AMB Laboratories DIY Audio

[Mullet]

This is a cross posting from over at Head-Fi. Figured I'd get more traction here. I've posted twice about my issue so I'll post both here...

"This evening I was playing two amps from the Gamma-2 portion of the DAC (one via RCA jacks and the other via the 1/8" jack) and I was fiddling with the amp (turning up the volume) that was connected to the 1/8" jack and I saw a few sparks because a metal 1/8" connector was touching the pot, which is also metal. This tripped the E12 on my CTH -- the other amp. I proceeded to unplug the first amp (A47) and then after firing up the CTH again noticed a buzz coming out of the left channel and proper audio coming out of the right channel. I tested the Gamma-1 portion and it plays music normally. So I know its the Gamma-2 with the issue ... Any ideas on what to check what died or is dying? U7, U5? What voltage readings and what pins should I check?"

"Upon using a DMM I'm getting these readings... I've decided to do the typical tests when done with a Gamma-2. Everything checks out with these voltages. Then I decided to take a look at the outputs of U7. On a known working Y2 I got these readings going into C20 (R) and C26 (L) -- 2.27v and 2.28v. With the bad Y2 I got 2.202v on C20 and 4.33v on C26. I'm assuming now that there is something wrong with U7 and I'll try and use some chipquik to get the chip out. I tested the points coming out of the U5 and I got similar readings of 2.2 to 2.3 volts for L and R of both sides of the DAC. So I'm thinking the DAC isn't the problem and it's most likely U7 that got severely damaged. Keep in mind -- no magic smoke and the right channel still works. Does this make sense?"

So to sum it up I partially fried something on the Gamma 2 from a short from one of two amps hooked up. I ordered a new U5 and U7 and plan on replacing the U7 first to see if it solves my issue. Once fixed I don't plan on having both hooked up and turned on at the same time anymore.

[amb]

Check the DC voltage at each pin of U7. These are not accessible when the boards are mated together, but look at the schematic to see what other parts these pins are connected to, along with the PCB layout image from the website, you could then check the voltages on the solder side of the board.

All pins should measure about 2.3V, except pin 4 (0V) and pin 8 (4.5-4.75V, depending on what type of voltage regulator was used for the analog rail.

[Mullet]

I'll post results as soon as I have them in the mañana. Mister X mentioned testing U5's VMIDR and VMIDL pins as well in the Head-fi Y2 thread. I'll do that as well. I'm assuming I can use the vias next to L+ and R+ to do this. Hopefully, this is just U7 acting up as it is easier to replace than U5.

[MisterX]

No it's not.
(those plastic caps don't like my hot air station and they're a pain to replace)

[Mullet]

I stand corrected. Yes, they both look difficult. Perhaps, I should have bought a few of those Wima caps as well. We'll see what materializes after replacing U7. If I melt a cap or two I guess its means an unwanted trip to Mouser.com.

[linux-works]

if you do an order, maybe order some chipquick (sp?) as well.

if hot air can ruin nearby things, you want to try a diff approach.

[Mullet]

Have some already when I originally started building Gamma 2s and 1s... I'm more concerned with not taking out any solder pads. I understand that chipquik has a lower melting point than 60/40 solder so I can dial back the temp to 150 degrees F. I worry about those box caps though. Too late the U5 and U7 have already been rolled into an O2 amp order.

[amb]

You can remove a few of the box caps to get better access to the opamp.

[Mullet]

Looks like it's definitely U7. I'm getting voltage readings of 2.3 and 2.29 out of U5's VMIDR and VMIDL pins. Then at C24 I'm getting either 3.58v or 4.5v depending on which lead I test where it should be close to 2.3v. At C26 I'm also getting 4.5v. Since I didn't buy extra box caps I'll have to carefully extract them using a big blob of solder or with ChipQuik. I'll post my results.

[Mullet]

So here's the skinny. Pulled C19, C25, C16, and C22. Found that a little blob of chip quik on each solder point and a temp around 650˚F worked best while going back and forth and pulling out the box caps with a set of flat nose pliers. When all was cleared I proceeded to apply some flux paste and two nice blobs of chip quik to the U7. Some directions I found online mentioned I could use a low temp when removing the chip so as not to remove the solder pads, but again I found 650˚F to work best. 250˚F didn't do too much. Once I removed the chip, I cleaned the solder pads with a solder sucker and proceeded to add more flux from a flux pen and re-soldered in a new U7. Finally, I put all removed box caps back in their respective positions. Upon firing up the DAC I tested C26 and C20 and now I'm gettings readings of ~2.35v and ~2.25v. I tested L+ and R+ and I'm getting similar readings. I figured that was better than my initial readings the other day, so I hooked everything up to my computer.......... drum roll...... Audio out of both channels! My only worry is that there is a slight imbalance in my voltages. Tomorrow, I'm going to reflow all joints. Anything else to test for? Could I have a bad cap or semi fried cap? I'd think I'd notice physical damage. I'll also measure the capacitance of all caps and see if that gives a clue.