Investor Engineering Discussions

[mongo]

How many of these extreme cold heat pump 'failures' do we know have been an actual hardware failure of any sort? From what I've seen the biggest issue seems to be sensors thinking something has gone wrong due to the vent being held open by snow build up and the software shutting the pump off.
Apart from that one broken pump in Russia, what other evidence is there that shows hardware is failing?

We (or at least I) have near to zero direct data since most repairs are handled by Tesla. So it's all reports by people who had their compressors replaced with no root cause visibility.
May be that the Russia failure(s) are rare or another issue separate from the general no heat complaint.

 

[MC3OZ]

We (or at least I) have near to zero direct data since most repairs are handled by Tesla. So it's all reports by people who had their compressors replaced with no root cause visibility.
May be that the Russia failure(s) are rare or another issue separate from the general no heat complaint.

We will soon find out if the software update makes a difference, for it to work the heat pump has to be operational.

Heat punps wearing out faster under these circumstances seems possible.

 

[dhanson865]

How many of these extreme cold heat pump 'failures' do we know have been an actual hardware failure of any sort? From what I've seen the biggest issue seems to be sensors thinking something has gone wrong due to the vent being held open by snow build up and the software shutting the pump off.
Apart from that one broken pump in Russia, what other evidence is there that shows hardware is failing?

https://www.reddit.com/r/TeslaModel3/comments/s51day

has a couple of pictures from someone on this side of the pond.

 

[Dangerous Fish]

https://www.reddit.com/r/TeslaModel3/comments/s51day

has a couple of pictures from someone on this side of the pond.

Where? I just see the same pic from Russia.

 

[MC3OZ]

Here is a thread on the issue with postings for Tesla owners in Canada:-

Recent heat pump failures - software issue?

I'm in Colorado with a July 2020 MYLR and the heat pump has always worked great. Yesterday got a VCFRONT_a531 error. The SC's estimate says they are going to replace both the compressor and super manifold. That seems intense. The heat is still working somewhat (I'm not sure exactly how 'bad' it...

teslamotorsclub.com

My impression is they are still unsure if heat pumps are failing, or if a software fix will do the job.

We should see some posts about heat pump being replaced in that thread, if it is happening.

 

[dhanson865]

Where? I just see the same pic from Russia.

I don't know if the picture is from Russia but the post I linked to was made by someone in Canada.

 

[Dangerous Fish]

I don't know if the picture is from Russia but the post I linked to was made by someone in Canada.

There is only one photo of a pump broken in pieces and it's come from Russia. Same photo the guy from Canada posted.

 

[mongo]

Normal if you open just to see why it stop working you broke it in part because this system is not for open it again when installed (it's because Tesla doesn't open it and replace it when not work ! I think it's no damaged but when they tried to open it they damaged the pump.

Automotive compressors can be disassembled nondestructively. Even if danage were to occur, it would not be the scroll surfaces themselves which would break. (Unless done maliciously)

 

[mongo]

When I learned that the Tesla system employ’s a two stage system (?) for higher efficiency at lower ambient temperatures, I was reminded of the potential (but ultimate failure) of a first generation attempt at that with the “Hallowell” dual stage heat pumps from the mid ‘00s.

Hallowell had two compressors in series and reportedly poor lubrication design.
Tesla is variable speed single compressor with a whole lotta potential operation modes.

 

[MC3OZ]

Here is a thread on the issue with postings for Tesla owners in Canada:-

So far there is no update from the Canada thread except for a Munroe video with footage of the shutters.

It isn't hard to believe ICE could freeze the shutters open..

Maybe that in combination with other factors causes excessive wear and tear on the heat pumps..

I'm tracking the Canada thread as sooner or later, I expect someone will indicate whether or not the software patch made a difference.

 

[MC3OZ]

The big unknown here is if Tesla would prefer to standardize on the 4680 format for all packs, or if there would be a format optimised for LFP.

Elon once mentioned something about more than one format.

With DBE at one time there was a limit on the minimum diameter. one approach would be to wrap the jelly roll around 2 poles spaced apart to create a squashed cylinder effect. That is 1/2 way between 4680 and prismatic. But even that change is a substantial change to the production line, all of the equipment from that point onwards needs to accommodate the new form factor.

Would DBE work with regular prismatic formats? IMO this is problematic.

DBE was only supposed to save 10% of the cell cost, but in terms of manufacturing equipment and Capex, we need to refer to Battery Day, but the difference is significant.

The is worth remembering, al of the Battery Day manufacturing processes were optimised around the 4680 format, building cells in another format is non-trivial.

I have thought about it more Prismatic could be laid down as a series of individual layers like a cake, but each layer of cathode or anode would need a tab.

I'm far from convinced that Prismatic is easier and cheaper to make than 4680s.

The remaining question is if prismatic LFP gives a significant volumetric energy advantages over 4680 and if that different is important.

For smaller cars and other compact forms of transport, the difference might be important.

 

[JRP3]

Prismatic definitely has a volumetric advantage over cylindrical.

 

[MC3OZ]

For me this is a required viewing...

