You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm suspecting that when using the alternate SCK pin on Teensy 3.2 and using hibernate something is not being properly shouted down before entering sleep. I'm calling SPI.end() before sleep but I see different current consumption if have previously called SPI.setSCK(14);.
Could you please confirm if the SCK pin is being applied the same configurations during SPI.end() if it is set for pin 14 like it is when it's using the default pin?
I'm using the SPI driver on the config block.
The text was updated successfully, but these errors were encountered:
I was seeing this behaviour in a word I design that has a lot more stuff in it. I have tried with the Teensy only and I'm not being able to replicate this very well. With the Teensy only I see around 52.2 µA while. I'm using a Agilent 34410A multimeter and powering the Teensy with 3.26 V.
I also notice that with time the consumption drops like 0.2 µA versus the consumption that it has as soon as it enters hibernation mode.
Hi,
I'm suspecting that when using the alternate SCK pin on Teensy 3.2 and using hibernate something is not being properly shouted down before entering sleep. I'm calling
SPI.end()
before sleep but I see different current consumption if have previously calledSPI.setSCK(14);
.Could you please confirm if the SCK pin is being applied the same configurations during
SPI.end()
if it is set for pin 14 like it is when it's using the default pin?I'm using the SPI driver on the config block.
The text was updated successfully, but these errors were encountered: