Featured Discussion: NexPower V3 Sodium-ion Battery Announcement

Just wanted to post my initial thoughts after watching @jacktheripper's V3 sodium-ion battery announcement:

First, congrats to Jack and team for their hard work completely redesigning their product. Not quite at the finish line yet (release date "2024AUG"), but today's product announcement is certainly a huge hurdle.

Given that I don't yet have physical V3 hardware to review, for now my comments below are my own opinions based on statements made in the above video. The feedback I offer below is identical to the questions I would ask a co-worker during an internal engineering design review process:

@1:04 "The lithium battery just isn't designed to operate in extreme conditions in some corner cases".
I agree that NexCell's specific lithium cells were ill-suited for the Prius' huge current demands. The 'corner cases' would be anyone who drove the vehicle hard and/or in harsh conditions. However, this statement shouldn't be applied to all lithium cell designs... just those that use undersized cells and/or lack a properly designed BMS.

...

@1:10 "In 2023 we went back to the drawing board".
As late as 2023DEC22 Jack and I were in talks to adapt my BMS electronics to his lithium NexCell modules. While it's certainly possible Jack was working on sodium in parallel, I find it odd that he would ask me to design a BMS for his existing lithium module design if he was already working on a replacement sodium product in 2023. In other words, I suspect "V3 is sodium" was at most a twinkle in Jack's eye circa 2023.

However, @5:27 Nick states that V3 "has survived nine months of rigorous testing", so maybe my hunch here is wrong. But then again, @7:28 Peter mentions that they performed their testing "in a matter of months, not years".

Why is this important? I'm skeptical that NexPower could properly design, prototype, validate, and verify a replacement sodium-ion solution in six months (e.g. 2024JAN-2024JUN). Obviously they do have working prototypes (as shown in the video), but I'm worried which engineering corners were cut given the super short development cycle. The most likely explanation is that they've simply placed sodium-ion cells into a snazzy looking mechanical enclosure, and have skimped out on a properly designed BMS (again).

Of course, the sodium chemistry is much more fault tolerant than lithium (LFP or NMC)... but if NexPower still hasn't added a BMS, then I have huge concerns regarding long term reliability. Na-ion cells will still run away if overcharged. While their failure mode is safer than lithium, they will still fail if overcharged. Hopefully there's a properly designed BMS. Maybe that's what Jack was asking me to do, but he just wasn't telling me what I was designing it for (i.e. sodium, rather than lithium).

...

@1:20 "We work with scientists and engineers from many cell manufacturers".
I'm certainly interested in determining which Na-ion cell V3 uses. Once I know the part number, the first thing I'll look up is the rated charge/discharge current.

@1:30 I haven't seen any Na-ion cylindrical cells this small that are rated to handle the Prius' huge current demands. However, I don't have access to all the behind-closed-doors specifications in this emerging field; certainly possible V3's cell is used within charge/discharge limits; impossible to tell without knowing the part number.

...

@1:26 "A robust solution that's going to last a long, long time in hybrid applications"
And yet NexPower is only offering a 1 year warranty to DIY customers, and 2 years to certified installers?
Commercial sodium-ion cells are bleeding edge technology. While BYD has released an Na-ion vehicle, they aren't yet in widespread use. Time will tell if "long, long time" is an accurate statement.

...

@1:30 there is text visible on a test cell, which I've reproduced below. It's not 100% legible, so any uncertain text is bracketed ('[' to ']'):

How well does the Prius handle this large voltage range?

Disagrees with data below... about twice as much capacity

Assuming the pack was tested at room temperature, this is a huge delta after one discharge event.

Assuming this is at the same 40 amp discharge rate, that's not much energy... maybe 9 Wh per cell?

At 150 amp load, each cell will self heat at 65 watts! Assuming there are QTY66 cells in series (250 volts / 3.8 VcellMax), this pack will self heat at 4.3 kW under heavy load. That's equivalent to placing QTY3 space heaters in your battery bay. For reference, a similar lithium system I designed self heats at only 0.9 kW under heavy load.

Yikes, comparing this to the above data yields a 20% ampacity reduction by increasing the current from 40 to 70 amps. Now we're down to ~7 Wh/cell. Assuming there are QTY66 cells in series, this pack is only 0.46 kWh.

So then the same 150 amp load will self heat each cell at 86 watts, which increases pack self heating to 5.7 kW! Yikes!!

