Weird temperature readings from 1 of 3 PWs

[youzer-name]

I was wondering if anyone here either has seen this before or has a likely explanation for the temperature data I'm seeing. I have 3 x Powerwall 2's. PW2 always runs the hottest, as they are stacked 3-deep and PW2 is in the middle, which makes sense. But for the past few days I've been seeing weird temp data from PW2:

PW1 and PW3 have fairly steady ups and down compared to PW2, which has a number of shark-fin looking peaks and valleys. Why do the readings look so different from PW2?

Possibilities:

• PW2 has some sort of fault with the battery itself
• PW2 has a fault with its temperature sensor
• PW2 has a fault with its cooling fan

I went back through about a month of older data and nothing I saw looks like the right half of the graph above.

Any ideas?

[jasonacox]

I agree, and the jaggies are a bit concerning. The big dip before the spikes feels like that would be hard to replicate with a fan so your sensor idea may be valid. But it could also be related to load. Do you see any odd Powerwall usage going on during that time? My system does show spikes but across both powerwalls and is directly related to A/C cycling (summertime). You can also see inverter fans kicking in about the same time.

Also, for comparison, I would look at the same time last year.

[youzer-name]

I didn't have vitals running for most of last year, so I can't do a direct comparison. I decided to open a ticket via the Tesla app and asked if they can run any remote diagnostics. I'll report back on what they say.

[youzer-name]

Thought I'd post an update on this. TLDR: No explanation for what I'm seeing, but Tesla support says everything is within normal limits.

PW2 is still showing temp readings that fluctuate in a very different pattern to PW1 and PW3 and in a different pattern than before around June 18th.

This is the last 24 hours where it is still showing shark fin-like rises and dips that the other batteries are not reporting:

I did open a ticket with Tesla and they said the temp reading represented ambient temperature and "The battery (pod) cells themselves are staying withing nominal operation range with no current signs of fan faults or low coolant. "

I then forwarded them some thermal images and screenshots from Powerwall Dashboard and they added "Although we can confirm a slight increase from 1 of 3 batteries ever since the 18th- the readings are not uncommon for the amount of use, sandwich placement, and unforeseen humidity." Note: the batteries are in my basement and there is a dehumidifier in the room, so ambient temps and humidity don't fluctuate as much as they would if they were outdoors.

Here is one of the thermal images I sent them:

The vented area behind the switch on PW2 is a few degrees F hotter than the other two.

The only other thing I noticed is that while my battery capacity chart has been fairly flat over time, right around when PW2 started showing these temp fluctuations, all three batteries showed a drop in capacity:

This is the last 90 days:

This is the last 2 years:

It looks like there was a slight drop last summer around the same time, but nowhere near as steep.

I was thinking maybe a firmware update resulting in different readings might explain things, but my last two firmware updates were on 6/6 and 6/28, so that doesn't line up with the changes starting around the 18th.

I've been keeping my batteries at 50% reserve since around March. I keep them higher in the winter since I can't generate nearly as much solar during the day to recharge them and I'm on a flat-rate plan, so they're being used for backups rather than TOU arbitrage. I might drop the reserve to 20% or 30% and then set it up to ~70% and see if different discharge cycles make a difference.

[jasonacox]

Yikes! that is a big drop and the thermal camera image (awesome by the way) seem to tell a pretty obvious story. I would think that you would have to be really pulling heavy kW from those to get that kind of degradation curve. It is also odd that it is all three experiencing the same thing but you only see PW2 having heat stroke.

[youzer-name]

Update on this. I got firmware 25.34.0 the other day and it seems to have changed the temp readings coming from PW2. The blue line is the firmware change... the difference in the yellow graph line for PW2 is pretty obvious.