Rainbow Prius Hybrid Powered by Lithium!

I could be wrong, but I thought each module equaled 2.5 LFP cells. In other words, 5 cells per block of two modules.

That would make keeping the isolation fuse thing not part of the circuit and require a workaround... if what I just said was correct of course.

 

If you have spare NiMh batteries, check out the thread here Re-hydrating the battery modules. | PriusChat and see if you can get them all back up to capacity. Then you could build 4 batteries, each 4 modules in parallel and 7 groups in series (4P7S) so the 4 batteries in series ads up to 28 modules. If I can gather enough NiMh modules, this is a project I want to try in the near future .... but that will require a thread on its own.

As far as I can make out, the low end cut off for the NiMh module is 6v but measured as a pr by the Prius system, so roughly 12v between the two sense wires. The LFP cells will handle 2.5v without long term damage, but 2.8v is a safer low voltage cut, so 5.6v for the 2 x LFP in series and 11.2v across the 4 x LFP cells in series that would be paralleled to the 2 x NiMh modules.
The high cut voltage appears to be variable, but also very short term. The LFP cells will handle 4v short term, 3.6v is the safe upper limit. The cable linking the NiMh module to the 2 x LFP cells in series will limit the current flow and the differential between the module voltage and the 2S LFP cells limits the amount of "push" (for want of a better word) to force the current down the cable.
If LFP cells are charged too fast, the voltage climbs rapidly as they reach the upper end of their state of charge, so any attempt to push the cell above 4v will effectively act as blocker causing the charging current is reduced to close to zero ( very little differential between the NiMh module voltage and the LFP cells voltage) and until the LFP cell can actually converter the charge current into stored energy, this voltage will remain high.
The rapid voltage rise is the result of the lithium ions released from the LFP plate reaching saturation point because the graphite plate has just about reached saturation so the lithium ions take much longer to fine an empty spot to attach. This means a sudden spike in the NiMh module voltage will not result in over charging of the LFP cells because of the limited current capacity of the jumper wire. The LFP cells would see this a a pulse charge and they can handle up to 20CA when pulse charging.

Even though the LFP battery is direct linked to the regen in parallel to the NiMh pack, the rapid voltagerise in the LFP battery will block the charging current while the NiMh battery can accept the high current for the short term.

The unknown is the affect the greater storage capacity will have on the upper charging voltage, it is possible the peak voltage will not be reached because the Prius system monitoring the voltage has more time to react and reduce the regen charge current and that way avoiding the voltage spikes.

On the discharge side, the LFP cells I linked to previously can handle a constant 30 amps discharge rate, so three cells in parallel can discharge 90 amp continuous. They also have a peak discharge rate of 6CA, or 42 amps per cell = 126 amps, add the capability of the NiMh battery that can deliver this amount of current and more, the available current far exceeds the likely maximum current demanded by the Prius inverter, that big torque burst you get when the traction battery is fully charged could be even higher because of the much reduced voltage drop, and maintained for longer because the added capacity.

So, how many readers are now sitting in front of their computer with their eyes rolled right back in their head :lol:

T1 Terry

 

The 2.5 LFP cells per module is only if you plan to use the LFP cells in place of the NiMh battery, because that means there is no where else for the high regen current to so the LFP cells need the added voltage "head room" to be able to handle those voltage spikes where the Prius computer doesn't cut the regen current back fast enough.

T1 Terry

 

I don't think it works like that nor are just voltage spikes the problem. The problem the actual charging voltage. NiMH cells are usually said to have a "full" voltage of 1.4V, but rather can be pushed up to as much as 1.54V. That's not a voltage spike, that's the charging voltage. The Prius will sense the pack can take on the current and will push it in at a higher voltage like this. If it tried to charge at 1.4V it would take forever because of the internal resistance. Only once the Prius BMS senses the battery is filling up does it start cutting back voltage so the NiMH battery doesn't charge up to more than some 75-80%.

It does this by comparing the current and voltage. As the NiMH battery fills it will reach a point at around 75% when the resistance noticeably starts to increase. Increasing resistance requires more voltage to push the same amount of current, otherwise the current starts to drop. At that point the Prius realizes the battery is getting there and starts cutting back on voltage.