Interesting to see the ESR is lower while charging. Probably measurement error (e.g. when the ESR was measured during the charge and discharge cycles).

...

@1:32 "We've updated the balance board design"
Looking forward to reviewing it.

@1:45 (image)
Looks like there's some kind of "Signal Soother" PCB. Impossible to tell what it does electrically based solely on the picture. It could be done safely as shown... but is it? Won't know until someone reverse engineers it. Send me a PM if I can borrow yours for a non-destructive review (I'll send it back when I'm done).

...

@2:05 "It has a super wide operating voltage range"
But does the Prius like this wide range? Or are you just using a small portion of the Na-ion SoC? Or is your "Signal Soother" lying about the voltage? Given NiMH's (and LFP's) flat discharge curves, I imagine you'd need to do something to modify the voltages seen by the OEM BSU. Can't comment further until I get the hardware.

@2:08 "Can even discharge to zero volts with minimum degradation"
Correct. Ideally, though, your BMS prevents this from occurring, right?

...

@2:18 "All the cells are connected internally"
Doesn't this make it difficult for the user to replace a single failed cell? Do they really have to replace the entire pack if a single cell fails?

...

@2:40 "We are now sourcing the cell from multiple manufacturers..."
Have you qualified each cell design from each manufacturer? In six months?

@2:43 "...instead of producing the cell in-house"
This is a wise decision. Manufacturing cells is exceedingly difficult, particularly in small volume. Congrats on this excellent decision.

...

@2:52 "But wait, there's more... introducing V3 GT"
OK, now I'm really confused. If you're keeping the existing (V2.x) form factor for your new performance V3 sodium product... why did you design a brand new architecture for the 'base' V3 product? I'm genuinely confused. Why didn't you just create the V3 GT product?

@3:00 "with a pouch cell" ... "ultra low internal resistance"
Again, if this is a better product, why release the non-GT V3 product at all? Particularly given my rough calculations on how much power the non-GT V3 pack will generate under heavy load. Why does the non-GT V3 product exist? So confused.

@3:15 "the mpg you will get with the V3 GT sodium ion is equivalent to the lithium variant, or even better"
I don't understand how this statement is true. The V3 GT appears to have the same internal volume as the existing V2.x lithium design. And yet sodium has considerably lower volumetric energy density. What am I missing?

@3:29 "We put a copper sheet between every cell to ensure proper cooling or heating".
Without a liquid coolant loop, a single copper layer between cells is going to have a hard time meaningfully dissipating 86 watts per cell (@70 amps discharge, see above calculation).

Finally, the V3 GT product appears to reuse the existing BMS and bus bars. Is the "Signal Soother-like" circuit board used on the non-GT product also used with V3 GT (but not shown in the video)?

...

@3:33 "Other than the research and development of the sodium ion technology, we spent most of our time..."
Wait, @2:40 you said the cells are now sourced from 3rd parties. So didn't they do all the R&D on the sodium ion technology? Or did you co-develop the cell chemistry? Huge if true.

@3:39 "We spend most of our time actually in testing"
Given the products I see in the video, I would assume you spent at least some time in R&D system design? As I will elaborate on next, IMO the testing you showed in this video isn't rigorous enough for a product like this. In-car testing is inferior to using proper lab equipment in a thermal chamber.

@jacktheripper, I will scientifically validate your sodium ion cells as I can get my hands on them. If you send me your sodium ion modules, I will perform this long term accelerated testing for free.

...

@3:46 "So with that I'm handing it over to our test engineer, Nick"
Besides being a 'professional courier' & 'rideshare driver', what are Nick's engineering qualifications?
He states that he is a 'test engineer', but does Nick actually have an engineering degree from an ABET accredited institution?

...

@4:07 "These tests are guaranteed to break a typical hybrid battery in the shortest time possible."
No... no they aren't. OEM hybrid batteries are tested properly, and in extreme situations. While manufacturers certainly perform harsh real-world tests, they also spend considerable time simulating the harshest conditions inside thermal chambers, so that they can gather data at an accelerated rate.

For example, a thermal chamber can heat cycle the pack across its entire temperature range hundreds of times in a month, while simultaneously charging & discharging the pack at its full rated value. And it's 100% repeatable, because they've controlled all the variables, except the cell under test. That's how you test "in the shortest time possible". Performing real-world test for six or nine months won't get you there... you need an accelerated tester to make the "shortest time possible" claim.

