[Jimbo69ny]

So this is a new problem to me. I am adding an additional battery pack to my EV and I didnt think it would be a problem to run two batteries in parallel. After all, they both use 18650 cells...

In a nutshell, I have a 2014 Rav4 EV. The OEM pack is made by Tesla and uses 2600 mah cells in a 92s48p configuration. I want to add a custom made 18650 pack with 3400 mah cells in a 92s31p configuration

Well, someone brought it to my attention and for the entire day I have been scouring the internet for a description of the problem AND a SOLUTION. After hours and hours of reading "it cant be done", "you're going to set your house, car, family, on fire" etc etc. Not only am I no closer to a solution, I am not even convinced that most of the people worried about it have even identified the correct problem.....

From what I can tell, having cells with different mah ratings isn't even the problem, its the resistance of the newer cells I am adding compared to the older cells. In most of the threads I have read, most commenters dont even mention resistance. Electricity follows the path of least RESISTANCE. So the pack with the higher resistance will feed less amps to the car. The pack with the lower resistance will be delivering more amps because its the path of least resistance. The pack with the higher resistance will most likely feed power to the battery with lower resistance in order to maintain the same voltage. Is any of that really even a problem?

If both packs have a BMS, the BMS will keep them both in line. They both will be at the same voltage and charging will stop once max voltage is reach. Again, whats the problem?

I always tell my employees to "bring me solutions, not problems".
So if it actually is a problem, how is this problem remedied? After reading threads for 10 hours today, I haven't seen a solution.
1. "Re-balance all of the cells in all of the packs so they both have equal resistance."
That's not possible for me because my OEM is a Tesla pack, there is no way of taking those cells out.
2. "Use a battery isolator?"
I have heard this being suggested a couple times but I haven't seen an isolator capable of handling 386 volts and 325 amps. Do they even make the that heavy duty? Has anyone done it?
3. "Use a DC to DC converter to charge the OEM pack off the secondary pack."
Even if the part does exist, I dont like this option. More losses and more parts to fail.
4. "Use a diode on both packs so power only goes out".
Ok maybe, but then you need two chargers to charge both pack separately. I didnt think diodes that large existed but I found this. https://www.ebay.com/itm/300A-NEGATIVE-STUD-MOUNT-DIODE-300AMP-1000V-ANODE-CASE-RECTIFIER-ECG6359-/121863675010
So maybe...

That might be about all of the excuses or reasons I have read today.

I very well might be wrong. But please tell me why and please give me a solution.

Thanks!

 

[kennybobby]

Cells in parallel will all see the same voltage, it can be no other way unless someone has repealed the laws of physics and we didn't get the memo.

The total capacity of the pack will be limited by the lowest capacity cell. To combine 2900 and 3400 mAh cells results in a pack limited by the 2900 mAh cells.

You can prove all this to yourself using 2 cells of different capacity--connect them in parallel and measure the voltage across each cell. Then add a load resistor and monitor the cell voltage and current during discharge. Pull a given amount of current for a given time, e.g. 0.5 Amps for 60 minutes. Now separate the 2 cells and measure the voltage individually. Unless one of the cells is damaged, then the resulting voltage will be identical.

Batteries function by chemical reactions, there is no internal resistor in a cell.

 

[Jimbo69ny]

Cells in parallel will all see the same voltage, it can be no other way unless someone has repealed the laws of physics and we didn't get the memo.

The total capacity of the pack will be limited by the lowest capacity cell. To combine 2900 and 3400 mAh cells results in a pack limited by the 2900 mAh cells.

You can prove all this to yourself using 2 cells of different capacity--connect them in parallel and measure the voltage across each cell. Then add a load resistor and monitor the cell voltage and current during discharge.

Batteries function by chemical reactions, there is no internal resistor in a cell.

No one is debating the laws of physics. I realize that voltage stays the same in parallel circuits. I apologize if I was misunderstood. Also, i said my OEM pack is 2600, not 2900.

Each cell does have a resistance and you didn't take that into account. Newer cells definitely have a lower resistance than older cells and that will effect the circuit. Until I am proven wrong, that seems to be true to me.

Is it your point that 3400 mah cells would be over kill? Judging by your post it would make sense to use 2600 mah cells and match the OEM cells.

 

[arber333]

No one is debating the laws of physics. I realize that voltage stays the same in parallel circuits. I apologize if I was misunderstood. Also, i said my OEM pack is 2600, not 2900.

Each cell does have a resistance and you didn't take that into account. Newer cells definitely have a lower resistance than older cells and that will effect the circuit. Until I am proven wrong, that seems to be true to me.

