Total Pack Power vs. Ownership Lifetime

[BGbreeder]

Aside from the form factor (2170), they are not the same cells. Elon has said many times the chemistry they use for Powerwalls/Powerpacks/Megapacks is very different from the auto cells, in order to allow for deeper cycles without the same level of degradation.

My original comment was really toward the advantage of using the same item in two or more products, based on performance, like CPUs, not all of which make the highest performance standard in each production batch. Being able to sell lower performance items, increases the overall profit. Someone from an oil company once remarked that the key to a profitable process wasn't making the product more efficiently, it was having a profitable market for the sludge.

I admit that things aren't super clear, but for Powerwall 2s, the chemistry would appear to be lithium Nickel Manganese Cobalt

Tesla Powerwall: Cost, Features, and Complete Review

Here's our detailed Tesla Powerwall review - check out the up-to-date Powerwall costs and product specs for this popular solar battery.

news.energysage.com

Tesla Powerwall 2 Vs LG RESU Vs Sonnen ECO Vs BYD — Clean Energy Reviews

We review and compare the Tesla Powerwall 2 with the popular LG RESU lithium battery which is considered the leading competitor to the Powerwall, plus the well known German Sonnen and the large Chinese manufacturer BYD.

www.cleanenergyreviews.info

also used in model 3s, though they are due to switch to LFP (lithium iron phosphate) in China per this article; Tesla's New 12V Li-Ion Auxiliary Battery Has CATL Cells Inside).
and this article makes the case for NMC and NMCA(NCA) batteries across the Tesla line up, though missing the Chinese market;

Tesla Batteries 101 — Production Capacity, Uses, Chemistry, & Future Plans - CleanTechnica

Some industry observers have in the past argued that, rather than Tesla being an electric car company, it's actually a battery technology company that happens to sell electric vehicles ... and lithium-ion energy storage systems ... and solar photovoltaic (PV) roof systems ... and ...

cleantechnica.com

The MegaPack is supposed to shift to LFP;

Tesla shifts battery chemistry for utility-scale storage Megapack

The move to cobalt-free lithium iron phosphate batteries for the 1 GWh product could signal supply chain shifts.

www.utilitydive.com

So,... as is often the case, Elon's comment that you refer to might be kinda of true, but does not appear to accurately capture the complexity of Tesla's battery chemistries today.

BTW: 21700 is a battery size definition (21mm dia., by 70mm length) and comes in several different chemistries, identified by the suffix; LMO (21700IMR), LMN(21700-INR), and LFP(21700-IFR)

All the best,

BG

 

[BGbreeder]

I found an Australian battery testing group that has been running accelerated life cycle testing. Their testing methodology is cycling the battery at 80% (90%-10% I think) charge three times a day at hot/cold summer and hot/cold winter levels. They issue reports every 6 months and the latest report from Sep/2021 says that the Powerwall2 chemistry is Lithium Nickel Manganese Cobalt and after 2110 of their extreme cycles they SOH, State Of Health, was at 79%.

2110 cycles would be the equivalent of 5.78 years, so to have the Powerwall be at 80%. Eyeballing the degradation slope indicates that it would be SOH would be a bit lower than the warranty of 70% after after 10 years, but this is a very extreme cycling schedule that should be unusual in a real word application. They also note that the first PW2 had an internal fault and was replaced in Sept 2018.

View attachment 805162

I have been following their testing for awhile now. I have not seen an explanation from them about the data points that deviate wildly from the general decay line. It isn't unique to the testing of Tesla batteries, but it does make me wonder about the general testing methodology and processes. Any ideas?

All the best,

BG

 

[bkp_duke]

My original comment was really toward the advantage of using the same item in two or more products, based on performance, like CPUs, not all of which make the highest performance standard in each production batch. Being able to sell lower performance items, increases the overall profit. Someone from an oil company once remarked that the key to a profitable process wasn't making the product more efficiently, it was having a profitable market for the sludge.

I admit that things aren't super clear, but for Powerwall 2s, the chemistry would appear to be lithium Nickel Manganese Cobalt

Tesla Powerwall: Cost, Features, and Complete Review

Here's our detailed Tesla Powerwall review - check out the up-to-date Powerwall costs and product specs for this popular solar battery.

news.energysage.com

Tesla Powerwall 2 Vs LG RESU Vs Sonnen ECO Vs BYD — Clean Energy Reviews

We review and compare the Tesla Powerwall 2 with the popular LG RESU lithium battery which is considered the leading competitor to the Powerwall, plus the well known German Sonnen and the large Chinese manufacturer BYD.

www.cleanenergyreviews.info

also used in model 3s, though they are due to switch to LFP (lithium iron phosphate) in China per this article; Tesla's New 12V Li-Ion Auxiliary Battery Has CATL Cells Inside).
and this article makes the case for NMC and NMCA(NCA) batteries across the Tesla line up, though missing the Chinese market;

Tesla Batteries 101 — Production Capacity, Uses, Chemistry, & Future Plans - CleanTechnica

Some industry observers have in the past argued that, rather than Tesla being an electric car company, it's actually a battery technology company that happens to sell electric vehicles ... and lithium-ion energy storage systems ... and solar photovoltaic (PV) roof systems ... and ...

cleantechnica.com

The MegaPack is supposed to shift to LFP;

Tesla shifts battery chemistry for utility-scale storage Megapack

The move to cobalt-free lithium iron phosphate batteries for the 1 GWh product could signal supply chain shifts.

www.utilitydive.com

So,... as is often the case, Elon's comment that you refer to might be kinda of true, but does not appear to accurately capture the complexity of Tesla's battery chemistries today.

