Hmm, I would expect our custom Pico 2 W firmware and Plasma 2350W to be broken in roughly the same way, but since they don’t share the same config it’s entirely possible one could be misconfigured while the other is fine.
Tbh vanilla Pico 2 W firmware should, in theory, just work on a Plasma 2350W and would demonstrate whether or not the hardware is at fault, or if it’s just a quirk of our current catastrophe of a software ecosystem.
I can’t see any obvious configuration difference, except that Plasma2350W’s build enables the new, as yet unmerged, and mostly only tested by us, dynamic CYW43 support whereas Pico 2 W currently does not.
@SirFico could you share the specific waveshare board and the code you’re using to run it?
Having missed the very obvious above, these are in very close proximity to GPIO29, which is the wireless clock pin. My wild guess would be that there’s a solder bridge between GPIO29 and GPIO28. This should be caught in testing (we run a self test on-chip, but maybe the CYW43 complicates this).
Could probably verify this by running the same code but changing MISO from GPIO 28 to something else. The screen wont work, but it’ll still be trying to work… assuming it’s write-only!
@SirFico: the MISO is only used for the touch-controller and SD-card, which you are not using because of missing pins. So you might get away with just not connecting MISO.
@bablokb
Yes, a miso connection is not required. I have tried not assigning a MISO pin (and also not connecting the plasma pin 28 to the screen), but I suspect in that case a default MISO pin is then assigned by micropython. Anyway without a MISO assignment the waveshare screen still works, but once the screen has been used, as before, the wifi still gives the same errors.
@gadgetoid
There are a limited number of pins to use for SPI and the only pin I found to use for SCK was pin 26 which is on SPI bus 1 according to the plasma pin labels hence using MOSI on pin 27 which is also labelled as bus 1. (and likewise the MOSO on pin 28)
I have tried to use the SPI bus 0 pins, but a SCK pin does not seem to be available, though there are 2 CSn0 pins (gpio 1 and gpio 21) Anyway I’ve tried some different combination of the available pins for SPI 0 and I end up with a bad sck pin or bad mosi pin. The only good combination appears to be the ones I used as per my post.
I used the same pin connections as before and it worked :) . The screen worked and so did the wifi. I’m not sure what I’m now missing as I did not use the plasma 2350 W version. Must be something I presume.
Thanks to @bablokb@alphanumeric and @gadgetoid for your assistance and I’ve now got the waveshare screen, wifi and an LED strip purring away. Now to find an i2c temperature display to attach to the Qwiic port… 🙂️ Next up, a go with the beta beta version when it newly arrives.
@alphanumeric I meant a sensor :) and I’ve found one I had lurking in my pile of electronic stuff. And here is a picture of the plasma driving the screen and a long LED strip. Now I need to make a pcb that can be plugged into the screen and have the plasma board plugged into it and loose all those jumper leads. (the screen was tested with some old dummy data in case you think I’ve been doing some timetravel :) )
My Plasma 2350 is driving a 144 RGB LED strip used as my desktop lighting. Nothing fancy at the moment, I will be adding some patterns for showing off. I use a Rotary Encoder Wheel Breakout to control the color and brightness.
Lots of environmental monitoring setups on the go here. Indoor outdoor temperature and humidity etc.
Could you share the 3D-model for the screen? I am using mine currently only with standoffs.
Another suggestion: don’t plug in your temp-sensor into your Plasma. Use a dedicated system for that and send the data using Wifi/BLE to the Plasma. Always-on systems are not ideal for temp-sensors, the readings will be off sooner or later.
@bablokb I give the stl files of the 3D model, but its not that good. Its just a box without a back and a hole in the top (or bottom) to attach the usb cable to the pico. The screen is then held in place by screwing a support strip (also 3d printed) to each side of the screen within the box. I give a picture of the back of the box where the screen is temporarily using jumper wires so you can see what I mean.
Unfortunately files that are not image files are prevented from being uploaded to this forum. If, after reviewing the enclosed picture of the back of the box you still want the .stl files, then you would need to let me have an email address.
No need to upload. I understand the idea behind the design and I think I can replicate it. Those frameless displays are always problematic. I always try to use fraction-fit if possible, but that takes many iterations to get it right (sufficiently tight but not too tight). The idea with the support strips is appealing and I will see if I can use it.
