Strange issue with MCP3561 about arduino_mcp3x6x_library HOT 3 OPEN

hi @y3nd
are you using main or develop? they diverged currently a lot. of course main should be stable to use, but I suggest develop for new projects as there are some breaking changes and you might run into trouble adopting afterwards... but honestly I can not guarantee develop is working currently.

I can not remember ever experiencing your issues... a dirty fix which came to my mind is to increase OSR

more likely it is an issue with _getValue - which data format are you using? as far as I remember this function changed as well between those branches... converting the raw value into signed 32bit dependent on the data_format and resolution and there are bitmasks which might explain your issue.

from arduino_mcp3x6x_library.

I forked the main branch in December, i'll try the develop branch next but I don't think it's a conversion issue from the lib, since printing the raw values from the SPI transfer are showing the same 00000 for LSB. I saw on the web issues with SAR, and AD has an article about "stuck codes" but it doesn't match with the "specific stuck codes" issue I have, and my board has practically no noise as the analog parts are quite decoupled with their own power-supply etc.. but it could be flawed somewhere.
It might be an issue related to impedance matching with our amplifier, or specific high frequency noise affecting the ADC I don't really understand.
I have contacted Microchip support too.

from arduino_mcp3x6x_library.

I circumvented the ADC spikes issue by using differential mode. No spikes or stuck codes now, but really high gain all the time, about x20 gain. Setting the gain just multiplies the always-present x20 gain (0.33x = 6.6x, 1x = 20x, 2x = 40x etc..) really weird. I've contacted Microchip and they have no idea of what's going on. It might be a counterfeit, but I bought it through JLCPCB...

from arduino_mcp3x6x_library.