I'm currently trying to bring up my first system using (mostly) period-correct parts.

I had a problem where the system would work correctly if PHI2 was 1 Hz or slower (single-stepping) however when PHI2 was any faster (even 10 Hz) I was seeing $ff on the data bus when there should have been other values whenever NVRAM was being read. I have some very weak pull-ups on the data bus so I suspect that it wasn't being driven when the NVRAM was selected. After some head scratching I noticed this:



The socket pin seventh down on the right hand side was bent when it was inserted in to the PCB for soldering.

That pin is /OE for the NVRAM and was supposed to wired to GND. The NVRAM was being programmed at 115,200 baud without issue but could only be read very sedately. After I patched that up, it worked just fine at 1 MHz.

I could have saved some time in this case had I e-tested my boards after assembly. Is there any way a hobbyist can take the drudgery out of an e-test? Some sort of flying probe on a 3-axis machine perhaps? Or maybe a program that organizes then steps through the netlist, shows what should be tested and then automatically proceeds when it gets continuity?

I think it helps a lot to have a partner in crime who can read out the expected results as you step over the board with a continuity tester - a tester with an audible beep helps too. Then neither of you needs to move attention from, respectively, the circuit diagram and the board.

If you're only making a handful of boards, it shouldn't take long to check.

(In your case, possibly a visual inspection would also have helped - I just read a recommendation to use magnifying eyeglasses, although I haven't done that myself as I have a big magnifier on a bendy stalk.)

Thanks for sharing your results, as others can learn from this. In that sense we can say you "took a hit for the team."

Certainly an e-test can help find something like this after it has occurred, but ideally we want to avoid problems. An ounce of prevention, and all that...

Congrats on getting the project working!

-- Jeff

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

It seems to me that some force would have been necessary to get that IC into place while folding over that pin. If the pins are straight the IC should slip right in the holes during assembly, using very little force. As alluded by Jeff, better to be meticulous during assembly and not have to endure the indignity of a lengthy pin-by-pin test looking for bad connections.

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

In many cases you can see under the IC, between the IC and the socket, from the edge of the board, to inspect for that kind of thing. It's also good to look at every pin from the top and make sure you can see where it goes into its place in the socket. I'm always super careful about that kind of thing, because I hate troubleshooting and fixing hardware! (And as a result, I very seldom have to.)

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

I wonder that you haven't noticed that missing pin during soldering. Or was it done by wave soldering?
When production rejects went back to engineering (and before SMT becomes regular) I used to took an edge look to the solder side across the pins, looking for irregularities. Next was pulling out all plugged ICs to look for bent pins AND (in case of "precision" sockets) to look if the socket were fully equipped with pins and these tiny golden contact inserts!