DIY Electric Car Forums banner
1 - 16 of 16 Posts

·
Registered
Joined
·
124 Posts
Discussion Starter · #1 ·
If I wanted to build a pack out of 20 of 3p4s modules configured in parallel (2 modules paralleled for 2p10s), I understand the BMS would be monitoring/balancing each pair of modules, so it can't balance the cells in the module. Would I want to first balance the cells independently of the final setup? What's the easiest way to do this?

Edit: or am I being dumb here, and the act of paralleling the cells causes the charge to flow between them and become balanced anyway.
 

·
Registered
Joined
·
1,071 Posts
So much to unpack here.

Two 3p4s modules paralleled does not give 6p4s much less 2p10s

It gives 2p(3p4s)

Are you connecting the two sets of balance wires to be monitored/ protected by a single BMS?

To balance cells with each other,disconnect from the pack,isolate all serial connections and put them all together in parallel at 1S.

So, you have 240 cells in total?

What total voltage do you need? What layout would you want if you just rebuilt it as one single pack?
 

·
Registered
Joined
·
124 Posts
Discussion Starter · #3 ·
2p10s from the BMS's standpoint is all I meant. I'm targeting 110-120v using the LG iPace modules, 11v nominal per module. And yes, monitored by a single Orion BMS (though one slave box, since there will be a front and rear pack). I'm trying to avoid rebuilding the modules for simplicity, but trying to hit in the neighborhood of 40kwh.
 

·
Registered
Joined
·
1,071 Posts
2p10s from the BMS's standpoint is all I meant.
So treating each module - or pair of modules - as if they were a single cell???

If there is any protective circuit or BMS within each module, they will need to be removed, go with just raw cells at that level.

Which of course means flying blind to what's going on within the modules,not a good idea.
 

·
Registered
Joined
·
7,970 Posts
If I wanted to build a pack out of 20 of 3p4s modules configured in parallel (2 modules paralleled for 2p10s)
It appears the plan to is to have 10 sets in series each consisting of two modules in parallel, each module being 4s3p. That's 10s(2p(4s3p)), or 240 cells arranged in 40s overall. A problem is that it is not 6p: the groups of 3 parallel cells in each modules are not paralleled with another 3-cell group in the adjacent module - only the end terminals of the modules are connected in parallel.

... I understand the BMS would be monitoring/balancing each pair of modules, so it can't balance the cells in the module. Would I want to first balance the cells independently of the final setup?
The BMS doesn't really handle modules - it handles cell groups. There would be 80 3-cell groups for the BMS to handle, and it's going to try to balance all of them to the same voltage per cell group; almost accidentally, the means approximately balancing every module voltage, without regard to which module is wired to which one.

... or am I being dumb here, and the act of paralleling the cells causes the charge to flow between them and become balanced anyway.
You're not planning to parallel cells at all; you're planning to parallel modules. If you just connect two modules at slightly different states of charge, they will equalize in overall charge, but if the cell groups within them are not already matched, they will remain mis-matched.


One other issue with the original description: the modules from the Jaguar I-Pace are not 3p4s (or 4S3P... there is unfortunately no consistency in this notation), they have 12 cells, but they have groups of 4 cells in parallel, three of those groups in series for 3S4P; that's why they have a nominal voltage of about 11 V (3 * 3.75 V = 11.25 V). Put those modules in parallel pairs and ten of those pairs in series and you have 10S(2P(3S4P)), or 30S overall for a nominal voltage of 133 volts. I assume from this:
I'm targeting 110-120v using the LG iPace modules, 11v nominal per module.
...trying to hit in the neighborhood of 40kwh.
... that the actual plan is for 30S overall.

I understand that you need to work with the modules which are available, but if what you want is 30S8P overall (for about 113 V nominal and 50 kWh) it would be a lot more straightforward to use an appropriate number of modules simply connected in series; unfortunately, this "VDA 355" size of module (which always has 12 cells) only seems to be available in 3S4P, 4S3P, and 6S2P... not the 2S6P that would better suit this low-voltage application (e.g. 15 modules in series for 30S6P: 113 V nominal and 37 kWh). No one currently offers this 2S6P VDA 355 module configuration because when strung together to reach the voltage used in modern hybrids and EVs it would be a huge battery; 48 of them would be 360 V 120 kWh, and for this overall size manufacturers want fewer modules for less complication.
 

