C652 is 0.1 uf (microfarads)
C623 is 0.22 uf (microfarads)
C653 is 10uf, 16V
C621 is.....I have no notes on that one, nor C622. (But see end of post)
First question: is the capacitor in question a series path from point A to point B, or a bypass path from point A to ground? For the above:
C652 is series, point A to point B
C623 is bypass, point A to ground
C653 is series, point A to point B
C621 is bypass, point A to ground
C622 is series, point A to point B
Bypass caps usually mean the cap can be temporarily removed; it will allow the circuit to work, but not necessarily properly or the most efficiently. However, if that cap is preventing the proper function completely, removing it and finding that you now get a partial circuit function is certainly a key indicator!
Series cap usually mean the cap absolutely must work for the circuit to work. You want to check these caps after you have eliminated any bypass caps as the problem. Reasons for that is, the series cap needs a bit more in-depth analysis. So, do the easier bypass caps first. In the case of the ASPCM circuit, both C623 and C621 can be removed for this analysis. Keep track of which is which, and remember to reinstall (or replace) them later!!!!
So, the next step is the series caps, right? Yes,.....but there is a little bonus for us in this circuit; C622 is used as feedback, which means the circuit can probably operate without it! So go ahead and remove it, too. Again, remember to reinstall/replace.
Since you have (presumably) removed the bypass caps, do a little testing...
Start with the simplest analysis first: A capacitor is going to be shorted or open. You want to check for both of these errors; if you can eliminate both of these possibilities, then assume it's good (NOT a guarantee, but a reasonable bet for first level troubleshooting).
[What about testing for a "leaky" cap? That's why I said "simplest analysis first". Determining if a cap is leaky can require a bit more work (unless of course you see electrolyte all over the board!), and the short/open checks can help you toward this.]
Now what?......
You want to check if the cap is shorted or open. You need to apply both tests, as you can get an ambiguous result and will need both data points to eliminate the ambiguity. Make a chart for each cap, check:
1. With power off, resistance from first pin to ground
2. With power off, resistance from second pin to ground
3. With power on and no ADPCM audio active, DC voltage on first pin relative to ground
4. With power on and no ADPCM audio active, DC voltage on second pin relative to ground
5. With power on and ADPCM audio active, AC voltage on first pin relative to ground
6. With power on and ADPCM audio active, AC voltage on second pin relative to ground
Note that, for 5 & 6, you may be looking for a few hundredths of a volt, so you need to use the most sensitive scale on the meter, and make sure you are NOT getting a false reading caused by any nearby electric/electronic interference.
Now you need to use the tool between your ears:
If the cap is shorted, then the DC voltage will be the same on both it's pins. Of course, be careful that the cap doesn't, by coincidence, connect two points that happen to actually have the same DC voltage. Usually, however, the whole point of the cap is specifically intended to keep the two voltages separate, so, if the DC voltages are different, assume the cap is not shorted.
If the cap is open, you will certainly have different DC voltages on it's pins (again, unless the circuits in use happen to have the same voltage). But a good cap will have this too, so check the AC voltages. You should have the same on both sides; that's the caps job, to get the signal from point A to point B. We will assume that point B is, itself, not the problem. If it looks like the cap IS open, it is safe to temporarily connect a second, new cap into the circuit...without having to remove the original, suspect cap. Just be sure to observe polarity, if any.
Shortcut:
In the case of the ADPCM circuit, U503 is used as a two-stage amplifier filter. This means that pin 1 has a much more powerful ADCPM signal than U502 or any parts in-between. So, you could try just measuring the AC voltage on pin 1, and compare the value of ADCPM audio off with the value of ADCPM on. But I guess you have done that, otherwise you would not be concentrating on the actual capacitors. However, as we have shown, C652 (and C653, which you did not mention) are the only two critical caps in this circuit. And if you do the simple short test (same DC voltage on both pins), and the simple open test (connect a second known-good capacitor across them), testing this circuit is easy.
As for C621 and C622:
I mentioned that C622 could PROBABLY be removed and the circuit could still work. This is because both of these caps are used as "filter elements" in an "active filter". This is a fancy way of saying they help "shape" the sound, kind of like an equalizer is used for audio systems. Of course, this equalizer only has one setting. But, without those caps, this equalizer circuit MAY have characteristics that, while still functioning properly, act like it isn't. So removing C622 if it is good (but you don't know that) has a small chance of actually making things appear worse.
As for the values, that depends upon what effect the equalizer is supposed to have; that is, what the audio is actually supposed to sound like. For test purposes, I would remove them both until you have (probably poorly) working audio, then reinstall them one at a time as a test.
Charlie
