Hi there!
Long story short, I spent some of my unemployment this year building a neat dashboard for my solar setup that reports data to me over Starlink. It shows output W, Input voltage (blue), battery voltage (red) and temperature:



Ignore the bug on the time scale at the bottom (those are hours of the day in 24hr format - the last digit is a bit broken cause I haven't bothered to fix the axis scrolling).

Anyway - my main question here is the weird temperature jumps you can see on the right. The physical setup is that I don't use the inductor thermistor, therefore it is floating a few inches above the PCB and the idea was to use it to measure AMBIENT temperature.

It looks to me like the temperature jumps with the solar input voltage, which is somewhat unexpected.

If it were watts, I'd expect the temperature to start dropping starting at 13:00 as absorb time goes on and power decreases as shown - but the opposite happens and the temperature is high as long as the voltage is high.

So, has anyone observed this before? I suspect it could have something to do with the reference analog voltage changing when the input voltage is high - maybe the DC/DC converter that powers the logic is moving a bit?

I'm not currently onsite to measure this but it's a bit of a mystery. Any thoughts?

Cheers,
Rafael
Edited 2025-10-28 09:45 by flyingfishfinger

My 1st guess was if the ntc was not secured in place, could it have drooped close to something warm? Possibly the voltage divider resistors for Vin?

If not, my 2nd guess is that from 13.00 in absorb (?) That the freewheel diodes are running hotter due to the high input voltage, and it's sensing this heat.

But I'm not sure I'm interpreting your plots right, as it looks to me like from 9.00 to around 13.00 it's in bulk mode (Vmpp) and subsequently in absorb (Voc), but I don't understand why Vmpp would be so much lower than Voc.

I can't see why the high input volts would affect the 5v rail, or one ntc and not the other.
Edited 2025-10-29 05:14 by mab1

  mab1 said  

But I'm not sure I'm interpreting your plots right, as it looks to me like from 9.00 to around 13.00 it's in bulk mode (Vmpp) and subsequently in absorb (Voc), but I don't understand why Vmpp would be so much lower than Voc.

My place is basically located at the bottom of a river valley and in the winter we get very little direct sunlight. What you're seeing is a clear (dark) blue sky with no direct sun until 13:00, so it won't even raise the voltage to MPP - not much energy in a clear blue sky. From 13-15 I still have direct sun.

It fares better on an overcast day, and I'm only running 1/3 of my panels at the moment.

When I go onsite next, I plan to move the thermistor fully outside the closet so it really is exposed to the environment - then I'll see how the trend changes.

Good point on the diodes, maybe it IS warming up the surroundings.

R