With an LFP I understand that things are different since they don't have the internal resistance that NiMH batteries have. Therefore they should be charged at a constant voltage of no more than 4V. To the Prius it's basically sensing the same thing, waiting for the current to start slowing by itself. The only difference is that with the NiMH batteries this is more of a function of resistance than resting voltage. With the LFP batteries it's more of a function of reaching the resting voltage. That's why people are finding that LFP batteries tend to fill up more, using closer to 100% capacity rather than about 40 to 50% like with the NiMH batteries. (This is also why I've said I don't think pairing another type of battery or ultra capacitor bank in parallel to the existing NiMH battery is a good idea since the Prius wants to force a high voltage into the NiMH battery and then let it fall back to a lower resting voltages whereas a LFP or ultra capacitor bank would end up at a higher resting voltage and then continue to feed current into the NiMH battery after the Prius decided the NiMH didn't need any more.)

1.54V x 6 cells per module = 9.24V per block under regen. When divided into 2 cells that's 4.62V! 2.5 LFP's, or 5 LFP's per block would be a much more comfortable 3.7V at an absolute maximum. Again, this is with or without a NiMH secondary battery. Unless you have an electronic means of controlling the current and voltage between the two batteries, just hooking them up in parallel isn't a good idea.

As far as voltage spikes are concerned I'm pretty sure it's not as big of a problem as people think. Voltage spikes are going to happen. That's the nature of starting and stopping large amounts of current. A battery is going to have a small amount of capacitance and is going to be perfectly capable of tolerating momentary voltage spikes up to a certain amount. And if voltage spikes are a problem, then you usually fix that using capacitors to assorb the spikes and then dissipate them back into the battery more slowly. But regardless, I don't think voltage spikes should be that much of a concern with batteries. Voltage spikes kill electronics, not batteries.

 

I've read through all three configuration scenarios a few times, and the explanations of the 3p9s +3p38s configuration at least 5 times since I first saw it around midnight local. While refreshing just prior to posting, I finally noticed what I thought might be a mistake
<

>
was actually my mistaken reading of 2S LFP cells as 25 LFP cells the first 5 time I read through this part of the post ... yes :lol: probably quite a few heads including mine!

I've been pondering as much as reading maybe even more pondering than reading, and each time I come back and reread, I gather something new with each reread. Thank heavens I finally corrected my misunderstanding of 25 LFP cells to what you actually wrote there, before I posted. including shouldn't that 25 LFP cells be 28 LFP cells?

The concept of using the multipule fuses, and the f links on each module interconnect as a poor mans BMS (so to speak) is elegant and fairly simple when fully understood. I'm not there yet, although much closer than I was before reading and rereading. I've put the main contactor connections on the back burner to come back to after if I can find info on the fuse and link types mentioned for the module interconnects. On that note, I believe I can get up to speed on the interconnect components as to my understanding at present is they are some of the most basic building blocks in electrical engineering. Though, to actually test the current flow I'd need to understand how to use more complex electrical components (that are mostly well understood by most any electrical engineer, like resistors, mosfits ( I mean mosfets ) , hall effect sensor, shunts, etc. which I don't have yet and struggle with to understand how most of the BMS projects I've seen progress through beta testing.

And that is one reason, I'm so interested in and the current success of @jacktheripper lithium replacement project and your OEM Nimh pack with Lithium enhancements.

I also remembered after posting previously about the other manufacturers high side low side pack split, that Jacks project videos also show the the Prius pack split which I was not aware of before seeing how Jack handles the pack split for his project, which made it much easier for me to understand your 3p9s + 3p38s configuration, even if it took a few rereading though to put the pieces together for the first time.

 

@Isaac Zachary isn't what your posted above the reason why when replacing an OEM Nimh pack with a LFP pack, that a customized BMS is a requirement? I thought that was clearly stated in reference to that configuration being discussed. Regardless of how the 3p9s +3p38s LFP booster pack might function when interconnected to the OEM Nimh pack. As to the best of my understanding at present, it's not been tested yet.

Kinda beyond the discussion, is the behavior of prius bms when the packs get older and weaker long after setting DTC's concerning the pack. And the different voltages reached on the high side when grid charging and probably also during regen.
The discharge side of the equation is noticeable even to inexperienced prius owners with weak packs, even if they still don't understand why the discharge capabilities of the car have changed so much since the days when the packs were healthier.