·
Registered
Joined
·
7,970 Posts
Why is the target voltage so low? If I understand the project correctly (I might not) this is for a HyPer 9 motor (in a Jeep Wrangler). The small modules are probably desirable for packaging, but it's hard to arrange them for such a low voltage yet have enough energy capacity.

Why not use the HyPer 9 HV (instead of the low-voltage HyPer 9) - the two versions of the HyPer9 have the same rotor and housing, with the difference just being stator winding to change the balance between current and voltage for the same magnetic field and performance. To match the higher 144 V nominal design voltage use 14 modules at 4S3P each (the I-Pace module) for 42S4P yielding 158 V nominal and 35 kWh? Peak battery voltage at 4.0 V/cell would be 168 V, likely within the controller's limit, but of course you would check that in the controller documentation.
 

·
Registered
Joined
·
124 Posts
Discussion Starter · #7 ·
Yeah, I've gone back and forth between the HV and non-HV possibilities. The only reason I brought up the non-HV in this discussion is that I though I could parallel to hit the non-HV with 20 modules, while to hit the max voltage of the HV (~176V) I'm limited to about 14 modules with them charged to 90% capacity (14*4.15v = 174.3v). And maybe that's enough at 36.4kwh.

Basically I'm just exploring ways to pack more kwh in on either Hyper9 motor. But it does sound like I need to improve my understanding of balancing on non-trivial cell configurations. And I should consider whether it might not be worth it to break the modules apart and rewire them.

At any rate this is all in the total conceptual design phase, so lots of time to learn and adjust.
 

·
Registered
Joined
·
1,071 Posts
My head is spinning

3p4s (or 4S3P... there is unfortunately no consistency in this notation)
Sure there is.

With wired connections, if the cell-level parallel grouping is done first at 1S level as is the norm

then those groups are series connected in a single string

that is 3P4S.

4S3P would mean 3 strings in effect sub-packs, those connected in parallel at the 4S level voltage

definitely suboptimal.

With spot-welded packs where strips and plates are used rather than cabling the serial and parallel connections are actually all connected together at the same time

so there is no "first" or "lowest level"

thus the layout notation ordering can go either way correctly.
 

·
Registered
Joined
·
3,096 Posts
Before you have a mega-amp flowing between module strings by connecting them together, you need to bleed them to equivalence with a resistor between them. Once the module string pairs balance in voltage, you can connect them together.
 

·
Registered
Joined
·
7,970 Posts
One other issue with the original description: the modules from the Jaguar I-Pace are not 3p4s (or 4S3P... there is unfortunately no consistency in this notation)...
Sure there is.

With wired connections, if the cell-level parallel grouping is done first at 1S level as is the norm then those groups are series connected in a single string that is 3P4S.

4S3P would mean 3 strings in effect sub-packs, those connected in parallel at the 4S level voltage definitely suboptimal.
Unfortunately not. With cell-level grouping in fours, and eight of those groups connected in series to make a module, most people would call that an 8S4P module, but as discussed in a recent thread LG Chem labels that 4P8S... and apparently so would you, John. No manufacturer would be stupid enough to build a module with four strings of 8 cells each (although some DIYer's would ;)) - the only question is the notation.

For an example of the notation that I (and others) expect, GM describes the Chevrolet Volt batteries as 96S 3P (first generation) and 96S 2P (second generation)... not 2P 96S and 3P 96S... and they use LG Chem components!
 

·
Administrator
Joined
·
6,443 Posts
Couple of points

You can connect cells in parallel - but that does not mean that they will be balanced -
what you have to do is connect them in parallel and then charge them up to where the cell voltage starts to curl upwards
That will get them "Top Balanced"
The other alternative is to discharge them to where the cell voltage starts to curl downwards
That will get them "Bottom Balanced"

You are using aftermarket cells?? - some of them WILL die on you
If you have paralleled up a bunch of cells and one dies it will probably take its mates with it

When I was using Headway cells after a couple of such experiences I used four separate strings - at least they only died one at a time that way

Finally - Voltage!
You need lots of volts to push decent current through your motor at higher rpm
I would say that 144 volts is about the lowest voltage for a road car
 
1 - 16 of 16 Posts
Top