Hello,

The documentation says #get_chip_temperature returns chip temperature in units of 1/10 °C for Master Bricks.

I assume this is a typo, and it actually returns 1/100 °C.

 

The chip temperature is given in 1/10 °C.

However, while looking up if 1/10 °C is correct i discovered that the SAM3 microcontrollers use a different formula for the internal chip temperature compared to the SAM4 microcontrollers. For the Master Bricks we switch back and forth between the two types (they both have the same pinout and hardware units) depending on availability.

I attached a firmware that differentiates between the two formulas depending on what microcontroller your Master Brick has.

One more thing: The accuracy is given as +-13°C full-scale and an additional +-7% error per degree change. So if it shows 10° while the processor is at 25° this is within spec.

The chip temperature should however increase and decrease with the actual temperature (with a relative error of at most 7%).

Edit: Veraltete Firmware entfernt.

 

Thanks for your quick help and clarification.

Since the units reported already vary depending on microcontroller, would it not be an idea to switch to 1/1 °C once and for all?

I am eager to test the new FW, but as I already have 1 out of 2 Master Brick 3.1's stuck in bootloader mode, I must postpone this. I'll get back to you as soon as I can.

I know this is a long shot, but:

1. My Master Bricks have a type of microcontroller that is supposedly a 1:1 replacement for a different type, but which turns out to be different in at least 1 very small detail: °C scale.
2. When I combine my Master Bricks into stacks with Wifi extension 2, their Wifi status LEDs act weird, and the stack becomes unresponsive when connected only via Wifi.

Probably just coincidence.

2.5.1 Beta FW fixes this on my Master Bricks.