Unit got zapped by 48 volt

I worked with a device connected to a 48 volt DC power supply (400 mA). The voltage levels on the components are 3.3 V.

My brand new Logic Pro 16 accidentally got zapped by the 48 volt DC. I believe one of the uncovered tips (of the harness) of the ground channels on bank/group 1 touched the 48 volt rail and I saw a small spark.

I checked, the unit seems to be still functional. It’s detected fine by my computer, and it’s still able to capture. Based on the spec, it can tolerate up to 25V.

Even though I set the current of the power supply to be low (400 mA), but I am concerned that it may have damage inside.

So, what should I do at this point? Is there a way to do a thorough check to make sure all of the channels are ok? Should I buy a new one? Is it covered by warranty?

Glad to hear the unit seems to have survived! Yes, this type of accident is covered under the warranty.

If you have a signal generator handy, there are 2 tests we recommend:

  1. slow triangle wave test. This is basically the same test we use at the factory to test & calibrate units. Setup a slow ramp, 1 second duration, Across the full range of your Logic Pro 16 (or the full range of your signal generator, if it can’t reach Logic Pro 16’s range. (Note, Pro 16’s range is advertised as +/- 10 volts, but in reality, it’s about +/- 11 volts.)

Turn on all analog and digital channels.

Test the signal on every input port. 1 channel at a time is fine, I recommend just leaving the software running in looping mode while you move the input signal from pin to pin.

You should run this test 3 times, once for each threshold voltage.

Verify:

  1. the triangle signal looks correct.
  2. the channels next to the triangle signal have minimal, or no cross talk.
  3. make sure that the digital channel transitions at around half of the threshold voltage. (e.g. 1.65V in 3.3V+ mode)

This will verify most of the functionality of the device.

A second test you can run is to feed in a higher frequency signal (e.g. 100 MHz square wave, or the max square wave frequency you can generate w/ at least 1.2V amplitude) and record that on the digital channels at 500 MS/s.

If you observe any abnormalities, please post them here or send them in to support: contact.saleae.com.

If you don’t have a signal generator handy, you can repeat the test w/ a few fixed voltages, e.g. 0V, 3.3V, 5V, and make sure the analog & digital readings are correct, and that there is no shorts input to input.

Hi Mark, thank you so much for the clear instruction.

  1. For the slow triangle wave test, with +/- 10 volts, it looks fine, except some cross talks when I connect/disconnect the signal.
  2. For the higher freq, I can only get a decent square wave up to 5 MHz (verified with my scope). I saw constant cross talk on the analog when I set it to 50 MS/s. And when I set the digital to 500 MS/s, it changed the analog to 12.5 MS/s - I only saw “dots” on the analog, and the digital is ok.

Glad to hear it! Some cross talk is unavoidable, but you can usually reduce it my keeping the other wire harness wires physically away from the wires in use. if you notice any problems with the device, don’t hesitate to contact us.

Thanks. Btw, how about the square waves appears as “dots” on analog lines, during the high frequency test?

A screenshot might help.

However, if you record a 5 MHz square wave at 12.5 MS/s, it won’t look like a square wave, for two reasons:

  1. Our hardware anti-aliasing filter has a cutoff around 10 MHz (if I recall correctly) which will heavily filter that square wave before it reaches our ADC.
  2. We run a digital filter on the recorded signal that (A) flattens the frequency response and (B) introduces a low pass filter at just above 1/10th the sample rate.

The result is that the analog bandwidth is 1/10th the sample rate, across all but our very lowest sample rates. Our advertised 5 MHz analog bandwidth is only available at the 50 MS/s sample rate.

Here is what a 3.3V 5 MHz square wave looks like with my Pro 16, at 50 MS/s and 12.5 MS/s.

50 MS/s:

12.5 MS/s:

As you can see the square wave is entirely removed by our filter at 12.5 MS/s.