 

I am by means no expert, but when comparing maximum voltages and currents the Prius BMS shouldn't hurt an LFP battery wired directy to each BMS lead, so long as you have 5 cells per block.

However, the Prius will not balance the cells. NiMH cells balance automatically since the more they're charged the more they self discharge.

So from what I understand you only need a way to keep the LFP cells balanced. There should be no need for limiting current, maximum voltage or minimum voltage as the Prius's own BMS will do that within the limits of the LFP cells, again, as long as there are 5 cells per block and the LFP cells are rated for the kind of current a Prius (or other Toyota hybrid) can deal out.

 

Ok, I get what your thinking concerning the buffer region for the LFP pack replacement. The self discharge of old weak packs can be the one of the issues that frustrates many if not all new module balancing / pack grid charging owners. It's not self evident to the inexperienced, just how important a part of the pack health is actually is.

 

There is a reason you need to reprogram the ECU(?), or one of the charging computers so it know WHICH
type of battery it is charging. If it is programed to charge the NiMH batteries, it will NOT charge any other
type of battery correctly.

 

I don't think this is true. Besides, how would you even do that?

The question is does the charging characteristics fall within the tolerable limits of a different type of battery?...

Prius will charge and discharge up to 150 amps...
Prius will stop charging at 258.1V max... (rarely gets that high)
Prius will stop discharging at 174.7V min... (rarely gets that low)
Is there a battery that fits that besides NiMH?

LiFePO4 can do 150A... Actually LiFePO4 with correct cells can do way more...
LiFePO4 can 3.7V × 70 = 259V... Actually LiFePO can do up to 4.0V, although less than 3.8V is better.
LiFePO4 can 2.48V × 70 = 173.6V... Well, some say 2.5V and others 2.0V, although 2.48V is pretty close to 2.5V. Again, the Prius won't normally go that far every day.

Can the LiFePO4 cells fall into the 14 blocks on the Prius so that you don't need the NiMH battery to keep the Prius's BMS happy?
70 ÷ 14 = 5, so yes!

 

Don't think, feel!

I forget who it was that also works with batteries that had said that. I also did a search on it and found info on it.
And I also knew when the Ni came out after the nickle cad battery had been out that there are different
battery chargers because of the type of material and how they charged and the rate, both voltage and amperage.

I can't explain all the details about it, too much info that for me, is not important. Maybe for some people.
So, the computer, which one actually, I do not know, NEEDS to know WHICH batteries it is charging so it can
do it correctly.

I personally would not take the risk of putting in a bunch of batteries that my system didn't know about.
Because I would be the one to have the battery explode and/or catch on fire at the worst possible time!!!

The Prolong worked very well for me! So I'm not expecting having to replace the battery any time soon.
But If I have to, then I will put in the same type that are in there so there will be no problem with the car
charging and discharging the battery.

 

Charging profiles are for the most part important when charging to 100%. But neither the NiMH nor the LiFePO4's will normally be charged to 100% on a Prius.

And even when two battery chemistries are charged to 100%, that doesn't mean that they absolutely have to have different charging profiles and voltages. LiFePO4's work as a drop in replacement for SLA's because the charging characteristics are the same at "12V".

IIRC, if you want to charge NiMH to 100% you first start out with constant current, then move to constant voltage once you hit a certain voltage, and when the current stops decreasing and starts to increase you stop charging. But on the Prius it never even gets to the constant voltage part of the charge profile. It just rolls at a high current, then hits a voltage and says "I'm done!" and slows down and soon after stops charging, leaving a NiMH battery at around 75% to 80% charged and a LiFePO4 battery at around 85% to 95% charged.

If it were going for 100% charge for the NiMH's, then on a LiFePO4, ya, that would be a disaster as it would be searching for that part where the voltage wants to drop or current wants to increase and would never make it there until the battery was fried.

 

"I don't think this is true."

It doesn't matter if you fully charge it or not, they receive a charge differently.
It doesn't really matter. Do as you wish with your car...

