflowing the joints again. Building a whole new board doesn't make much sense for what seems like a minor problem.

[Mullet]

Yup. Obviously. So I've reflowed U7 like 10 times and now there is no difference in readings. I'm wondering if some residual chip-quik would have anything to do with this. I tried cleaning it off and by now I'd think with all the reflowing it'd be burnt off. At least the Gamma-1 works I could try pulling the parts again and re-soldering them. What's weird is that the DC ratings on the U7 pins are still the same from when I initially fixed the Gamma-2, but the AC RMS has changed. I feel like this board is cursed.

[Mullet]

So I decided to not take the shotgun approach and redo the whole Gamma-2 section. Good thing! I tried reflowing again and this time around pulled my C20 and C26 and now I'm getting what seems to be more reasonable readings...

New Reading with new MUSE caps…
S to T = .896Vac
S to R = .911Vac

My other Gamma-2 (ELNA caps) is getting these readings for reference.

S to T = .721Vac
S to R = .721Vac

I would think that the solder joints at C20 and C26 were less than satisfactory. Otherwise, I think the slight imbalance is something else down the line. Perhaps if I reflow the 8741 DAC chip things might even out.

My only reservation is that the other Gamma-2 has distinctly different values for RMS readings. Any reason why that could be? I'm using MUSE caps in the recently fixed build and ELNAs in the other build.

In the end, I'm satisfied that there is no noticeable imbalance at this point. Getting the readings closer would be more along the lines of obsessing.

[Mullet]

There's something more going on here...

Hrmmm... now it's back to its old self again. I haven't done anything except unplug the Gamma-2 and do the same tests with my newly built GrubDAC then plug the Gamma-2 back in and now it's .453Vac on the left channel again. I think it's time to invest in a new Gamma-2. Bleh ~$70 fix. D'oh!

[amb]

Reflow the WM8741 pins.

By the way, at 0dBFS the γ2 should output close to 1.4Vrms. Make sure your Anti-clipping switch is OFF and any software volume control is set to maximum for this measurement.

[Mullet]

Just realized the "good" Gamma-2 had Anti-Clipping on. New readings are ~.901Vac for both channels. The "bad" Gamma-2 never had the Anti-Clipping on with those readings. Will try and reflow U5 and I'll post my results. If that doesn't work out I'll most likely be putting in an order to you. With the TI samples program, amb.org store and Mouser I can get the total cost to around $60.

[Mullet]

Well it appears after finally reflowing U5 for the first time since doing this build and U7 again... I'm back in business.

Measurements are .906-908 from Sleeve to Tip and .910-.911 from Sleeve to Ring. I think this is as close as I'm going to get it. My theory is the channel imbalance was always there even before my U7 catastrophe.

Hope this helps someone in the future.

[amb]

Congratulations.

As I've said many times before, and I'll say it again here... 90%+ of γ1/γ2 problems are related to soldering. Most of the time it just takes reflowing of the joints to fix it.