Lithium replacement pack nearing market. Sounds great

Yea, buy new old stock or some other form of old tech in a new shape and color.

Or....wait a bit and you will have new tech that is lighter and will save even more MPG.

https://plugoutpower.com/

Hybrid battery diagnostic and repair tool for Toyota and Lexus



Home - Pulstar Spark Plugs

 

Studying the charging characteristics and block voltages, it seems that NiMH cells can be directly replaced with LiFePO4 cells without any modification to the stock system. The things needed would be a battery warmer and a balancer. You won't get the full rated capacity of the cells without modification to the Toyota system, but you don't with NiMH cells either. And in the end, using less capacity means better longevity.

Here's one guy who did the conversion himself to LiFePO4 and it was simple and worked. But he hasn't reported back since doing it.
Lil Asian Grocery LLC - YouTube

Personally I'd be interested in increasing capacity rather than reducing weight. This is because of the mountainous terrain I live in. The OEM NiMH batteries fill up quickly, meaning there's no real regenerative braking benefit for mountainous areas.

For my kind of terrain a PIP or Prime would be better suited. Or maybe a plug-in Pacifica.

But in the end, I don't have the money to buy another vehicle right now, nor an expensive experiment. And I know what to expect from an OEM battery. Toaster's also look pretty good. So...

 

I would guess that replacing an in series group of NiMH cells with an in series group of LiFePO4 cells (with number set to match the same total voltage) would not work well and would quite possibly go up in flames. This is however much better than the traditional Lithium ion batteries which absolutely would go up in flames in short order if the cells were just connected in series.

At first the LiFePO4 cells would most likely charge up and discharge OK, especially if they never hit any of their limits (high or low), but in a car they will hit those limits and then the cells would get out of balance. The logic the Prius would use to try to balance things would not be appropriate for the new chemistry. To work properly the new chemistry would need a cell level charging controller. See item 3 (self balance) here:

How to charge Lithium Iron Phosphate lithium ion battery packs including packs with high current and High Capacity.

So I'm not saying that a battery based on this technology is not possible for a Prius. However, it is going to need its own battery controller, and it would probably have to synthesize a set of fake voltages to feed back to the car, both so that the Prius won't throw a code, and also as a way to control the car so that it charges and discharges in the manner that controller requires.

 

Correct. And that's only one aspect of why this is fire material.

Either you keep the OE ECU and add an inbetween translator and hardware BMS or you create a new BMS that emulates the old ECU and is plug and play.

Both approaches have been tried and succeeded for the Gen2 and Gen3 for commercialising lithium packs a decade ago.

The pack under discussion does not have the required hardware.

 

We have done winter testing and the pack is working great at temps below freezing, but I've directly only tested a small amount below freezing, not super extreme freezing, but perhaps some of the other people doing winter testing have taken the extreme further? @jacktheripper would be the best person to ask. He'll likely see this comment and reply to it sometime this week.

 

Unlike Li-ION batteries, LiFePO4's won't go up in flames even if shorted out, severely overcharged, punctured or crushed. So flames wouldn't be any more of a problem then they are with NiMH batteries.

Looking at NiMH charging and discharging characteristics as determined by the Toyota Hybrid system, the maximum and minimum block voltages seem to stay within 14.2 to 18.4V. Divide that into 5 LiPoFE4 cells and you get between 2.8 to 3.7V per cell. This is within the limits of a LiFePO4 cell. And even more so seeing how these can tolerate even more voltage (up to around 4V) and 2.8V is usually the lowest resting voltage (some even specify 2.5V). So under acceleration discharge, 2.8V would still be well above the lowest possible voltage under that kind of amperage because once discharge current is cut off the resting voltage will be higher.

In other words, as long as the cells are big enough for the amperage and the battery is kept balanced by a separate battery management system, there's no technical reason why you couldn't hook up blocks made up of 5 LiFePO4 cells for a total of 70 cells. In fact, I'd be willing to bet that as long as the cells are occasionally bottom balanced, there may not even be any need for a balancing circuit. Just hook up 5 per block and go.