...

@5:00 Your "postal worker" test is a good real-world datapoint. Certainly verifying that the ECU disables the pack "at 120 degF as a safety feature" is valid.

However, it would be much better to also perform this test in a thermal chamber. I'm not trying to discount the testing efforts NexCell made, but they're not up to generally acceptable engineering quality standards for a product like this.

As I mentioned previously, I would like to properly long term test your sodium ion cell with my equipment. I will do this for free as soon as I can source the cells (either from NexPower, a customer, or directly from the cell manufacturer).

During this test, Nick also mentions that after nine months of testing, "life expectancy still shows over 100% capacity". This doesn't make any sense, and almost certainly means the initial capacity calculations were calculated incorrectly (e.g. errant data, etc). Operating the "postal worker" test QTY3 times a day for nine months would probably cycle the pack around QTY1000 times, which is certainly a measurable percentage of the cell's overall cycle lifetime. I am extremely skeptical about this claim. Of course I can easily verify the claim with a test cell and my accelerated lifetime tester.

...

@5:50 Your "happy camper" & "diarrhea" tests attempt to discharge the pack as much as possible. @6:34 I see the pack drop to 190 volts. Assuming there are QTY66 cells in series, that would mean you've only discharged each sodium ion cell down to 2.9 volts or so. Assuming the fully charged pack is near OEM (~250 volts), that would mean you're cycling these cells from ~3.8 to ~2.9 volts. If my several assumptions are true, then you are only using ~45% of the cells' SoC. Unless your 'BMS' is lying to the Prius computers about the actual pack voltage, that would mean the V3 product only allows the Prius to use half the energy stored in the pack.

As I mentioned previously, sodium-ion's discharge voltage curve is huge compared to LFP and NiMH. I suspect the Prius simply won't allow the pack voltage to droop as low as Na-ion operates, hence you're using a small portion of Na-ion's SoC.

At the end of this test, Nick claims @6:55 "If this doesn't damage the battery, nothing will". This statement is undeniably false. Jack explained why earlier in the video when he mentioned that you can discharge Ni-ion cells down to zero volts with minimal consequence (which I agree is true). Maybe Nick's statement is in regard to the entire test suite (not just the "happy camper" & "diarrhea" tests)?

...

@7:15 Peter Batch mentions "wind tunnel testing". I'm glad to see some sort of lab testing, but alas the testing shown in the video doesn't represent how the pack is cooled while installed in the car. Hopefully the actual wind tunnel testing performed is different than as shown in the video.

...

@8:15 "this journey wouldn't have been possible without the incredible support from the Prius community".
I know I've been a thorn in your side, but I just want to reiterate that my goal here is to inform Prius customers when safety issues exist. While I've only seen one member publicly reveal that they had a V2.x pack fail catastrophically, I've had two other members privately tell me that the same thing happened to them. In both cases, Jack went above and beyond to make the customer happy, but that doesn't solve the safety issues. Hopefully V3 is safer, but we won't know until someone (e.g. me) gets the hardware for evaluation.

...

@8:45 "V3 available 2024AUG for only $1500" & "V3 GT available 2024SEP for $1800"
Kudos for offering V3 at this excellent price point. If the design is safe and reliable, I will gladly endorse it in my review. I still don't understand why both 'GT' & 'V3 GT' exist, particularly given their small price difference. In other words, since V3 GT is a better product using the existing form factor, why completely redesign a 'GT' product with such high cell ESR?

...

@9:00 "It is a dream to finally bring the product to life after ten years".
So is this an admission that the existing V2.x design was bad?
Also, what is the upgrade path for your thousands of existing V2.x customers? Do they all get a discount to upgrade to V3?

 

While I was writing my the first post in this thread, a more general discussion thread regarding the V3 design was started in parallel. Moving forward, I'll keep this thread for my detailed brainstorming/analysis, and will post key points in that 'main' V3 thread.

 

Sure, can you please send a link to what I actually wrote?
And also let me know how you want me to elaborate? For example, are you questioning my desire to profit from selling a product?

Note that this isn't a 'review' per se, but rather my initial thoughts having just watched Jack's V3 release video. I make this distinction for legal reasons.

