Hello everyone.

Today I brought my Acolyte SBC's into the office, preparing for Math Day tomorrow. I got them hooked up to monitors and running games, etc. I left them on for hours because that's the plan tomorrow anyways.

One of my older revisions worked just fine. The latest one would lock up after some time, randomly! I don't know exactly what was happening, I didn't have a scope, or anything else with me. It would just run for about 30 minutes to an hour, then lock up. I'm guessing it put itself into some type of infinite loop or something. I tried two different places in software as well, same effect.

Why?

After work, I brought my board home, put it where it normally is, and left it on. It's been running for almost 5 hours straight now, no locking up.

Why?

My first theory is that it could have been some type of interference. There was a LOT of cables on my desk at the office: 120V power cables, USB cables, VGA cables. They were nearly laying on top of the SBC itself. I had brushed them off to the side thinking that laying on top of it was wrong, but that didn't change anything, it was still locking up. And though the cables were not on top of the SBC, they were close, very close.

So my question: Could there be such interference with power and USB cables hanging around that it could mess with my exposed SBC? If so, are there ways to prevent that, given that cables would be inevitably close? Are there design choices I could go with to minimize the effects? Traces and spaces? Would putting the SBC in a plastic or metal enclosure help?

The ONLY other thing that was different was the VGA monitor I had it plugged into, but that's an out-only, right? Could VGA monitors feed back into a system and cause disruptions to code???

Any thoughts are appreciated. Thank you everyone!

Chad

Are you powering it via USB?

I've found my Ruby boards to be susceptible to low USB voltages - and I've found my desktop PC has lower USB voltage on the front ports than the rear ports! Also my laptop has lower USB voltage (round about 4.7v!) which causes my system to not initialise properly... When out and about with my Laptop I use a separate USB power bank/charger thing to power the board which works great.

-Gordon

_________________
--
Gordon Henderson.
See my Ruby 6502 and 65816 SBC projects here: https://projects.drogon.net/ruby/

Yes, I am. But I was using the same wall-wart that I always use here at the house. Still, if I see some weird stuff tomorrow I could at least try different power sources and see what works best. It's a start! Thank you Gordon.

Chad

Do you have any unprotected EPROM windows that could be catching UV from the classroom fixtures?

_________________
Got a kilobyte lying fallow in your 65xx's memory map? Sprinkle some VTL02C on it and see how it grows on you!

Mike B. (about me) (learning how to github)

No, I only use FlashROM's: SST39SF010.

But both of your responses worry me. I figured it could be cable interference but you both gave hints towards things I never considered!

Thank you Mike.

Chad

A few to consider beside potential power supply issue:
* Is all your electronics plugged into the same outlet?
* Are there difference in temperature between your home and school? Maybe board heated up during extended operation. TTL speeds up when hot but CMOS slows down when hot.
* Flex your board while running to make sure there is no hidden intermittent solder joints
* Tap your board while running to make sure there are no loose conductive particles
* Make sure all components are firmly seated in sockets.
Bill

Good list Bill!

1) Yes, would that then force each device to draw less power? Or could they be different voltages from different outlets?

2) Haha, long story, but essentially 'no'. Our building's A/C has been broken for 5 months now, it is *terribly* hot in my office most of the time, but recently they have supplied each room with a mini A/C.

3 & 4) That is a good idea! I will indeed be testing my boards like that now!

5) I do test that fairly often.

Thank you Bill. Your list yet again makes me wonder if perhaps it was not some kind of interference, but maybe something else. Maybe a bodge wire was pushed or bent strangely during transportation?

Anyways, update: During the Math Day event, both of my boards ran without ever locking up for at least 8 hours straight. The board that had locked up was fine. BUT, I did not want to test anything with it, so I connected it to the same power supply, same keyboard, and same monitor as it always has here at the house.

Thank you everyone!

Chad

Update: I set up my computer here at the house, and was starting to use it and... it locked up within a minute! So, I started trying things Bill had said, flexing the board, checking chips.

And it got worse, and worse, and worse. Soon it wouldn't even reset!

There are a TON of bodges on the bottom, and somehow, somewhere, one or more of those is just not happy at the moment. The board is very sick.

I think this draws conclusions: While transporting my boards to the office, I must have moved some of the bodge wires and it was marginal. When I brought it back home, I must have fixed it magically. It worked through Math Day, that's a miracle. Brought it home, and it's marginal. Flex the board and it's now broken.

Ever have a computer broken during transportation? Yup, this isn't the first time for me either.

Thanks everyone!

Chad

To me that sounds like it's time for some new botchless PCBs

I would certainly inspect & wiggle all bodge wires first, but this problem has the feel of an intermittent contact made with bare minimal pressure that shifted gradually with prolong operation or transportation or flexing. So I would also remove every IC, inspect its legs for oxidation and inspect the corresponding socket for springiness.
Bill

Im getting the same impression from the last few posts. we had this Keno machine once and it had this sort of thing, was fine in one location, moved, consistently got worse. Turned out to be a bad solder joint coming from a connector, I think the bill changer or something.

Its the 'it was fine, i moved it, it wasn't, I moved it again it was fine... thing', plus didnt this happen after the board was flexed?

@sburrow, check all your pads?
especially on the bottom where you said it had snagged on something possibly.
you may just need to re-solder like "one connection" with any luck.
however, if you had a lot of these 'bodge wires', flexing the board could have snapped a trace, I mean if they are 'that bad'.

let us know what happened, and have you had this experience with the PCB maker before?

Sure enough!

There were TWO broken bodge wire connections, one was an address line, another was something with the interrupts I believe.

I turned it on, and it was still not resetting right. So I lifted the bodge wires off the board a bit with a screwdriver and now it finally works! I think the reset bodge wire was pressing into some DIP pins on the bottom, causing weird results.

Whelp! Thank you everyone. Now I have learned. And yes Proxy, I think it's about time for a bodge-less board

Chad