Is it your point that 3400 mah cells would be over kill? Judging by your post it would make sense to use 2600 mah cells and match the OEM cells.

I have done it! It worked for a time. I even had a different chem cells.
To summarize; i used 38x 200A LiFePo cells as promary pack. Then i added 36x LiPo cell pack in paralel with a disconnect contactor.
It worked and my range went further. There were weird things however. There was some current flowing inside the system from one pack to other and changing by the SoC of each pack. You couldnt notice it after the fuse but i put ammmeter on both + lines and scary amps were flowing. From 200A to 400A back and forth.
Both pack could charge together since both SoC voltage levels were in sync. So 80% SoC was the same voltage on both packs.

BUT! One day a cell went bad in LiFe pack. I took it out and forgot about voltage balance! $$$ mistake! LiPo pack was feeding current to LiFe during charging trough contactor that was operated by my BMS. So in theory i was safe.
In practice it meant that when LiFe were full contactor was opened and charger would soon go to CV since LiPo had stabler plateu after 4V.
What happened was that charger would charge LiPos with CC as usual, but LiFes were allready at 3.6V since one cell was missing. BMS opened contactor and protected them sure! But soon after voltage dropped a bit and BMS opened contactor..... LiPos has some 3V potential difference and probably 100A went in the pack! BMS was switching contactor furiously.... It cooked LiFes overnight. THE END.

That practical example shows you should paralel cells one after the other NOT parallel the packs. Murphy says something will happen and you will forget your arrangement.

In worst-best case you should make a change over arrangement so packs are not ever connected by + line.

I think best case would be if there would be a Prius style controler (charge pump) between packs. CC style. That way pack imballance would not be a factor. Also you could use two different voltages/chemistries.

my 5c

 

[Jimbo69ny]

I have done it! It worked for a time. I even had a different chem cells.
To summarize; i used 38x 200A LiFePo cells as promary pack. Then i added 36x LiPo cell pack in paralel with a disconnect contactor.
It worked and my range went further. There were weird things however. There was some current flowing inside the system from one pack to other and changing by the SoC of each pack. You couldnt notice it after the fuse but i put ammmeter on both + lines and scary amps were flowing. From 200A to 400A back and forth.
Both pack could charge together since both SoC voltage levels were in sync. So 80% SoC was the same voltage on both packs.

BUT! One day a cell went bad in LiFe pack. I took it out and forgot about voltage balance! $$$ mistake! LiPo pack was feeding current to LiFe during charging trough contactor that was operated by my BMS. So in theory i was safe.
In practice it meant that when LiFe were full contactor was opened and charger would soon go to CV since LiPo had stabler plateu after 4V.
What happened was that charger would charge LiPos with CC as usual, but LiFes were allready at 3.6V since one cell was missing. BMS opened contactor and protected them sure! But soon after voltage dropped a bit and BMS opened contactor..... LiPos has some 3V potential difference and probably 100A went in the pack! BMS was switching contactor furiously.... It cooked LiFes overnight. THE END.

That practical example shows you should paralel cells one after the other NOT parallel the packs. Murphy says something will happen and you will forget your arrangement.

In worst-best case you should make a change over arrangement so packs are not ever connected by + line.

I think best case would be if there would be a Prius style controler (charge pump) between packs. CC style. That way pack imballance would not be a factor. Also you could use two different voltages/chemistries.

my 5c

Thank you for your reply. It sounds like you had some issues. I would guess it was because you used different chemistry batteries with different discharge curves. You think? It sounds like you had a 36s1p arrangement, yeah, thats not good. Because as you found out if you dropped one cell the entire voltage of the pack changes. If you had a 36s2p arrangement with fuses on each cell you would've been fine. Capacity would've dropped and amperage for that one cell would go up because its partner died but thats all.

My system will include 9 modules. Within each module there will be 10 blocks in series. Then each module will be in series with each other. That will make up the battery pack. That pack will then be connected to the main battery.

As for on/off's, I purchased this switch. http://www.evwest.com/catalog/product_info.php?cPath=2_42&products_id=267
Are you saying that it should be on the negative line instead of the positive line?

I also have Tesla EV200 contactors that will be at my battery pack. If I can find the signal that Tesla sends to the OEM contactors then I will piggyback the signal from those to mine. That way my pack engages when the OEM pack engages. That will only happen while charging or driving. When the car is off the contactors will open.

The car has a BMS but I dont speak in CAN so I cant connect to it. So I will be installing a BMS to my battery too.