BTW: 21700 is a battery size definition (21mm dia., by 70mm length) and comes in several different chemistries, identified by the suffix; LMO (21700IMR), LMN(21700-INR), and LFP(21700-IFR)

All the best,

BG

Correct, the 2170 chemistry for PW2 is NMC, but it's not the same "sauce" as what you have in the 2170 batteries used in the cars. Elon has been very clear that they "tweaked' things depending upon the product.

Forward moving, yes LFP looks to be ideal for stationary storage (can charge to 100% with no consequences, much lower chance of thermal runaway, and no one cares that LFP weighs more per kwh than the prior chemistry).

Also, 2170 is used as layman, no one refers to them as 21700 (but you are technically correct that they are dimensions and nothing more).

 

[Redhill_qik]

I have been following their testing for awhile now. I have not seen an explanation from them about the data points that deviate wildly from the general decay line. It isn't unique to the testing of Tesla batteries, but it does make me wonder about the general testing methodology and processes. Any ideas?

All the best,

BG

I was thinking about this myself and I think that the outlier dots are testing/control errors. The cycling methodology isn't well documented for the important bits. They cover the temperature profile in detail and the times for charge, discharge and rest, but how much they are discharging is opaque for some reason. My best guess is that they think that the Powerwall reached 100% charge before discharging down to 5-20%. I want to assume that they know that the app doesn't report this correctly and they are getting it correctly from the API. Even with that, if the charge cycle didn't fully charge the Powerwall and they measure only 5kWh was discharged to drop the SOC to 20% then they calculate the SOH is 5/.80 = 6.25 kWh and you get a dot well below the trendline.

BTW, the footnote that they have on the Tesla Powerall page https://mikesgear.com/2017/12/07/monitoring-teslas-powerwall2-on-pvoutput-org/ is no longer valid.

 

[ParaAdBellum]

• Install date: Dec 2019
• Today's Date: Nov 2022
• Total Powerwall(s): 2
• Expected Energy (1x = 13.5kWh): 27.0kWh
• Current Reported Energy: 21.87kWh
• Percentage Lost: 19%

 

[zƬesla]

@ParaAdBellum That's your nominal_full_pack_energy and not nominal_energy_remaining, correct? The latter is how much the PWs currently hold (i.e. if they're not fully charged at 100%).

I had 2 PWs installed in Spring 2020 and one more a year later. My current nominal_full_pack_energy for all 3 PWs is 41352.

What kind of charge/discharge do you have that their capacity is down 19%?

 

[ParaAdBellum]

@ParaAdBellum That's your nominal_full_pack_energy and not nominal_energy_remaining, correct? The latter is how much the PWs currently hold (i.e. if they're not fully charged at 100%).

I had 2 PWs installed in Spring 2020 and one more a year later. My current nominal_full_pack_energy for all 3 PWs is 41352.

What kind of charge/discharge do you have that their capacity is down 19%?

Yes, it's nominal_full_pack_energy. I've discharged 10.45 MWh lifetime from 2 powerwalls that were installed in December of 2019. That's ~390 discharge cycles assuming 27kwh capacity. 10450/27 ~= 387.037. I only started paying attention to its' BMS capacity estimates around a year ago (see my earlier posts to this thread), and even a year ago degradation seemed high. Given that my power walls are not even 3 years old, and I already am at 19% degradation, I think there is a fairly good chance I'll be >= 30% before 10 years.

 

[h2ofun]

Yes, it's nominal_full_pack_energy. I've discharged 10.45 MWh lifetime from 2 powerwalls that were installed in December of 2019. That's ~390 discharge cycles assuming 27kwh capacity. 10450/27 ~= 387.037. I only started paying attention to its' BMS capacity estimates around a year ago (see my earlier posts to this thread), and even a year ago degradation seemed high. Given that my power walls are not even 3 years old, and I already am at 19% degradation, I think there is a fairly good chance I'll be >= 30% before 10 years.

And how did you check again?

 

[ParaAdBellum]

And how did you check again?

Check what precisely?

