#SHDN and FREQ signals need voltage level translator

sinara-hw / Thermostat_EEM

Thermostat with 4 TEC and 8 sensor channels in EEM form factor

2 stars 1 forks source link

#SHDN and FREQ signals need voltage level translator #56

Open maciejprzybysz opened 2 years ago

maciejprzybysz commented 2 years ago

Same issue as in Thermostat standalone version, details described here: Thermostat Issue #110

nkrackow commented 2 years ago

Thermostat-EEM has the 10k pulldown, 2k to MCU resistors on the #SHDN pin, leading to 2.7V at the pin.

The same issue applies to all previous versions, but despite the difference in voltages, it has worked fine so far. Since release 2.2 additional 10k pulldown resistor R94 has been added to project (https://github.com/sinara-hw/Thermostat/issues/94 ). This pulldown in combination with a 2k series resistor (R31 & R85) on the #SHDN line made a voltage divider and a high level voltage dropped below 3V. As a result, MAX1968 stopped turning on.

Removing the 10k pulldown makes the driver go out of shut-down and work. @gkasprow

gkasprow commented 2 years ago

right, but I'd add 100k pulldown to make sure the driver won't turn on when the CPU is not initialized.

gkasprow commented 2 years ago

I would even replace R120 with a diode and add a pullup. That would work both as a voltage shifter and wired-and gate. To make sure the driver won't start when MCU is not initialized I'd add pulldowns at the STM PWM outputs @maciejprzybysz

nkrackow commented 2 years ago

Ok, thanks.

jordens commented 2 years ago

If you add pulldowns on the PWMs, make sure to add them at the CPU pin before the lowpass filter. Otherwise there will be an error.

nkrackow commented 2 years ago

Actually for developing now I'd only switch the pulldowns on #SHDN to 100k. I'll do more modifications as necessary.

jordens commented 2 years ago

Yes. I meant for design changes.

nkrackow commented 2 years ago

Some more observations:

No pulldown resistors on #SHDN leads to the following behavior:
- If the CPU stops (I don't know what this means exactly, I simply stop execution by terminating the probe-rs process) the output current goes to the value set by the DAC (which is persistent).
- When the board is powercycled with the CPU in reset the output slowly rises to the maximum 3A. (There must be some glitch coming from the CPU that goes above the threshold voltage of the #SHDN pin)
- The PWM limit pins and the #SHDN pin creep to about 1.5V. (The PWM pins currently also have no pulldowns)
- The DAC output goes to 0V which means maximum negative current for the TEC driver.
A 300K pulldown leads to a high level of 3.30V on the #SHDN pin, turning on the TEC at room temp. Stopping the CPU or powercycling wit CPU in reset leads to ~50mV and the TEC output is turned off.

gkasprow commented 1 year ago

added level shifter and pulldown. This should solve the issue.