The point is that with 5 LiFePO4 cells per block the voltage and amperage limits fall within the same voltages and amperages that blocks of 12 NiMH cells do making it rather impossible to exceed voltage and current limits. If you don't go out of limit, there's no reason it wouldn't work.

Man! You're tempting me!

I do know that some brands, such as the Winston batteries, are rated to -40°C (or -40°F). But I'm skeptical of that. Still, a trunk full of 40ah Winstons would make for a rather generous capacity gain. And adding a battery warmer to all the cells wouldn't be too difficult.

 

What chemistry was involved in #14 ?

 

NiMH are balanced from above and Lithium are balanced from below. Contrary to @2k1Toaster 's claim when you're running a pack wired in series the OEM's BMS ability to balance a NiMH pack doesn't exist unless you go up to max charge and the Prius BMS is designed to protect the pack from being charged up that high so it can never do a legitimate balancing on its own in the first place.

So anyone who claims the BMS for Prius pack balancing is not compatible with a Lithium pack is full of it because a Prius requires an aftermarket system like Prolong or Maxx Volts to do legit balancing.

As for how to keep the Lithium modules balanced over time, the developer is several years into this effort and those original modules still aren't in need of balancing. And will that change after 10 years? Perhaps... But at that point it's safe to assume that their will be an aftermarket system that will balance these modules from below. Hopefully @jacktheripper will see this post and further elaborate on concerns with module balancing.

 

My Gen 1 definitely ran occasional cycles up to max charge, which I watched on the ScanGauge at least three different times. One of the early times I got spooked and thought maybe it was a bug or the BMS was locked up in a particular charge mode and I interrupted it. A later time I was off on a leisurely drive and I just watched it through each different stage of the process. I think I posted about it that time too, but I'm not finding that post just now.

I don't know that later generations still have that behavior programmed in, but I don't know that they don't, either.

 

Lipo stands for Lithium Polymer. It doesn't state what chemistry is used. But seeing how LiFePO4 (lithium ferrous phosphate polymer) batteries have a nominal voltage of 3.2 and many RC LiPO's have a nominal voltage of 3.7 it's obvious that they are different chemistries.

Exactly. A true battery balancer has a lead to each cell's positive and negative electrode so as to be able to directly make all the cells exactly the same voltage. NiMH cells tend to self balance, as do lead acid, due to having a higher self-discharge rate at higher levels of charge. Slowly increasing the state of charge up to near 100% allows the cells to self balance even better. But it's kind of hard to directly change an individual NiMH cell's voltage since they are sealed up in modules of 6 on the cars we're referring to.

From what others have told me, LiFePO4's can be bottom balanced and used that way for a long time since their self-discharge rate is negligible at any state of charge and at the same time have a generous upper limit. So even if one cell has a smaller capacity than the rest of the cells it will tend to discharge back down to the same voltage as the rest even though it will charge up to a higher voltage. For an example, start out with 2.8V/2.8V/2.8V, if the cell in the middle has a little less capacity they might charge up like this: 3.6V/3.9V/3.6V. But once you discharge them they'll all be at 2.8V/2.8V/2.8V again. And since they're all between the 2.5V and 4.0V limits nothing is damaged.

 

I forgot to add there are a lot of videos of LiFePO4 batteries being purposely mistreated in many different ways in order to demonstrate their safety. At worst they do bloat and off gas a little. But not enough to cause any flames. NiMH would do the same in the same situation. Even just charging NiMH modules without them being squeezed together in the battery box will let them bloat. (Ask me how I know).

 

Yes, that's precisely why Jack spent so much money on as many as 10 final prototypes for winter testing after several years of extreme hot weather testing in Southern California. If you have any torture testing ideas we've not yet considered let us know...