 

[h2ofun]

Check what precisely?

storage levels in battery. i forgot how to see via my computer

 

[Redhill_qik]

storage levels in battery. i forgot how to see via my computer

You log in to your TEG using it's IP address that is something along the lines of https://###.###.###.###
then using the same IP address the REST API page is https://###.###.###.###/api/system_status

Or you can connect directly to the TEG's wifi and then use https://teg and https://teg/api/system_status

Note: You will likely get a warning that the certificate can't be authenticated and you shouldn't go to the site, but use the Advanced link to proceed to the site.

 

[h2ofun]

looks like my batteries are down about 10% in 18 months

 

[zƬesla]

@h2ofun - is that in aggregate? might be interesting to look at them individually. All 3 of mine seem to be at ~102% !?

 

[h2ofun]

@h2ofun - is that in aggregate? might be interesting to look at them individually. All 3 of mine seem to be at ~102% !?

this is aggregate looking at the ones installed 18 months ago, not the 2 new ones, which are full value

 

[miimura]

An update on my system installed 2/28/2018
/api/system_status selected fields
"nominal_full_pack_energy":26545

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13237
"energy_charged":7454910,"energy_discharged":6361450

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13308
"energy_charged":7594600,"energy_discharged":6415450

I think those figures are pretty good considering that they will have been running for 5 years at the end of this Winter.

 

[power.saver]

An update on my system installed 2/28/2018
/api/system_status selected fields
"nominal_full_pack_energy":26545

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13237
"energy_charged":7454910,"energy_discharged":6361450

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13308
"energy_charged":7594600,"energy_discharged":6415450

I think those figures are pretty good considering that they will have been running for 5 years at the end of this Winter.

I have a similar 2 PW system installed in May 2018 and here are the numbers:

"nominal_full_pack_energy":27848

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13887
"energy_charged":7832500,"energy_discharged":6798870

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13961
"energy_charged":7884570,"energy_discharged":6831590

 

[zƬesla]

Tesla did an update a while back after which my PWs went from ~98% to ~102%. Not sure if that was just to make the number look better or if there was an improvement in measurement. So I take a lot it with a grain of salt.

 

[cwied]

An update on my system installed 2/28/2018
/api/system_status selected fields
"nominal_full_pack_energy":26545

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13237
"energy_charged":7454910,"energy_discharged":6361450

"PackagePartNumber":"1092170-03-G"
"nominal_full_pack_energy":13308
"energy_charged":7594600,"energy_discharged":6415450

I think those figures are pretty good considering that they will have been running for 5 years at the end of this Winter.

I'm a little suspicious these measurements are just rough estimates. My system was installed two weeks earlier:
"nominal_full_pack_energy": 27652

"PackagePartNumber": "1092170-03-G"
"nominal_full_pack_energy": 13850
"energy_charged": 7773100,"energy_discharged": 6664120

"PackagePartNumber": "1092170-03-G"
"nominal_full_pack_energy": 13802,
"energy_charged": 7821810,"energy_discharged": 6682610

This is with reserve set to 33%, time-based control and export everything turned on (so it hits the reserve most days). What I've noticed is if I lower the reserve to 20%, the nominal full-pack energy decreases. How is your system configured?

 

[power.saver]

I'm a little suspicious these measurements are just rough estimates. My system was installed two weeks earlier:
"nominal_full_pack_energy": 27652

"PackagePartNumber": "1092170-03-G"
"nominal_full_pack_energy": 13850
"energy_charged": 7773100,"energy_discharged": 6664120

"PackagePartNumber": "1092170-03-G"
"nominal_full_pack_energy": 13802,
"energy_charged": 7821810,"energy_discharged": 6682610

This is with reserve set to 33%, time-based control and export everything turned on (so it hits the reserve most days). What I've noticed is if I lower the reserve to 20%, the nominal full-pack energy decreases. How is your system configured?

I'm sure these are just estimates, but they seem pretty good and consistent across multiple installations.

I use TBC all the time, but my reserve changes seasonally from 20% in the summer to 60% in winter. I have noticed lower nominal full pack energy when the reserve is 20% and it discharges to this level. But it recovers to the values I posted above in winter when the reserve is higher.

 

[miimura]

I'm a little suspicious these measurements are just rough estimates. My system was installed two weeks earlier:
"nominal_full_pack_energy": 27652

"PackagePartNumber": "1092170-03-G"
"nominal_full_pack_energy": 13850
"energy_charged": 7773100,"energy_discharged": 6664120

"PackagePartNumber": "1092170-03-G"
"nominal_full_pack_energy": 13802,
"energy_charged": 7821810,"energy_discharged": 6682610

This is with reserve set to 33%, time-based control and export everything turned on (so it hits the reserve most days). What I've noticed is if I lower the reserve to 20%, the nominal full-pack energy decreases. How is your system configured?

I use TBC but I have a small solar system so I set the Reserve to 75% in the Winter because it makes no sense to bounce off the Reserve every day at a low state of charge - might as well bounce off the Reserve every day at a Higher state of charge.

Hopefully, I will get another 6kW of solar installed before the PW's reach their 5th Anniversary. It took a bit to find someone who would do what I wanted and work with what I already have installed.