I have a lot to say. Maybe think of my initial post as a stream of consciousness, that I would ideally shorten and make into a video... which would also probably be too long. Again, I have a lot to say.

 

Let me know if you decide to take the time to find the post you're referring to. I just want to make sure you're quoting me correctly, and in context. No worries if you don't want to look it up, as I'll describe what I think you read next.

The LiBCM Honda Insight project is a labor of love. I never expected to make money doing it, because there were so few G1 Honda Insights made (QTY17020), and even fewer left on the road today (someone went through DMV records and estimates ~QTY8000). For that reason alone, I knew I would never profit from developing the LiBCM project. However, since I personally drive a G1 Honda Insight, I decided that since I was already doing this project for myself, I might as well offer it to the community. I actually just sold my QTY250th unit last month. As you can see, my volume is quite low.

I suspect you're referring to a public statement I made regarding how many more Prius vehicles Toyota made... IIRC there were 2 million gen3 Prius units sold worldwide, which means for each G1 Honda Insight sold, Toyota sold ~QTY115 Gen3 Prius units. So then there is a much larger addressable Prius market, hence my sales would be higher selling a Prius product, compared to the G1 Insight. Econ101 stuff.

Further, I don't actually own a Gen3 Prius, hence I wouldn't be designing a passion product for myself, that I then offered to the community (as I had with the Honda Insight). The conclusion that I suspect you're implying here is the "mudder is only in this for the money" angle I keep hearing. Yawn. I don't see any shame in wanting to make a profit selling a well designed product. Capitalism 101.

Regarding the fact that Jack likely started NexCell as a passion project is commendable. I concede that I wouldn't enter the Toyota space for the same valiant reasons I entered the Honda Insight space. However, my proposed Toyota product (LiBSU) is essentially identical to my existing Honda project (LiBCM). Hence the passion I had for that project is built into any Toyota projects I move into. It's a nearly identical parallel space, and I am in fact passionate about releasing a well designed Toyota product. I'm passionate about nearly every project I take on; at this point I don't take on projects unless they're exciting. Yes, one goal is to make some money selling a good product, but I don't see how that discredits my motivations behind designing a safe BMS for the Prius.

Can you post a link to this thread. Very difficult to fully address your concerns regarding a thread that's not immediately available to read. I might have posted in the thread you reference, but I post a lot in a lot of places, so I'd need to see it to fully address your concerns.

You haven't quite come out and said it, but I suspect your implicit accusation is that I created my PriusChat membership solely to discredit Project Lithium? Am I describing what you're trying to say correctly? I will answer this accusation in the next paragraph.

Yes, I created my PriusChat membership initially to highlight my concerns regarding the Project Lithium product. I had previously expressed these concerns to Jack, and then when he disregarded them I posted a single youtube video outlining my concerns publicly. And after that I left it alone, because I was busy with other things and since nobody had publicly announced a failure yet, I figured "OK, I've said my peace... time to move on."

And then @sworzeh posted her My Project Lithium Battery Caught Fire Yesterday thread on reddit. Given my extensive knowledge on this subject, I decided to break my reddit vow of silence and wrote this response (my reddit username is 'redditmuder'). Several reddit members sent me messages stating that I should write more, but since I opt not to use reddit, I instead decided to post in @sworzeh's similarly-named thread here on priuschat. So yes, I created my priuschat handle specifically to address the safety issues I identified in the Project Lithium NexCell design.

Given that I don't own a Prius, I'm not sure what else I would focus on writing here on PriusChat.
Oh wait, there's my LiBSU project thread that I started on PriusChat.

Linsight is certainly taking up time, but I don't need you as a gatekeeper managing how I spend my time. I do what I want whenever I want. That's certainly not for you to decide.

Sure, yes, I found a severe design defect in the Powerwall 2 design. Yes, I properly reported this to Intertek, the CPSC, and Tesla. They're in the "we'll take it from here" mode, hence I don't really have any updates on it. Notably, since Intertek is Tesla's customer, they're not at liberty to discuss corrective actions Tesla has taken with me. If they ultimately decide a recall is required, I'll read about it on the same news sites you do. Not sure what else you want to know regarding this issue.

FYI: I doubt the issue I discovered would result in a Powerwall fire. Almost certainly unrelated. 'My' issue poses an electrocution hazard to the user, but (probably) wouldn't result in a fire.

