Jeez this thread really blew up in a short time!
So. Hi, "that guy" here. Unfortunately I must embrace it, and recognise I'm the party pooper. Sorry.
There's some really really good advice in the OP, surrounded by some really false details, and also a bad suggestion or two. This does not eliminate the fact there are good points though, which I've summarised at the bottom.
I struggled with posting this, as it
seems combative and/or argumentative? But that isn't my intent.
@SomeJoe7777, this ain't your fault for "getting wrong" so please don't see it that way. Seems like you were fed a bad explanation IMO. This actually happens a lot from Tesla Service employees, unfortunately (as they're pressured into offering
some explanation with
their current understanding). You brought new information to us, which is very appreciated!
The open circuit voltage is indeed important, this has been known for a while but good to talk about directly (it gets lost in the misguided "balancing" threads every time). The voltage swings quite wildly when driving, but your battery gauge is resilient to these swings while still providing a very accurate report of remaining capacity. This is an admirable component of Tesla's battery meter.
There's much for the BMS to learn during a drive (changes during discharge and regen can reveal a lot), but a non-sagged open-circuit voltage is fairly representative of the battery state especially when battery temperature is also known (and indeed, it is to the car).
---
The bleed resistor thing sounds all wrong. Balancing setups generally aren't like this (maybe never?), for many reasons:
- The bleed resistors, for balancing, cannot (reasonably) be permanently draining
- By definition, the battery is not open-circuit if there's a reasonable load across any of the cells (e.g. for balancing)
- Being the "primary cause" of vampire drain measured in the wild would require some unrealistic values
- Stats says ~60W median for LR. 96 series bricks means each brick drains 0.625W
- Implies chonky 1W resistors. These would be visible on the BMS boards. I see no such resistors. I don't even see 1/4W ones.
- At 4.0V (reasonable guess of average SoC, ~73%), that implies a ~25 Ohm resistor. That is a very, very low value for a "bleed" resistor.
- If the bleed resistors were 10x higher (and always engaged), voltage would reach steady state faster, not slower. Higher resistance equals less power draw. Less power draw sags the battery voltage less. Less sag means being closer to steady state or "open-circuit".
- That would also imply 2.5 Ohm "bleed resistors" on S/X. This is more into current shunt territory than bleeds, and would indeed cause massive vampire drain if always engaged. But S/X are not that bad.
- Most importantly, bleeding all bricks simultaneously just wouldn't be an effective way to balance - no balance is actually occurring this way, just a fixed drain no matter what the capacity of that brick is, which would actually induce imbalance over time.
---
I find 3+ hours to stabilise voltages
incredibly hard to believe (yet I have no data to support this). What's more believable to me is that this is an upper bound for some common tasks Model 3 is doing before going to sleep:
- Sometimes the car fetches new firmware and/or map data after a drive. It seems to remain fully awake while it downloads. As downloads are large and the car often gets spotty WiFi (garage or outside), these take a while.
- Sometimes the car will upload data after a drive. See above, but note upload speeds are generally much slower on home internet.
- The blower will continue to run for a surprisingly long amount of time if the AC was on, to reduce condensation leading to the moldy smell.
- When the battery is very warm, it seems to stay awake longer (not necessarily using the chiller to cool it off, just running the pumps and going through the radiator).
Now, stabilised could've meant "balanced" I suppose, which
kind of makes sense in the post? But this doesn't seem right?
---
Anecdote time. It just so happens that ever since switching to 240V charging (months and months ago), our charging habits would be ideal for this OCV theory. Especially in the pandemic where we leave the car unplugged at various percentages now, it has more of a chance to gather more readings.
The result of that? The reported capacity plummeted nearly all at once,
no gain. It also still varies +-1% despite all this seemingly good data it should be getting. And where it ended up seems about right for a healthy battery, reporting about 5% degradation on a car with 40,000km on it.
What I mean to highlight by this is that, if true, it can go both directions. One person may "gain" range, another may "lose" it.
---
I'd put these as the "good takeaways" of the OP.
- Let your car sleep. This has multiple benefits, but accurate measurement is one!
- Yes, don't run Sentry 24/7. Car wasn't designed to be a dashcam, and does that job poorly.
- Any "calibration" is just that, and hasn't increased the actual capacity of the battery.
And don't do these:
- Don't: Charge to 100% to "balance". This happens at lower percentages anyways as well, and balancing to an uncommon state actually offsets it in a non-ideal way. Balancing would also take a lot of time, so doing this right before a trip doesn't accomplish that anyways.
- Don't: Skip charging to provide more data points. While I've found some vampire drain benefits to being unplugged, not having the capacity when you suddenly find need for it is not good. There is no sense in stranding yourself because you left the car unplugged at 20% just so the battery meter is very very slightly more accurate. That's like 50mi of range of a brand new SR+ in ideal conditions on flat ground.
- Don't: Take TeslaFi/Stats/etc. graphs as accurate. There's so many problems with the way these report capacity in general, and can paint pictures that don't reflect reality.
- Don't: Take Tesla Service's guidance on everything regarding batteries. Some do have some generally good advice, but some have their own wild theories. They are often not battery experts, they are mechanics. That said, don't take some internet dude's advice either, me included!
---
Rapid question answering time! (it seemed like most of these weren't responded to)
Yes. The AC and heater are both high-voltage components and will draw from the battery at all times if on.
My suggestion would be to not leave the climate controls on. Turning them on even just 2 minutes before you get in the car does wonders, if possible (I realise connectivity can be an issue in some places).
If you have Cabin Overheat Protection on (it is by default), that could explain some of your drain (but it doesn't run that if it hasn't been driven for more than 12 hours). The computers are actually liquid cooled, and the AC wouldn't need to run for that (just passive heat rejection via the coolant loop and radiator is way more than enough).
Mostly, yes. Small effective capacity gains may happen via balancing, but that's a whole separate topic IMO. If there are gains to be had by balancing, they are small. If you have significantly reduced capacity due to balance issues, you have a physical battery issue that is beyond correctable by balancing.
The hole may be this: are you sure it's exactly 80% with 268 miles reported? Probably not, it's really hard to get it to hit 80% exactly.
Yes. Up to about 3% before you see the "snowflake", and presumably more after that. It's gradual.
Replacing the breather valves would have nothing to do with your battery capacity, measured or real. I don't think they even do a high-voltage disconnect for this procedure, so your car may literally not notice anything being done to it.
This is absolutely the case and understanding outside this thread, yes. Most "calibration" procedures, recommended either here or by Tesla, do technically wear the battery more than otherwise.
If it's not made clear by this thread, there are too many variables at play to make a blanket suggestion for charging habits that is beneficial to everyone.
The singular clear thing is that charging above 90% routinely is almost never good, and
that is the one recommendation they're aggressive about (the car will even warn you if you charge to 100% multiple times in a row). Anything further recommendation needs to account for too many local variables and personal use-cases.
I suspect the unreliability of the below-0 capacity is due to voltage sag. They start cutting max power
heavily at low SoC to prevent voltage sag from going outside the boundaries of heavy damage (made worse by the reality of higher sag at low SoC). My guess is they cut the battery out the moment a cell exceeds or touches a voltage threshold. You'd be especially at the mercy of cell variances at this point too.
