RP2040 sampling incorrect

[Volhout]

I used version ula_0.0.2 and also version ula_0.0.3 on 2 different genuine RP pico's (no clones).
With PulseView (nightly build, the last "stable" build 0.4.2 does not connect to the RP2040).

When pico is connected to a stable square wave generator signal of 100kHz, 3.3V high, 0V low, (GP15) withg good ground connection between generator and RP2040 I see 2 artefacts

1/ Sometimes data from channel 15 shows up on the channel 7 waveform in pulseview. This can very well be pulseview causing it.

2/ When I use sampling rate 1MHz, I see a nice square wave. When I use 10MHz the waveform is irregular (some pulses are not the correct pulse width). When I sample at 100Mhz the 100kHz square wave is not recognisable. In each pulse there are missing samples (low in stead of high).

Note, I am not using triggers, just RUN in pulseview.

[kukrimate]

I seem to be having the exact same issue.

It seems like a Linux-only sigrok bug to me however, as it only happens under Linux, not macOS, have tested both.

As an absolutely wild guess, this might be related to the Linux serial port code opening the USB tty device in cooked mode for some inexplicable reason, and some bytes being treated as ANSI escape sequences.

[SebKuzminsky]

I'm seeing something similar here.

µLA @ 100 MHz shows glitchy signals (esp clock):

fx2lafw on the standard cheap USB LA (slower sample rate, but showing the SPI bus itself has no glitches):

[Gustavomurta]

I built a test circuit using a 74HC4044 binary counter with an input clock of 4.9152 MHz. Therefore I obtained digital signals with low frequencies from 19200 Hz to 2400 Hz. I tested the capture on a PC Ubuntu 22.04. I compared the capture with a Salae Clone analyzer. The capture of the ULA Logic analyzer is not at all reliable. I think the RP2040 code needs corrections.