What I meant by “dots” during my high freq test last week was that the analog signal was flat, with “dots” (like the dots on your sine wave - which may indicate sampling points). I understand that I was at 12.5 MS/s thus it’s “flat”, but I don’t know about the “dots”. Because at 12.5 MS/s yours is “flat” and there is no dots.

Do you know why there were dots on mine?

So, I repeated my test, not sure why I got a different result because I kept the setup and equipment the same. So here is the screenshot - I got nothing on the analog at 50 MS/s. Do you know why?

Sure thing,

First off, the dots. The dots are sample points, and they only appear when you zoom in near the limit. When you zoom out, the sample points are too close together to draw.

Second, the missing data. Could you re-run that test? Usually I only see that if I enable the channels after I record without them. Any chance that’s the case here?

If you enable the channels, then record data, you should see data on those channels. It looks like you’re fully zoomed out vertically too - so there is no chance the voltage is off the screen to the top or bottom.

If you’re sure there should be data, please send us a copy of the capture.

Note, there is no hardware issue that could cause a problem with analog samples simply not being shown. If analog is enabled, it will parse analog data from the data stream, no matter what.

Thanks Mark.

So, that’s not the case, I had enabled those channels before running it and then pausing it to take the screenshot.

Btw, I run it on Ubuntu 22.04 machine. So, I tried it on Windows machine, and I got the same result as you (sine wave with dots, not just flat with dots). So it confirmed it worked fine. Please see the screenshots below.

Now, it seems that the issue is with my Logic 2 running on Ubuntu machine. I am puzzled as it worked before (with the exception of that flat line with dots) - which I believe it’s related to the way it works/displays/captures in Linux.

So, I would great;y appreciate if you can give some suggestion in how I can run it smoothly/correctly on Ubuntu 22.04 machine, especially the analog part.

Hi @peaceguy,

That helps narrow this down a lot.

Our application is failing to draw the analog graphs, which it does using webgl. We see this type of failure from time to time, and it should be quick to fix.

There are a few things to try.

First, please run Logic from the terminal, capture some analog data, and then send us the output from the terminal. We’re specifically looking for any graphics related warnings or errors.

Next, let’s try clearing the application’s shader cache. The simplest way to do that is to delete ~/.config/Logic. You may want to backup ~/.config/Logic/config.json first. That file contains all of your application settings, including any saved presets.
Then run and test the application again.

Pass or fail, you can restore config.json once the application has been closed. (running the software once will recreate the ~/.config/Logic directory, including the default config.json file. You can over-write it with your backup)

If that doesn’t fix it, you should try updating your graphics drivers. If you need help with this, please let me know what graphics vendor you use, and I can send you some pointers. I usually use the “Software & Updates” > “Aditional Drivers” tab, and select the latest proprietary, tested version of the graphics driver.

Note, after updating the drivers, there is a small chance you may need to delete the shader cache again, following the procedure above. We’ve really only seen one graphics driver update that required that though. You probably should also reboot the computer.

You can also run the software with hardware acceleration disabled, or with in-process GPU, the latter of which can get around some GPU related issues.
You can launch Logic2 with the --disable-gpu flag or the --in-process-gpu flag.
Please let me know if either of these work.

One more thing: You can load this page in your regular browser to check to see if it has any problems with webGL 2: https://get.webgl.org/webgl2/

If that’s successful, and you’re using a chromium based browser (like chrome) that might indicate that the problem could be fixed on a newer version of chromium.

-Mark

You beat me to it. I found this article and was verifying it when you posted your reply.

https://support.saleae.com/troubleshooting/analog-waveforms-do-not-appear-on-linux

I tried the solution there by running this below, and it worked:

rm -r ~/.config/Logic/GPUCache

Thank you so much for such a detailed and comprehensive explanation. Much appreciated.

A quick note (hopefully it’s useful for others): it helps not only with analog disappearing issue, but also when the analog signal looks weird.

Fantastic! Glad it was the simplest fix!

Please let us know if anything else comes up that we can help with