 

Out of curiosity, how many people contributing in this thread are battery engineers/developers?

 

Everyone, because... Google.

 

I won't believe that unless I see it!

Oh, right.

 

It's a website dedicated to DIY, not engineering & design... aka: social media

We're just hanging out talking about stuff.

And if you want to put the "you're not qualified enough to talk about this" turd in our punch bowl go ahead... We've all drinken more than enough Kool aid and probably don't need to drink anymore if you float a new bad flavor into the mix.

 

After looking into this a bit more, it does appear that a battery pack consisting of serial LiFePO4 cells should stay in balance for a reasonably long time - so long as it is kept within strict ranges for the voltages. It isn't yet clear (to me) if that is still true when car level (ie, very high) currents are being moved. Nevertheless, sooner or later, the pack will drift out of balance. The car cannot fix this. NiMH cells can be safely overcharged at low currents, so the low voltage cells in a serial pack can be brought up without destroying the fully charged ones. If the Prius has any ability to rebalance the pack it would be by using some variant on that mechanism. As I understand it LiFePO4 cells cannot be overcharged because they pretty much stop conducting once they are 100% charged. So whatever the car can do would not work here.

What might work though would be to keep a couple of fully charged extra cells in the pack, and have line selectors to attach these one at a time to cells in the pack. If a Prius pack is ~202V and a LiFePO4 cell is 3.3V then it would take 62 cells. Imagine a wire running to the negative electrode of the "bottom" cell and a wire running to the positive electrode of every cell (63 wires), and two big mechanical selection wheels on the same axle which can be used to select any one cell for charging from the extras. Of course this would be insanely dangerous because there is so much potential between the top and bottom of the pack. So it would be better to break the pack up into let's say 8 regions of 8 (or 7) cells, each with its own extra cell for balancing purposes. When the car was off this sort of replacement battery could scan through the cells in each cell, and the whole pack, and decide which ones to charge, and by how much. This doesn't sound beyond the realm of possibility. Would it use mechanical selector wheels? Probably not, probably it would use MosFets, like the ones in solid state relays (turned on optically, to keep the grounds separate).

Now then, does the balancer really need to be built into the car? That is the $100 question. If a LiFePO4 pack remains balanced for a year or so then it would be perfectly fine to have a big plug on the case into which one plugged an external balancer. If on the other hand it would only stay balanced for a week, or even a month, it would need to be built into the pack.

So technically, I don't see anything insurmountable. Economically I'm not sure that anybody can build a pack like this which is price competitive with a NiMH pack from Toyota. Then again, it is a lot more likely that the price of LiFePO4 cells will fall than that NiMH prices will fall. Bottom line, I think this approach can probably work.

 

Wow, that was a wild ride! Seems wiring a pack up partially in parallel is what you're getting at here... I think Tesla has done a little of this to get rid of a need for separate 12v, but never heard of it for the purposes of balancing. Still have yet to hear of a lithium powered car failing due to lack of balancing, but it's for sure a problem with NiMH due to self discharge problems that get worse over time.

 

That general idea could technically work and you have some good observations.

But I do believe that 62 LiFePO4 cells wouldn't work, you need 70. For one, the maximum voltage we expect to get from a Prius is nearly 258V, which divided into 62 perfectly balanced cells would be 4.16V. That's a bit too high, especially for LiFePO4's. (Like you said, when LiFePO's reach 100% they stop taking current). But divide that maximum battery voltage into 70 cells you get a little less than 3.7V, which should be plenty safe. 62 cells may be ok for some other Li-ION batteries though.

Also with 70 cells you get to divide those into 14 blocks of 5 cells. This means you can hook them up to your Prius' battery management system and make it happy without this simply being and addition to the NiMH battery or without some sort of circuit that makes up pretend voltages. It also means the Prius will throw a code if one of your LiFePO cells fails, which is desirable IMO.

 

No argument from me, it was just a ball park number.