 

That is why my wife writes the info on our website "T1lithium.com.au" I answer her question a number of times until she gets what I'm saying, then writes it a plain language and then hands it back to me to read to make sure she understood it :lol:
I also have a private lithium forum set up over 10yrs ago to assist the first 125 people who wanted to do their own lithium conversion in the motorhome or caravan. You think reading through that post was a mammoth undertaking, all 125 people on the forum said they read it all multiple times, gaining a bit more each time using sone part of it that they now understood but didn't before .... ever one of them came to the "penny dropped" stage, then reread the whole lot again before starting their install. I must have got it right, not one of them has had a battery failure over the 10 yr period and all still have 100% plus capacity doing the 0.5CA test (discharge over 2 hrs to 0% SOC and the cells remaining above 3v while under load) I did destroy over
$10,000 worth of my own lithium cells doing the trial and error testing to find out just what the cells can handle and what is outside their capability resulting in catastrophic failure.
Was it worth it ... you bet, every thing you learn can not be measured in $$ and cents, every bit of knowledge you can share is priceless .... I'll never be rich in $$ terms, but I'm happy with the value I feel I have gained and can share with others.
These forums all run on this shared learning thing and that is what makes them so good

T1 Terry

 

Does it use some sort of different octane rated electricity or something? How does one battery receive a charge differently than another?

You can't fry things electrically that are within safe limits. People get way more out of their LiFePO4 replacement 12V batteries than 12V SLA's, even when charging them on the SLA profile. There seems to be enough evidence that the charging profile for a NiMH Toyota hybird will work for LiFePO4's, and several people are doing it and I haven't heard of any mishaps yet. Have you?

I'm not saying that charging profiles aren't important. Don't charge SLA's as you would a non-sealed LA battery, especially when it comes to trying to keep it desulfated. NiMH's voltage starts to dip when full, unlike Ni-Cads and most other batteries, so don't charge NiMH's on a Ni-Cad charger or viceversa, unless the charger is rated for both. Lithium-ion-anything has a maximum voltage and you don't want to constant current them above that voltage, but rather constant voltage, unlike others batteries than can be trickle charged at a constant current indefinitely.

In theory charging LiFePO4's on a NiMH would spell disaster. But Toyota chose a charging profile that just so happens to meet both and it seems to work plenty fine, and lots of people seem to be proving this.

 

Hi Isaac,
All these posts are really excellent and very helpful to me. Thank you all.

I'm just scratching the surface of all this. A couple of my Priuses have multiple problems so I've rigged up very crude grid chargers using Mean Well 350mA LED power supplies, and inline old household light switches (!!! hahaha). What I have noticed is if I leave the grid charger on overnight and/or for extended periods the overall voltage of the (probably bad) HV bat may start off low - around 180-205V - and then gets up to around 235V after a relatively short while - then goes down to around 228V after a longer period. Like I said, my system is very crude, without any diodes, and doing this in Hawaii there is a lot of ambient heat so I try to do it mostly at night. I do make sure to switch off the connection so as not to "backfeed" the charger. Do you think it's reasonable to conclude that the HV bat voltage dropping off a little is indicative of it reaching something approaching full charge? (Your explanation of NiMH charging is the first I have see regarding voltage dip at near full charge). I'm just curious. I don't expect miracles, or to "revitalize" my HV bats but I'm interested in learning as much as I can about what I'm doing. My impression are that these HV bats and the cells in them can tolerate a lot more abuse than we might think (without blowing up ;-). Thanks again all!

 

That could be. When you look at any graph depicting the charging curve of a NiMH cell, you'll see that at a constant current, the voltage increases until the point of full but then will decrease after that. So it's not a good idea to let them decrease after that.



On the other hand, I'm not completely sure what's happening with your battery without being there. 235V shouldn't be high enough to reach full capacity in all the cells. Maybe there's an imbalance or even a dead cell. When I've charged and discharged mine I took off the cover and measured each module individually.

But still, your findings do seem to coincide with the charging characteristics of NiMH cells overall, even if there is a bad cell or an imbalance. I would recommend, at any rate, to measure the charging voltage often and when it does start to drop slightly then stop charging.

 

Whatever......... in theory, anything can work.

 

And rules of thumb aren't the gold standard of possibilities.

 

Waaaaiit ... isn't the whole point of useful theory that it lays down the line between what can and can't work?

I mean, Pauli was able to predict neutrinos in 1930, even though it was 1957 before anybody had the means to spot one ... because without them, theory showed that β decay couldn't work.