That Rigol mixed-signal oscilloscope I bought a couple of years back in an end-of-year capital-spending splurge?
Been using the logic-analyzer function for debugging that CPLD thing, what with there being rather a lot of signals in need of observation.
Sometimes, what I'm observing Just Doesn't Make Sense.
Well! Seems that some of the doesn't-make-sense stuff doesn't make sense because it's, as we engineers say, wrong.
As in: some of the logic-analyzer channels are acting like they're AC-coupled, and I just see a little spike of a few tens of nanoseconds on a low-to-high transition.
I thought at first that some of the probe wires weren't making proper connection to my test points, but that turns out not to be the case.
Running probe calibration gets some apparent error messages flashed briefly on the screen, and eventually a message that probe calibration failed, with no apparent way to retrieve details - thanks, firmware guys!
Opening the pod reveals that signal conditioning is handled by 8 chips, at 2 channels/chip. All solder joints look fine (the symptoms seen consistent with a bunch of loose connections, resulting in rather minimal capacitive coupling in place of actual conduction, but I'm not seeing those loose connections).
The cable shows signs of having been bitten by a cat at some point, but doesn't appear to have any broken conductors
Testing manually, by displaying all the channels, hooking up a ground pin, and stepping across the inputs, reveals life on channels 0, 1, 2, 4, 5, 9, 10, 11, and 15. This leaves 3, 6, 7, 8, 12, 13, and 14 being out of commission.
No obvious pattern to this, and it doesn't seem consistent with cat-bite damage. Dunno if the problem is in the pod, the cable, or the scope itself. Looks like a replacement pod is (at least from convenient sources) only available as part of a DSO-to-MSO upgrade kit. Which presumably has to include the bit internal to the scope, no? Probably a fix, but almost $300. Not today, I think.
Oh, well. I can stagger through with 9 channels, even if the scope won't let me selectively display just those. Maybe I'll fix it later. Also, I really should get a proper logic analyzer someday. (Hey, for up to maybe 200 Msa/sec, I could build one myself, around a Spartan 6 plus an MCU to handle the computer interface. And have lots of flexibility with regard to triggering. But that's, again, not a today project.)
And I could probably narrow things down by running some simple test signal into a few inputs with the probe cover off, and poking around with one of the regular analog probes to see if the signals are making their way through the inside-the-probe signal paths. Hm. Again, though: not today.
Actually, what I should be doing today is finishing up the schematic review, which doesn't require any test equipment at all. What it does require is rather more mental focus than working on tangible hardware.
[Hmmm. And how many hours do I need to erase from my timesheet? Guess I never got around to logging yesterday's time, so there's not really anything already logged that was, in hindsight, just me chasing my tail due to a problem with my own equipment. Everything up to yesterday wasn't affected by this.]
Update: So here I am, back to reviewing the schematic, and some parts are a bit questionable, calling for a spot of component research. Turns out Digi-Key is having a Very Bad Day; trying to get to anything beyond their main page gets me an error message. Oh, well; that's why there are other resources, but Digi-Key is so convenient when it's working.
Update 2: So here it is Tuesday morning, and Digi-Key is partly back up, in that I can look up components... but viewing an individual product page gets price-not-available errors. I would not want to be one of their IT guys right now!
Meanwhile, I've come across a possible-looking USB logic-analyzer-and-other-things dingus: Digilent's Analog Discovery 2. Small, not horribly pricey, should more or less get the job done (only 100 Msa/sec, but that's good enough for many purposes), and the software is available for Linux. Dunno about library support, talking directly to it using libusb, etc., but then I haven't been looking at it very long. Might even be something I could employ embedded in test fixtures. Oh, and it's based on a Spartan 6. Not ordering one today, but it's flagged for future interest.
Recent Comments