 

[arber333]

Thank you for your reply. It sounds like you had some issues. I would guess it was because you used different chemistry batteries with different discharge curves. You think? It sounds like you had a 36s1p arrangement, yeah, thats not good. Because as you found out if you dropped one cell the entire voltage of the pack changes. If you had a 36s2p arrangement with fuses on each cell you would've been fine. Capacity would've dropped and amperage for that one cell would go up because its partner died but thats all.

My system will include 9 modules. Within each module there will be 10 blocks in series. Then each module will be in series with each other. That will make up the battery pack. That pack will then be connected to the main battery.

As for on/off's, I purchased this switch. http://www.evwest.com/catalog/product_info.php?cPath=2_42&products_id=267
Are you saying that it should be on the negative line instead of the positive line?

I also have Tesla EV200 contactors that will be at my battery pack. If I can find the signal that Tesla sends to the OEM contactors then I will piggyback the signal from those to mine. That way my pack engages when the OEM pack engages. That will only happen while charging or driving. When the car is off the contactors will open.

The car has a BMS but I dont speak in CAN so I cant connect to it. So I will be installing a BMS to my battery too.

My mistake was just that i REMOVED one LiFe cell but didnt change charger voltage back one cell! Had i done that i would have been OK. Less capacity but healthy cells.

Heh OK = 0 Killed! get it?

I definitely recommend using one EV200 one the + line so your BMS has control over it. And make it an interlock disconnect (remove gnd with one relay and +12V with another). On the - line you can have manual disconnect, i used simple fuse.

Well i have my own BMS and it had done its job. I should have had an interlock circuit so contactor wouldnt be able to switch back on.

A

 

[Jimbo69ny]

Tested! You CAN run two different packs, with different internal resistance, in parallel. https://youtu.be/itXovf2J10w

 

[kennybobby]

That's a great test you have made with interesting results.

Did you happen to measure each cell voltage before the test--that would have helped show what is happening when the cells were put into parallel.

Also it would have been nice to see a volt meter connected across the resistor--that would show the current by the IR drop of the load.

What does the current read if you connect the two blue wires together without touching the load resistor--this would be the charging current from the new cell to the old one if the meters can read in both + and - directions.

So by your test is the internal resistance of the new cell, ~73 mohms, and that of the old cell about 174 mohms?

Next time maybe a selfie-stick on a tripod would help so both hands are free, otherwise excellent video.

 

[Jimbo69ny]

That's a great test you have made with interesting results.

Did you happen to measure each cell voltage before the test--that would have helped show what is happening when the cells were put into parallel.

Also it would have been nice to see a volt meter connected across the resistor--that would show the current by the IR drop of the load.

What does the current read if you connect the two blue wires together without touching the load resistor--this would be the charging current from the new cell to the old one if the meters can read in both + and - directions.

So by your test is the internal resistance of the new cell, ~73 mohms, and that of the old cell about 174 mohms?

Next time maybe a selfie-stick on a tripod would help so both hands are free, otherwise excellent video.

The voltage was somewhere around 3.9 each cell.

Yeah I guess I could've tested voltage at the resistor but I didnt think I needed to show that.

I connected my meter to both cells and showed that only .02 amps were transferring from one cell to the other.

I could never bring myself to buy a selfie stick. hahaha

I could do another video if there was more you want to test. I'm just limited by crappy meters.

 

[digsys]

These are some results I found from making up Li race packs for years, some of this may help -
The worst thing you can do is have different types / ages / conditions of cells in a SINGLE string - that is a 100% NONO.
Other than that - You will get away with paralleling 2+ packs of the SAME chemistry type BUT note :
a/ If one is lower capacity, then the charger may/will force higher than ideal currents into it - as the charge voltage is fixed for both.
b/ In drain situation, the "lower" pack won't supply as much current as it is capable of - due to having a higher internal impedance (some call resistance)
c/ As the lower capacity pack ages, it will be "stressed" more and more.
It is IMPORTANT that you monitor both with a BMS ! - that can alert if any cell (block) goes too high or low in voltage (c/w a median voltage).
Bonus - the main pack doesn't (usually) suffer any ill effects.
In our packs, we have 5-7+ cells hard wired together per module. The reason is - in 99.9% of all cell failures, they go O/C or high impedance, thereby removing themselves nicely from the (single) string and block. The BMS picks that up easily, so we re-reduce load / charge profiles.
Note: These are normal parameter failures, not catastrophic failures, where anything can happen
In summary - you'll "get away with it" - how long the 2nd pack will last depends on many factors.