Need A Logic Analyzer? Use Your Pico!

Screenshot of Pulseview showing capture and decode of some digital channels

There’s a slew of hardware hacker problems that a logic analyzer is in a perfect position to solve. Whether you’re trying to understand why an SPI LCD screen doesn’t initialize, what’s up with your I2C bus, or determine the speed of an UART connection, you’ll really want to have a logic analyzer on hand. People have been using a Pi Pico as a logic analyzer in a pinch, and now [pico-coder] has shared a sigrok driver that adds proper support for a Pico to beloved tools like Pulseview.

The specs offered are impressive. Compared to the $10 “Saleae” clone analyzers we are so used to, this thing boasts 21 digital channels with up to 120 MHz capture speed, 3 ADC channels at up to 500 KHz, and hardware-based triggers. The GitHub repository linked above stores the driver files out-of-tree, but provides build instructions together with an easily flash-able uf2 firmware. It’s likely that you’ll soon see this driver in a stock Pulseview installation, however, given the submitter-friendly attitude we’ve witnessed on the sigrok mailing list. However, if you need a logic analyzer ASAP, you should follow the caringly offered quickstart guide.

We’ve covered Pulseview being used in combination with cheap accessible analyzers before — a must-watch if you need to get yourself up to speed on the value they provide to a hobbyist. If an oscilloscope is what you need and a smartphone is what you have, perhaps you’ll enjoy the Scoppy firmware for the Pico.

We thank [mip] for sharing this with us!

18 thoughts on “Need A Logic Analyzer? Use Your Pico!

  1. Cool, I have a pico banging around that I’ll probably utilize for this once it gets folded into Pulseview’s master tree.

    One thing to keep in mind is the RP2040/Pi Pico does have a hardware bug (slightly mis-specced internal capacitances) that create ~9LSB non-linearity (0.22% of measurement range) at 4 points. Not a huge deal for most, but if you’re trying to debug some sensitive mixed-signal hardware with it, it’s something to be aware of.

    1. @fiddlingjunky said: “One thing to keep in mind is the RP2040/Pi Pico does have a hardware bug (slightly mis-specced internal capacitances) that create ~9LSB non-linearity (0.22% of measurement range) at 4 points.”

      Yes, you are right. The ADC HW bug was discussed widely around February 2021. I assumed it was covered here on Hackaday, but when I searched through the RP2040 posts here I did not come up with anything. If you will excuse me, I’ll post some info and links here in my reply to your comment:

      * Raspberry Pi confirms there is a hardware flaw in the RP2040 ADC. Refer to the RP2040 datasheet build-date: 2021-11-04, build-version: 150df05-clean, sections “4.9.3. ADC ENOB”, “4.9.4. INL and DNL”, and in-datasheet errata “RP2040-E11”.[1]

      There are Differential Non-Linearities [2] and Integral Non-Linearities [3] caused by mis-matching of on-die ADC capacitors.

      1. RP2040 Documentation and Datasheet

      https://www.raspberrypi.com/documentation/microcontrollers/rp2040.html

      https://datasheets.raspberrypi.com/rp2040/rp2040-datasheet.pdf

      2. Differential Nonlinearity (DNL)

      https://en.wikipedia.org/wiki/Differential_nonlinearity

      3. Integral Non-Linearity (INL)

      https://en.wikipedia.org/wiki/Integral_nonlinearity

      —-[OTHER RELATED LINKS]—-

      – ADC has very high DNL spikes and low-noise mode has more noise. #91

      https://github.com/raspberrypi/pico-feedback/issues/91

      – Big DNL spikes caused by an apparent mismatch between capacitors used in simulation and production, James Adams confirms.

      https://www.hackster.io/news/raspberry-pi-confirms-it-is-investigating-a-flaw-in-the-raspberry-pi-pico-rp2040-adc-95c393b55dfb

      – ADC Differential Non-Linearity Sat Jan 23, 2021 2:54 pm

      https://forums.raspberrypi.com/viewtopic.php?p=1802658

      1. I’m hoping for a cheap FT601Q-based logic analyzer. Shouldn’t be too difficult to add an iCE40 to buffer samples and maybe even do some RLE compression or triggering.

  2. Interesting….makes me wonder if it’d be possible to do something similar and get a 4 channel 150Mhz-166Mhz using Teensy 4’s high speed GPIO ports from the RT1062.

  3. I do wish the limitation of libsigrok to have the same sample rate for analog and digital were not still around.

    That made the Google Twinkie ( https://www.chromium.org/chromium-os/twinkie/ ) nearly useless with sigrok, because anything fast enough to sample the logic would lead to a MASSIVE file that would crash pulseview because of the huge pile of ADC samples.

    I have an ongoing on-and-off effort to reverse engineer Sony’s E-mount camera/lens comm protocols – among other things, it appears that a lens can request either regulated 5v or unregulated vbat on the motor power pin – I would only need a very slow sample rate for this, but since the digital pins can go up to 1.5 Mbits/sec async serial, you need a 6 MHz sampling of the digital I/O channels to get clean data. (running right at the Nyquist limit has always failed for me.)

    1. Oh cool, is there some documentation out there? How would you make an adapter to sniff? Broken camera and broken lens for the contacts? Another camera for a raspicam-mount :D

    2. The Nyquist limit is the bare minimum to recover a perfect sinusoid. As soon as you have any kind of modulation – you need a much higher sample rate to be able to capture it. You might be able to see the problem if you plotted some modulated data, and then only displayed discrete points at the sample rate.

  4. I’ve been using a $10 clone for years – the main use being i2c debugging – and it works well, and at 400K i2c you don’t need anything faster.

    That said, interesting project – but the key for any of these things is the software given the hardware is close to trivial..

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.