If my above responses didn't answer your implied question here, you will need to ask more specific questions. You aren't going to hurt my feelings. I am almost always very open with my plans and actions.

 

Failing to provide a concrete accusation is item #1 in the Strawman section of the Logical Fallacies 101 Handbook.
Standing by for more ambiguity, special pleading, false causes, loaded questions, ad hominem attacks, anecdotal evidence, circular arguments, appeals to emotion, tu quoque, burden of proof attacks, etc... all key elements in a non-scientific rebuttal. Meanwhile I'll be ready at the keyboard to analytically respond to whatever drivel Team Jack drums up.

You get a gold star!

I missed this the first time around: @vvillovv, are you suggesting that I surreptitiously staged a Tesla Powerwall fire and then reported it to the media? If that's what you're implying, can you link to any post I made afterwards wherein I discussed the dangers I had outlined leading to said fire? If that's not what you're implying, please clarify your statement with a concise accusation I can actually defend.

Not sure why my motivations keep coming up, but I'm not hiding anything here. Yes, I'm a potential competitor, but that's not why I'm doing it. There were safety issues in the previous V1.x & V2.x designs. To help protect the community, I want to evaluate whether the V3 design is now safe.

I concede that if the V3 design is bad, then yes my poor review would probably affect Jack's business... However, if V3 is safe and reliable, my review stating that would certainly increase his sales... particularly because I wouldn't release a competing product if his already safely served the same purpose.

 

Here are some observations after sorting through 32140 Na-ion cell datasheets for an hour:
-Every cell I looked at in this form factor is nominally 10 Ah.
-Most cells have a recommended maximum 3C discharge rate (i.e. 30 amps assist max recommended).
-I found QTY2 cells that have an absolute maximum peak 15C discharge rate (i.e. 150 amps assist max). Note that you typically don't want to design a product that hits "absolute maximum" ratings... specs like this are basically saying "this will break immediately if you ever do it".
-Most cells I looked at have a recommended 2C charge rate (i.e. 20 amps regen max recommended).
-I found QTY1 cell that has a recommended 3C charge rate (i.e. 30 amps regen max recommended).

Given that the Prius routinely exceeds 30 amps assist and 30 amps regen, I have strong concerns regarding a V3 design that uses a 1P series string of 32140 cylindrical cells. Please correct me if I'm wrong, but my understanding is that the Prius can easily assist at 150 amps and regen at 75 amps. So then specifically I have the following concerns:
1) 150 A_assistPrius > 30 A_assistCell, and;
2) 75 A_regenPrius > 30 A_regenCell.

I don't know for certain that the V3 NexPower design is 1P, but based on the graphics shown in their video announcement today, it seems likely. For example, a 2P pack would have ~QTY132 total cells, which is about twice as many as are shown. Therefore, I suspect the V3 pack is 1P, which certainly causes concerns for long term cell reliability.

Rest assured, I will test these cells on the bench at the maximum assist and regen current they should expect to see in normal use. It's an abusive test, but every cell I've tested has survived it for months, except for the V1 NexCell cell I tested earlier this year.

@jacktheripper: If you'll let me know the cell model and overall pack configuration, I can make more definitive statements.

 

Sadly, @jacktheripper has asked me not to contact him again. I will honor his request, so it looks like I'll need to source the V3 hardware from a 3rd party.

So then I am offering the following bounties to the first community member that provides the following:
-$500 (plus your actual costs) for a 'non-GT V3' NexPower pack, and;
-$500 (plus your actual costs) for a 'GT V3' NexPower pack.

Please send a PM if you're interested. You can have whatever hardware I don't destroy back (for free) after I'm done testing. At my own expense, I will review these products for both safety and longevity. If they are good I will say so in my review. If they are really good, I won't see any reason to release LiBSU. I certainly can't compete at Jack's $1500/$1800 price points. Honestly I don't think Jack is charging enough.

 

I'm having a blast! You can follow along in any of the public LiBCM threads over on ic.net.

Over the past couple days we've had one customer who failed to address a critical safety warning that LiBCM continuously reported to the customer for over 15 months... during that time, customer never reached out for support. When they finally did, we were able to get the pack out of their car and safely discharged. I will be making some firmware changes to the code that will more safely handle the failure mode this customer experienced. Specifically, LiBCM will force the contactors open if the customer ignores initial soft warnings. Honestly they are very lucky a thermal event didn't occur.

