pico-sdk
pico-sdk copied to clipboard
raspberrypi
Unstable programming interface: DAP init failed
The pico RP2040 hardware dev board (not wifi). The latest version SDK 1.4.0 release The OS's it occurs: Windows and Ubuntu Architectures: AMD64 and ARM64
Issue: Unstable programming interface
In the last month, a lot is the unstableness of updating the microcontroller, when connecting a new circuit board with the PICO programming/debugging is not an issue, slower when writing the binary than it was months ago, but it works. After some time the pico does not want to respond anymore when you want to program it. If you solder thick wires it is programable again, the programmer I use was always very stable and reliable.
When I check with my oscilloscope I see the clock line and data line communicating with the programmer. The weird thing is that it hangs it the error DAP init failed.
Making a thick connection with GND did not solve this issue.
The problem occurs with multiple different boards and different programmers.
I presume you are talking about the SWD (Serial wire debug) interface? You should always ensure the shortest possible path from the host to the Pico when using SWD. Don't use any sort of breadboard in between, this will cause the problems you are describing. So, short wires, directly connected from, for example, the debugging Pi to the Pico.
This is covered in the Getting Started guide.
Yes, I am talking about SWD, and of course shortest path possible. I did not experience any issues with communication before. Even lowering the clock did not help. The paths on the end board are gold plated.
So what has changed in your setup? The HW hasn't changed, and i don't think anything in the software stack that would cause this has changed.
I guess windows, maybe a simple driver issue? I am not sure what the exact conditions are to trigger DAP init failed
If you can reproduce this with a Raspberry Pi and two Raspberry Pi Picos, as detailed in Appendix A of https://datasheets.raspberrypi.com/pico/getting-started-with-pico.pdf I guess that would give a good "common starting point" for further investigation... but if you're using your own custom-built programming hardware, I guess the Raspberry Pi engineers can't really replicate that? :man_shrugging:
@lurch I want to understand what can trigger it, and the RPI engineers will possibly have more knowledge on this topic than I would have. Having to program hundreds of pico's the setup described in the getting started is not really an option.
The trigger is usually too much resistance or capacitance in the cables which degrades the signal and makes communication erratic.
Have you tried reducing the SWD speed with your current setup @martijnvwezel , to see if that makes things more reliable?
The title of this issue is very unclear.
Is the issue actually, that your version of a picoprobe appears to work less well when compiled against SDK1.4.0 vs SDK1.3.1?
I reduced the speed of the SWD, but it did not help. I double-checked the GDB output log and there it stated it was set to 1KHz. Weird this is/was during measurement that the scope saw normal signals coming to the PICO. I am not sure how slock skew has an impact
closing as hasn't had an update in a while - this doesn't specify what you are using to connect to SWD; still better asked in the forums https://forums.raspberrypi.com/viewforum.php?f=143 or in the picoprobe repository