mystories ([personal profile] mystories) wrote2018-03-10 11:13 am
Entry tags:

More electrical debugging

I was working on an interface to a radar digitizer, and some things weren't working right. The head person on the effort was a freshly-graduated electrical engineer, whereas I had no degree but many years of hands-on experience. We hooked up a logic analyzer, and everyything looked fine. I wasn't convinced, so I hooked up an oscilloscope as well and watched the transaction that wasn't working correctly. I pointed out a signal that went from 0 volts to about 0.2 volts, then back to 0. "That's not right", I said. It should be 0 volts or around 5 volts, not some voltage in-between. The EE said that couldn't be the problem, 0.2 volts was a valid logic zero. I agreed that while it was a valid logic zero, it shouldn't be changing voltage like that. This is the sort of analog detail you can see with an oscilloscope that's hidden by a logic analyzer. I explained that normally waveforms like that show up when there's "bus contention", when one circuit is trying to pull a signal up and another circuit is trying to pull it down at the same time. For technical reasons, TTL signals like these have more strength pulling down than pulling up, so you get a lowish intermediate voltage. "That's not supposed to happen", the EE said. And he was right, it's not supposed to happen, but that is evidence that it is happening, so our next step was to figure out why. Once he'd accepted that we had a bus contention issue with that particular signal, we were able to quickly work out why it was happening and fix it.

Afterward, he admitted he'd been annoyed that I had been assigned to work with him on EE issues when I didn't have a degree, but seeing how I analyzed a problem showed him the value of real-world experience and the ability to challenge assumptions.