@calin 在 M5Stack MultiApp Firmware 中说：
Any other suggestions ?
Since oscillo works, you can also graph FFT and from that detect Hz +THD.
Benefit: detects the quality of the electricity and helps troubleshoot and predict issues.
I started a project on that, measurements are out of wack, maybe we can work together on it?
https://github.com/laurentopia/M5-Signal-Multimeter
Thanks for pointing me to the right way to access the SD card.
Samsung 32GB passes the write/read hello.txt
It's this model: https://www.amazon.com/gp/product/B06XWN9Q99/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1
It took me 2 days to pull the trigger on it because when the cpu is $40, it feels wrong to pay $12, but look at the adapter, and the speed, if the SD controller isn't the bottleneck it's well worth it.
@jimit thanks I'll do that.
@m5stack the warning in the get started can be removed http://www.m5stack.com/assets/docs/index.html
The DSO112A generates signal that's ok clean.
This M5Stack oscillo graphs something far noisier.
Why is that? Can this be fixed without lowering the sampling rate by smoothing the analog signal?
EDIT: I connected to the wrong wires and didn't ground. Once grounded, disconnected USB and pinned to 35, the signal is very clean. Weird that I'd get a squarish signal when connected to AD
It looks like someone already did it
https://github.com/Bodmer/TFT_eSPI
@pkourany inside user_setup.h comment out the generic setups and uncomment m5. Then most examples work.
Result: same performances as m5 lib.
EDIT: even with SPI_FREQUENCY at 80,000,000 ... weird
answer: use the TFT_eSPI library, you can pass transparency color in the sprite draw functions