I built up another ZG6502 as a testbed to check out a number of used 6502 and 6551 that I purchased a while ago. ZG6502 design is proven, but the new board won't work, even when populated with working parts from a working board. After a good bit of inspection and troubleshooting, I finally discovered two broken traces on the problematic board. These broken traces were under the solder mask, so they were broken during the manufacturing process. Here are side-by-side comparison of the assembled board vs bare pc board. The broken trace is the one pointed with the white arrow. The trace should pass through pins 3 & 4 of U3 as shown in the bare pc board, but you can see the trace was broken and pushed off at a 45 degree angle in the assembled board. There is another broken trace but it is under the socket and hard to photograph.

PCB defects are quite unusual with JLCPCB, but it can happen. It is something to keep in mind when troubleshooting.
Bill

Edit, Inspected other pc boards from the same batch and found no problem. Build up another board and it is working fine.

Indeed they are rare, but can still happen. I remember early in my electronics journey that I had a similar issue. It even looked like they connected, but it simply was not. As soon as I connected them myself it worked fine. Thankfully it was a disconnected color line, so it was pretty obvious what the issue was. Never had a problem like that again.

Thanks for the update Bill!

Chad

I thought JLCPCB did 100% flying probe test on each board.

Greg

I believe they do, but that I think has to happen before the solder resist and silk screen. So e.g. a probe catching a thin trace before moving might indeed have checked good, and then broken before further processing.

Neil

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I had the same thought that the trace breakage occurred after electrical test. The trace was pushed aside, so it was caused by some kind of mechanical tool.

Pc board used to have lots of problems, so we would fully inspect a board before assembly. We even had a special prism probe that showed the inside of plated through barrels.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I even bought a desoldering station for that purpose after accidentally ripping some traces by pulling out a DRAM (I was repairing lots of dead C64s at that time, occasionally salvaging some precious SIDs or socketing them.) Desoldering gun makes it effortless, and I never used any other tools ever since.

There's one downside though: when you have a desoldering station, everything looks like it could use some desoldering.


Totally. Cheap PCB manufacturing seems to be a highly competitive market nowadays, so I'm sure they will be happy to get that feedback, possibly even compensating you for damages.

_________________
/Andrew

deck65 - 6502 slab with screen and keyboard | ПК-88 - SBC based on KM1810VM88 (Ukrainian i8088 clone) | leo80 - simple Z80 SBC
nice65 - 6502 assembly linter | My parts, footprints & 3D models for KiCad/FreeCAD