Note that LiBCM is still in beta, and one of the terms of that agreement is that my beta testers are ultimately responsible for verifying all cells are in a safe operating area. LiBCM reports this data to you via the LCD screen, and there's even a permanent message printed below the screen that lists the max and min safe cell voltages. And if those are violated, the screen starts flashing and LiBCM starts playing a really annoying beeping sound, which in this case occurred every time my customer drove the car for fifteen months.

Even though this customer ignored these warnings for more than a year, LiBCM still (somehow) kept the pack from catching on fire. And now that I understand what happened, I can change the firmware to better handle this failure mode. You can't say the same for NexCell's V2.8 and previous hardware, which lacks the capability to tell the customer when a single cell has failed. This is probably also true for the new V3 hardware, but I can't say that for sure until I review it.

I'm glad my customer owned his mistake, is generally in good spirits, and has decided to continue with the project. I even offered him a full refund, which he politely declined. All of this is publicly viewable in the LiBCM Open Beta Support thread. Note that I don't offer private beta support; everything I do is out in the open, so that the community can evaluate my product from their armchairs.

What did I say that makes you think otherwise?

Are you referring to the fact that I wouldn't choose to develop a Toyota product that I wouldn't make money on? LiBSU isn't a passion project, so if it's not lucrative I don't see the point. If it's clear I won't generate profit, I'll just take on another passion project. I'm confused why people fail to grasp the simple concept that LiBSU would exist to generate a profit.

 

I found some information about the Sehol E10X, which is an electric vehicle sold in China that is offered with a sodium-ion variant. This car uses a 340 volt 7P 32140 sodium ion pack, which is rated to 72 Ah. The peak assist current is 45 kW, which means the maximum pack discharge rate is 1.8C (45 kW / 340 V / 72 Ah). Therefore, the Sehol E10X is probably using 32140 cells within their maximum recommended discharge rate (3C).

On the other hand, based on what I've seen, I suspect the NexPower V3 pack's maximum discharge rate is 15C (150 A / 10 Ah), which probably exceeds the 32140 cell's maximum recommended discharge rate (3C) by a factor of five.

...

To use 32140 cells within their recommended maximum discharge specification, the V3 pack would need to have five cells in parallel (e.g. 66S5P), whereas I suspect they're using something closer to 66S1P.

What does this mean if my assumptions are true? It strongly suggests that people who drive 'non-GT V3' packs hard will experience accelerated cell failure. All the more reason to ditch the cylindrical design and only sell the pouch variant... of course, I still need to analyze the pouch variant, too... but for now I'll take Jack's word that it performs better than the cylindrical variant.

 

FYI: NexPower is hosting a pre-recorded Q&A tomorrow to answer the following questions:
00:25 - Difference between V3 and V3 GT
04:24 - Why 0v tolerance is important?
07:12 - How do you make sure V3 last a long time?
11:11 - Who is your sodium-ion cell supplier?
13:05 - V3 installation
14:47 - What is the business impact from the 2024 tariff against import battery?
15:27 - What is your next project?

 

It's really none of your business, but here's a small part of my response to Jack's request that I "do not reach out to [him] again":

In case it's unclear, that means I've left the door open for Jack to contact me whenever he wants. However, I won't directly contact him (e.g. via phone/text/email/private message/etc).

Note that I didn't quote any other portion of our interaction. Therefore, you are missing some of the background information regarding why Jack doesn't want me contacting him again. I am not interested in discussing this background publicly, particularly with you. You are of course welcome to solicit this private conversation from Jack, but until you understand the details you really have no business doing whatever it is you're doing right now.

In another thread you told me to stay organized.
I recommend you do the same, as I suspect you meant to post this in 'Jack's' thread instead of mine.

Regardless, you are gatekeeping hard, which is laughable because you're an unaffiliated, uninformed 3rd party to this matter.

I hope you are able to understand that I intend to continue having on-topic discussions anywhere I so choose on this public forum. Team Jack really likes gatekeeping!

 

With the same caveats I outlined in post#1, here are my thoughts on Jack's latest "NexPower V3 Q&A" video:

Jack's explanation on why the non-GT V3 product exists (and not just the GT version):
-installation time: 5 minutes vs 30-60 minutes; better for installers
-higher internal resistance, but better cooling (due to V-shaped air channel).
-has higher capacity than GT version. Jack mentions the GT cell is 7 Ah, versus the non-GT is 10 Ah.

