Strange Sound Issue. Looking for diagnostic/fix suggestions

[+cubanismo]

I have a strange issue on one of my Lynx II units. The sound sometimes works, sometimes doesn't. I just recapped it hoping that would address the issue, and I'd previously rebuilt the power supply when installing the BennVenn kit, but the issue persists. It existed before installing the BennVenn screen as well. The issue occurs both on the internal speaker and headphones. When the sound is working, it sometimes later cuts out with a sound that reminds me of a static-filled 80's laser sound effect... "PEEeeeewwww-*static*-*static* <silence>." Sort of a slippery sound to it if that makes sense.

 

I noticed while reassembling the unit today after the recap that if I press very hard on the top (Right-handed grip) A button, the sound comes to life when it's not working. I took it back apart and there doesn't appear to be anything even close to making physical contact with the PCB or speaker, so all I can imagine this is doing is flexing the PCB a bit, and I don't know why that would do anything unless there's a cracked solder joint somewhere under the shielding where I can't see it or something. Before I go disassembling that mess, has anyone seen/heard something like this before? Any advice on alternate ways to narrow down the issue?

[+karri]

I did have a similar issue on my Lynx 2. For some reason there was no connection between the Mikey and the audio input of the amps. I never figured out why/where the break was. Soldering in two new wires fixed the issue.

[+Stephen Moss]

5 hours ago, cubanismo said:

The issue occurs both on the internal speaker and headphones.

That would suggest the issue is with U12 or at an earlier stage, but you really need access to an oscilloscope to check you that part of the circuit as a multi-meter won't be sufficient other than to check the supply voltage of U12, without an oscilloscope I think you might just have to try replacing U7, U12 and VR2 to see if that resolves the problem.

 

Although your description of the sound returning does sound like a dry joint issue, however I am not certain you would get enough flex to affect anything under the shielding due to how the PCB is supported, where the pressure is being applied and that the shielding itself should reduce the amount of flexing. Could that have just been co-incidence or is it very repeatable?

[+cubanismo]

It's very repeatable. I can toggle sound on/off over and over by pressing harder and softer on A.

 

I do have an oscilloscope. What am I looking for?

[+cubanismo]

I took the damn thing apart again to mess with it some more. It's definitely something to do with flexing the PCB that makes it work, though it used to work without any physical fiddling (sometimes had to turn it off/on a few times or fiddle the volume to get sound) most of the time before today's recap disassembly/reassembly.  This rubber/plastic strut thing is resting over C60 in the current assembly. I keep forgetting about it, so it fell out of place when I disassembled it to do the recap, and hence I don't know if it was in exactly the same place before, but it's from somewhere over there, and seems to be providing a bit of support for the A/B buttons:

 

 

I can reliably restore sound by pushing down around that area and pulling up on the area of the PCB near C61/C58 as shown here:

 

 

If I release it *very* slowly, the sound will keep working. Definitely seems like some kind of nearly-broken trace or bad solder joint, but I can't see anything obviously questionable in the area even with a little magnification.

[+karri]

That is exactly the spot where I added the wires. But I could not find the error either.

[+Stephen Moss]

22 hours ago, cubanismo said:

I do have an oscilloscope. What am I looking for?

I see U12 is on that corner of the PCB, so it make more sense now that PCB flexing could cause an issue rather then it being somewhere under the shield.

You mentioned the sound also cut out when using headphones, presumable it was both channels cut out, if so then we would be looking for something that is common to both channels which would probably mean U7, U12 and VR2.

 

I am not sure where U7 is but it does no appear to be close enough to the area in question for the issue to be there unless the problem lies with a break in the power rail.

When there is no sound...

Check the power pins of U12, VCC will be easy but ground would be a little more tricky

I was going to suggest checking the inputs and outputs of U12 for anything resembling an audio signal. You could still try that, but I just did and the signal was the same whether audio was playing or not, which logically can't be right. Perhaps a USB oscilloscope is not good enough to capture it or taking GND from D11 was causing problem, consequently I am currently unable to suggest what you should be looking for.

 

 

[+cubanismo]

Turns out it was a weak solder joint on the bottom of L21 (Top in this picture):

 

I measured VCC<->GND on the LM386 (U12 and U5), and sure enough it was nearly nothing when sound wasn't working, ~4.65 when it was. Seemed to be floating or nearly so in the bad state. So I went here:

 

https://atarigamer.com/pages/atari-lynx-model-2-pcb-stencil-scans

 

And traced the VCC circuit back. Took me a minute, but I eventually realized L21 is where the voltage was getting into that trace, and sure enough, the solder was prettttttyyy thin on both  sides of that component. I pushed on the one side while sound wasn't working and sure enough it sprung back to life. I managed to dip my soldering iron in there and gob a bunch more solder on each side, and sound is steadily working now no matter how much I tweak the PCB. Close-up after picture:

 

 

Not pretty, but best I could do without removing the top cover entirely again. I hate wrestling with that BennVenn display ribbon cable in the little snap-on connector, so I wanted to avoid that.