Muff's Modules & More

[bananeurysm]

So I've gotten along pretty well for a long time identifying module problems by simply looking closely at the board and wiring, and spotting stupid mistakes. But now I need to learn how to actually troubleshoot.

I built 2 VCSs. 1 of them works perfectly, one of them does literally nothing and won't even pass signal while powered on. I've double checked all the wiring, done close solder point inspections, and double checked all the chip and polar cap orientations - everything obvious. In the past - something in there has always been my problem - but not this time.

I have a meter, I have a scope. I'm assuming it's time to actually learn how opamps work and test voltage at various points on these chips? Could someone point me in the right direction here? I really have no clue where to start.. I guess I could go pin by pin and compare to my working circuit.. but that's sort of cheating - not to mention very tedious sounding.

How do I check basic functionality of the chips (and transistors) in this circuit? http://www.cgs.synth.net/modules/cgs75_vcs.html

Or is there a better way to go about this?

[Revok]

If it's not working at all, I'd start with a continuity test using your multimeter. It's a little tedious but not too time consuming.

[bananeurysm]

I've done continuity tests here and there - and checked voltage at the pin inputs - all of that looks good as far as what I understand.. I'm thinking I should start testing voltage at the outputs of these opamps? Does that sound right?

Or are there some specific points you're recommending a continuity check? I'd really love to learn how to approach this - I always feel pretty haphazard with my continuity checks.

[keninverse]

If one is working and the other isn't, the easiest thing to do is to trigger each module and compare the voltages of each of the modules. Start first at the input and work your way to the out. I'm willing to bet you'll be able to find something along the way. There's a lot going on with the VCS but I'd start first following from the trigger in. If everything looks okay there, start measuring from the regular input.

[appliancide]

What I do:

- check orientation of semiconductors (also make sure they are seated in sockets if you are using them) and polarized capacitors
- check power on all semiconductors
- stare at pcb through magnifying loupe, checking for shorts, opens and bad solder joints
- continuity check any connections that are suspect
- send triangle wave to input and use scope to trace the signal through the circuit.

Having a working circuit to reference should be a big help as well.

[marvkaye]

If you haven't already, please, by all means, read Ray WIlson's (MFOS) Troubleshooting 101. He lays out lots of hints and tips to help you get to the bottom of a problem. AAMOF, if you do what he recommends there prior to firing up your module for the first time it is entirely likely you'll get the module working faster than if you didn't follow his recommended procedure. Works for me, just thought I'd pass it along......

<marv>

[Revok]

keninverse described a good method of working your way through the circuit. I find it helpful to visually separate the different areas of the circuit kind of like Ken did here:
http://www.cgs.synth.net/modules/cgs09_vcd.html

[bananeurysm]

Thanks guys - lots of great advice! A lot of it is part of my common practice already - but a few key bits are not!

How do I check power on the semiconductors - probing the + and _ voltage pins is really just ensuring continuity from the power inputs of the pcb - to that point on the pcb (which I've done). Is there a way to test the chips themselves without blindly swapping them?

Working my way from input to output makes a lot of sense.

If I start with the trigger input - I guess I should see the trigger on my scope at pin 13 of the 3900 if I'm understanding this correctly. That's about the point at which I get lost. Should the trigger also be present at pin 9 of the 3900?

Similarly - should the input signal be present both at pin 6 and pin 7 of the tl072?

Of course it only gets more complex from there - but that might be away to at least test that those 2 chips are passing signal?

I will take keninverse's advice and compare the two circuits - but it would be nice to be able to read schematics a little better - especially for when I don't have the luxury of one working module for comparison!

[GrantB]

If you haven't, you may want to look at Tim Stinchcombes superb article explaining how the VCS works: http://www.timstinchcombe.co.uk/index.php?pge=vcs

Serge is a twisted genius who used the LM3900 chip a lot, most famously in these slopey circuits. It's a Norton amplifier which means it operates on current feedback rather than the voltage feedback used in most op amps. As someone who barely understands regular op amp theory, I couldn't comprehend the DUSG at all until I read Tims article about 5 times. The LM3900 app notes help too.

To quote Jimmys mom in Quadrophenia: "IT'S NOT NORMAL!"

[bananeurysm]

Thanks GrantB - I was actually JUST looking at that page and realizing it had a bunch of nice probe points for me to play with!

[Dr. Sketch-n-Etch]

I don't have too much advice to offer on troubleshooting that can't be found in the sources already cited. However, I will offer this: whenever I have a module that won't work, it almost (99%) always boils down to a dead chip. (And, chips almost always die because they have been plugged in backwards. )

Presuming you have extras (or just pull chips out of a working module) you could try replacing every chip one at a time. Before you do that, though, plug in the broken module and systematically do a "fingertip check" on every chip. In my experience, broken chips almost always run hot.

[daverj]

My methods are similar to what has been described above.

Without power:

- check that multi-lead parts (ICs, transistors, etc...) are oriented correctly. Check based on the documentation, not just looking at the sockets.

- Inspect the board for solder splashes/bridges. Inspect traces for cracks. Inspect solder joints for unsoldered pins, or pins where it is not clear that the solder bonds to both the pad and pin.

- check wiring that leaves the board to panel parts or power

With power:

- Use a volt meter to check the power pins on each IC. Not at the board pad, but right where the pin enters the body of the chip (to rule out pads, sockets, etc...)

- If you have a working one, power both units up, feed the same signals into both, set knobs/switches the same, then scope points on each tracing the path of the signals through the circuits. Compare everything, including places where you don't expect a signal, just to make sure the voltages match.

[bananeurysm]

Great advice - thanks everyone! I'm now confident I'll figure this out.