Registration is free and it won't be recorded due to Jeff's request: Electric Vehicle Society - Canada Talks Electric Cars Webinar Episode #21 - "Recent Advancements to a Million Mile Battery"

Jeff Dahn will talk and describe recent advances in batteries work Feb 1st 4:30pm Pacific

EDIT: Corrected time to 4:30pm, thx @MP3Mike !

I'll be interested to see what Jeff has to say,

It was interesting that the Million Mile Battery didn't feature in Battery Day, but IMO there could be a good reason for that.

Entry level Robotaxis and Energy Storage could use LFP, but the Semi is unique in needing high-nickel batteries with long life.

Jeff is known to be an expert on NMC. and previous Million Mile Battery work did a lot of testing on NMC.

Speculation1 - The Million Mile Battery might be NMC 8:1:1, that would mean it contains Cobalt, costs more than what was revealed on Battery Day and is slightly different product. Elon once mentioned that a Million Mile Battery was more expensive. I could see Panasonic making the Million Mile Battery cells at GF Nevada specifically for the Semi. IMO Tesla will trust Panasonic with the Million Mile Battery formulation, but it will be closely held.

Time will tell if I'm accurate, or way off the mark, there is no point in bold speculation after the event. I admit being way off the mark is possible, but my confidence level about the Million Mile Battery being important for the Semi is high. I'm less confident about the guess around NMC.

Speculation2 - I've always seen the Semi as 12 Model 3 battery packs - 12 x 50/75 kWh.
Previously Megacharging could have been 4 x 250 kW, now it might be 6 x 250 kWh.
Rather than modules I see Semi packs as largely 12 independent mini-packs, with the ability to swap out an individual mini-pack. This is more for the reasons of reliability and redundancy, with a faulty mini-pack being able to be turned off, so the Semi can keep travelling. But I also think this for reasons of assembly of packs and management of packs. A single large pack is hard to build and it is a lot of cells to recycle/re-use when the pack is faulty.
A more modular approach does have overheads in terms of some additional cost and weight, but might be worth it.

For this particular speculation, the line between a mini-pack and a module can be blurry, a lot depends on whether Tesla finds it desirable to be able to swap out what I call a mini-pack. It also depends in part on how Megacharing is done.

While it is possible that a Semi pack could be a structural pack, we are talking a very big pack under enormous forces, The trucking industry is likely to be very conservative about things like that. Show the trucking industry redundancy where a mini-pack or motor can fail, and the Semi can keep driving, and they will love that.

 

[JRP3]

It was interesting that the Million Mile Battery didn't feature in Battery Day, but IMO there could be a good reason for that.

I think it was implied. 2,500 cycles for a 400 mile pack is 1 million miles, that is already within the range of current chemistry. A million mile pack was already possible even before battery day, it just needs to be more cost effective.

 

[MC3OZ]

I think it was implied. 2,500 cycles for a 400 mile pack is 1 million miles, that is already within the range of current chemistry. A million mile pack was already possible even before battery day, it just needs to be more cost effective.

Yes, I know my thoughts are speculation, that is one reason why they are in this thread.

However, Jeff is still working on something.

 

[JRP3]

Jeff is always working on something, he's been keeping things close to the vest in the last year or so, I hope he reveals something interesting.

 

[petit_bateau]

Speculation2 - I've always seen the Semi as 12 Model 3 battery packs - 12 x 50/75 kWh.

Rather than modules I see Semi packs as largely 12 independent mini-packs, with the ability to swap out an individual mini-pack. This is more for the reasons of reliability and redundancy, with a faulty mini-pack being able to be turned off, so the Semi can keep travelling. But I also think this for reasons of assembly of packs and management of packs. A single large pack is hard to build and it is a lot of cells to recycle/re-use when the pack is faulty.
A more modular approach does have overheads in terms of some additional cost and weight, but might be worth it.

For this particular speculation, the line between a mini-pack and a module can be blurry, a lot depends on whether Tesla finds it desirable to be able to swap out what I call a mini-pack. It also depends in part on how Megacharing is done.

While it is possible that a Semi pack could be a structural pack, we are talking a very big pack under enormous forces, The trucking industry is likely to be very conservative about things like that. Show the trucking industry redundancy where a mini-pack or motor can fail, and the Semi can keep driving, and they will love that.

The extent of the modularity, and that blurry line between mini-pack and module, may be related to how the charging is achieved. Multiple connectors would tend to lend themselves to mini-packs with a crisper distinction being a more natural outcome.

 

[Mengy]

Prismatic definitely has a volumetric advantage over cylindrical.

But prismatics also have a thermal heating/cooling disadvantage over cylindricals too. It is much easier to control cylindricals thermally.

 

[mongo]

Have we an idea how SpaceX handles that issue? Obviously that is tiny volume compared with Cybeertruck.

SpaceX uses coils. Each ring is a single piece 28.2 meters long.

 

[JRP3]

But prismatics also have a thermal heating/cooling disadvantage over cylindricals too. It is much easier to control cylindricals thermally.

Less of a concern with more stable chemistries such as LFP, and even NMC. NCA is the only chemistry I'm aware of that is never made in prismatic.