Channel Imbalance in DAC -Seeking help / advice

Have you measured the output of the DAC (say, via a DMM while outputting a 60Hz sine wave) to make sure that the channel imbalance is really in the DAC? There are lots of other possible sources of channel imbalance.

Quote:

Originally Posted by mizunoblackox 

Tested all the resistors and noted 2 resistors' readings were out compared to the other channel (i.e. 1.2Kohm vs 2.2 kohm). Based on the color codes, should be 2.2kohm. THought I spotted the problem , repl the "1.2kohms" with new 2.2 kohms but the problem persist. The new resistors still reads 1.2koms and the desoldered ones seems ok reading at 2.2. i.e. the resistors seem to have almost halved the value when mounted on the board !

That's an interesting observation.

If you'd just said that the resistor measured wrong, I'd have thought that you were a novice who didn't appreciate that a resistor on the board may not read it's printed value because there are other components in shunt (parallel) with it.

The fact that the other channel measures correctly, however, means that there is a definite problem. The problem MUST be in something that is connected to the 2 resistor terminals, either at the pads or somewhere on the tracks that connect to them. That could be a chain of components, or a single component. There should be be an open circuit between the pads for the problem resistor (when it's removed), but instead there is a resistance of ~2k6 (by calculation).

The most likely cause is a solder string somewhere, probably introduced when you replaced the caps, but it could be a tiny piece of loose solder or cut component lead, or a faulty cap. I've even seen (detected) an invisible short due to a string of copper embedded in a PCB, but this is unlikely in your case, since I guess the DAC worked OK at some stage in the past.

Keep looking, you'll find something. Good luck.

w

Did you notice the imbalance after the caps were replaced (3 months ago), or was it present before that, leading you to replace the caps? If after cap replacement, then as wakibaki noted, the 1.2K reading could be due to a solder problem around the caps.

A schematic would help :) Barring that, you could follow the traces from the 2.2K resistors to determine what all they connect to. It has to be something effectively in parallel, which should not be there in this case.

Which resistors are you referring to? The hilighted ones?

I would take voltage measurements (to ground) on both sides of C105 and C205 (I think, not real legible) and compare between channels. You might also remove the offending resistor in both channels, and then ohm each resistor pad to ground and compare (power off for that one).

Here is a slightly more legible schematic:

Another view of the front portion of this:

This thread has some additional info which might help:

http://www.diyhifi.org/forums/viewtopic.php?p=28103

OK, unfortunately,

1. I can't read the schematic

2. I can see enough to tell that there are several paths in parallel with the offending resistor.

You've got some options.

First check the caps that are obviously potentially a problem, there's a 220pF close at hand in series with a resistor whose value I can't read, the both in parallel with the problem resistor. All caps should read open circuit (megohms unless electrolytic, in which case certainly hundreds of k). If you can find one that isn't, that's a good one to replace.

You can check voltages to ground comparing channels with the DAC idling. If you can find a clearer version of the diagram, write the quiescent voltages on it and post it. If you can find voltages at odds with the good channel then the components on whose pins you find them become suspect.

You can pull the active components and check them one-by-one. If you have a meter with a transistor tester (hfe), so much the better. The fact that you're getting something out of the channel tends to indicate that none of the transistors is dead though. Transistors read as back-to-back diodes, so you should see 0.6 ~ 1.0V with a diode tester from the base to both the collector and emitter (in one direction, open circuit in the other). You can identify the base, collector and emitter by googling the transistor part number and looking at the datasheet, or just test them at random and figure it out.

You still haven't told us when the fault appeared. I still haven't given up on it being something associated with the cap change.

w

Quote:

Originally Posted by wakibaki 

 

1. I can't read the schematic

 

Right click the big one, open in new tab. 

Check for shorts around C107. specifically that the non-grounded side is shorted to ground. 

I'd also check for shorts around "Lout+" but I'm not sure if there is a safe, or dangerous way to do it. 

Quote:

Originally Posted by wakibaki 

1. I can't read the schematic

Or, right click (or control click on a Mac), and save the image to your hard drive. Then open it in whatever preview or graphics package you have, and blow it up to your hearts content.