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?

 

<Reserved for future use>

 

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.