FYI: Since I know little about the Prius' OEM NiMH cell characteristics, I concede to anyone who provides different data than what I've reproduced next.
Based on my web search, a new OEM NiMH blade is typically around 19 mOhm (at unknown discharge current).
Based on the test data visible in Jack's first video @1:30, each Na-ion cell is 3.8 mOhm (at 70 amp discharge).
Since each V3 "blade-equivalent" (V3 doesn't use blades, per se) is probably QTY5 cells in series, the comparable figure is also 19 mOhm. So then Jack's statement might be true. I'd need more data to definitively confirm or refute this statement. Probable the same ballpark.

Once I get my V3 unit, I will gather concrete data and then evaluate these claims.

I don't have any feedback regarding NexPower's design.
Yes, sodium-ion is vastly superior to any lithium chemistry in regards to dropping to 0 volts.

Wow, I'm surprised to learn that Toyota allows the pack to drain to empty when the car is in neutral. Bad design.

I'm going to give Jack the benefit of the doubt and assume he complied with the measures outlined in:
49 CFR § 173.189(b) & (c) & (d)
49 CFR § 175.10
I suspect that eventually sodium transportation laws will catch up to the much more rigorous existing lithium laws, or even just lead acid laws. Right now the technology is too new to have been properly regulated. Overall, I do agree that it is technically safe to transport sodium ion cells at 0 volts, but that would also technically apply to lithium cells discharged to 0 volts (even though this will destroy a lithium cell, it's still illegal to transport large lithium cells in this safe state).

I suppose this is as close as an admission as we're going to get regarding the V2.x circuit board fiasco.

Notably absent from Jack so far is how they've managed to fix the circuit board issue. I certainly hope V3 electronics have a proper BMS that can disable assist and regen if any single cell voltage gets too high. Jack not featuring that as a banner spec makes me skeptical it exists... we'll know for sure once I get the V3 hardware.

Great to see that Jack agrees the lithium cells in the V2.x product were ill-suited for use in a hybrid traction battery.
If you were able to break your old product with those tests in just two months, imagine how much faster they would have failed with a proper lab test gauntlet in a thermal chamber.

7 months ago was 2023DEC06, which is a date I can believe regarding when Jack started testing sodium ion cells. This agrees with my previous "sodium was a twinkle in Jack's eye in 2023" comment.

I'd love to see your test data.
I'll show you mine once I get the V3 pack.

What are CATL's C charge/discharge ratings at either extrema?
Does your BMS enforce these lower limits when the pack is very hot or cold? Or are you just hoping the OEM system limits assist and regen using the NiMH-specific temperature rules?

Jack is strongly implying that they are using a unique, proprietary cell chemistry. I find this hard to believe, because NexPower is very low volume compared to EV manufacturers, who might purchase several hundred million cells per quarter. Why would CATL develop a better sodium product and then only offer it to Jack? That makes zero sense unless Jack owns the patent.

I'm not contesting Jack's statement that they had a manufacturer create a custom cell specific for NexPower's use case. Yes, this is expensive, but it happens all the time; a customer wants a unique property with an existing cell chemistry, and pays a ton to get it.

I agree: CATL is a great company. However, your wording makes it sound like you might not actually be using them in production V3 units. The other companies are lesser well known, but fine. I'm still confused why you would actually source cells in production quantity from half a dozen different companies. Did you just engage with all these companies, or are they all actually going to provide production volume going forward? Seems like having just two good companies would be a better use of time, and still allow for competitive price reduction over time. Maybe you have one or two manufacturer(s) doing your cylindrical cells and another couple doing your pouches?

Here's a much better view of the pack. Thanks for clarifying that the graphic shown in the previous video was only half the pack.
Based on this image, looks like the V3 pack in total consists of maybe QTY140 cells (~14*5*2 cells). In that case, V3 is probably a 70S2P, which would improve its performance under load (compared to my previous analysis when I thought the module shown was the entire pack, whereas it was only half). Probably still abusing the cells under heavy load, but not as much as I had calculated before. Since the "cell is proprietary", I'll have to wait until I test it on the bench to see how well it holds up to my version of the 'postman' test.