Current measurement is finnicky

[mryall-mawson]

Low-pass filter doesn't seem to be working properly.

[mryall-mawson]

Could measure resistance of heater trace over a range of temperatures.

[mryall-mawson]

Notes from discussion with @JDMawson, Mon 19 Feb:

• Agreed ADC sampling frequency should be increased by switching to MCLK (5 → 16 MHz) and decreasing sample hold time from 256 cycles to 96 (SHT0_8 → SHT0_5)
• Agreed we should align PWM with ADC sampling, so samples are taken immediately after (not during) MOSFET transitions
• Discussed PWM requirements for Hestia 2W:
  • UTS need configurable power range of 0-2W in 0.1W increments (20 gradations).
  • Boards are likely 6-7W, so PWM range will be 0-30% duty with resolution of 1 percentage point.
  • Frequencies lower than 600 Hz should be fine, as long as power source (USB battery pack) doesn’t switch off. (Is our input capacitance sufficient?)
• We have flexibility to use calculation instead of measurement if we like. So if measurement is not possible or inaccurate due to PWM, we can statically measure resistance (at various temps) on each board and use this to derive an expected duty/temp-vs-power curve.
  • I’m anticipating calculation might be best approach for low duty loads (<5%), where only one measurement can be taken.

Next steps:

• Test aligning PWM with ADC sampling
• Add signals for tracking end of interrupt handler and main loop execution, to ensure that these are only running when expected.
• Use electronic load to check ADC current measurements against a known voltage+current.

[mryall-mawson]

Voltage measurement reliability is now fixed by porting over the changes made to Hestia 2W, where we make sure the ADC measurements and PWM switching is timed appropriately.

Current measurement is still problematic on v2.2 due to the op amp issue, which is not fixable unfortunately.

Laryssa is going to use some resistance measurements taken at various temperatures to calculate power dissipation based on the voltage.