Parrallel battery modules safe way of doing...

Hi Everyone,

I have a small project I hope to complete over the next little while. It is retro fitting a smart car battery into a model A ford.

In order to run at a safe voltage that the ev converted ford needs, I must parrallel the 3 x 31s modules found in the 2013 smart car battery.

What would be the safest way to do this? Should I parrallel all the modules cells with say 14gauge wire then connect things to a bms?

It should be possible to pull 600amps or more @ 120v minimum. Pleanty to get the truck around with its big warp 11 (overkill i know).

I need to figure out the specs of the smart cars batteries in order to know the exact limits of power it can provide.

https://youtu.be/wYhI0Fl-Xnk

Thanks
-Steveo

If you think 14awg can handle 300A longer then you can blink your eyes you might want to skip this project.

flippy said:

If you think 14awg can handle 300A longer then you can blink your eyes you might want to skip this project.

Click to expand...

I don't think you understand the question, or know who your replying to. But the way I put it, if you don't have anything good to say don't say anything at all.

I understand the question. If you use 3 different packs and parralel up each pack and if one of the bms on a pack triggers during a discharge you will see all the current go into the parralel connections.

You can prevent this but you must strap all 3 packs as a whole and use a single bms for all 3. And i would use 4 or 6awg for the parralel connections. That can handle a 600A peak without a problem for a short while if something bad should happen.

I haven't been able to watch your video from start to finish, but am I assuming you don't want to break up the 3 x 31S batteries into 93 modules to make a single 31S3P battery.

If that's the case, then yes, Big heavy 00 AWG cables on each battery, (I.e. each set of 120v), and 14AWG cables between each row of cells, then one single BMS on the end will suffice.

Flippy... As helpful as he can be, sometimes makes assumptions that don't hold true for a project. In this case, he seems to be thinking you're using 14AWG everywhere.

Flippy - By using 00AWG on the whole battery and 14AWG for the inter-cell balancing, only balancing current will be going over the 14AWG wires. If all cells are perfectly balanced at all time, then no current at all will go over the wires. You can load up 600A across all three packs, and if you clamped an ammeter over the 14AWGs, it'd show zero current.

Edit: I should note, you probably want to check internal resistance and capacity of each cell before you settle on wire size. What you propose is safe if the cells are at least similar. If you had one cell which had say 10% of the capacity of its two peers, what will happen is that during heavy load, the faulty cell will want to sag, and the other two cell will try to supply the energy to prevent it from sagging, and in that case, the 14AWG will carry a fair bit of current.

Like wise, if the internal resistance is high, it will sag during load, and during the bounce back, its voltage will be higher than the others, and try to rush current backwards.

But if the cells are even vaguely close, 14AWG and lengths measured in inches/cm, it should be completely fine.

Sunder said:

I haven't been able to watch your video from start to finish, but am I assuming you don't want to break up the 3 x 31S batteries into 93 modules to make a single 31S3P battery.

If that's the case, then yes, Big heavy 00 AWG cables on each battery, (I.e. each set of 120v), and 14AWG cables between each row of cells, then one single BMS on the end will suffice.

Flippy... As helpful as he can be, sometimes makes assumptions that don't hold true for a project. In this case, he seems to be thinking you're using 14AWG everywhere.

Flippy - By using 00AWG on the whole battery and 14AWG for the inter-cell balancing, only balancing current will be going over the 14AWG wires. If all cells are perfectly balanced at all time, then no current at all will go over the wires. You can load up 600A across all three packs, and if you clamped an ammeter over the 14AWGs, it'd show zero current.

Edit: I should note, you probably want to check internal resistance and capacity of each cell before you settle on wire size. What you propose is safe if the cells are at least similar. If you had one cell which had say 10% of the capacity of its two peers, what will happen is that during heavy load, the faulty cell will want to sag, and the other two cell will try to supply the energy to prevent it from sagging, and in that case, the 14AWG will carry a fair bit of current.

Like wise, if the internal resistance is high, it will sag during load, and during the bounce back, its voltage will be higher than the others, and try to rush current backwards.

But if the cells are even vaguely close, 14AWG and lengths measured in inches/cm, it should be completely fine.

Click to expand...

Thank you for the help, this is exactly what I had in mind for setup. I will do my best to update everyone on progress.

-Steveo

Steveo has certainly been around and will surely come out with something good.

One false assumption made by flippy was that individual BMS's would include cutout (not so on full size EV)
Small Ebike BMS's include solid state primary output control
Big kid BMS's do this differently

If I were Steveo...
I would not wire them all up in parallel
(but its a fine way to do it and I am stoked to see his progress develop)

If this were a perfect world (which it is not)

Steveo would be able to utilize the previously embedded *monitoring and balancing* boards - which are a part of - the overall BMS scheme.

If Steveo had 3 "modules" and he wanted to run them all in parallel
The three modules could be paralleled at only the primary contacts
No contacts at the cell level required~!

Each module balances itself
Modules are "forced together" at the primaries... which leads to the obvious problem... of "what if one fails"

Well...

Each module has the ability to communicate*

There is always a "Buss Master" (an Arduino) which collects and interprets the communications (CAN or isoSPI)
Bussmaster controls the Primary Contactor or Contactors

Buss Master may control a single contactor - where all three packs are hard wired together at the primary terminals
(A little dangerous)

Buss Master may control 3 contactors, which tie together the three packs in [strike]Series[/strike] parallel
(Or 6 contactors... where we break at both the (+) and (-)
(This method introduces additional failure modes)

The failure mode we are most concerned with is a single "cell" (or pack of local cells in parallel) fail
If we have 3 modules
If 2 are "good"
If a 3rd has a single weak link...

And all three are tied together in parallel at the primary (not at the cells)
One can envision many failure modes

Realistically... you have to integrate over a time period to properly assess the risk
Before time Zero
Time Zero
After time Zero

Since we are running a BMS master (your Arduino...) we KNOW (not suspect..)
We KNOW
The state of the individual modules

They are GOOD or BAD
(or heading toward bad)

Batteries blow up (99.99999999999%) of the time during
CHARGE
or
DISCHARGE

A single primary contactor controlled by a BMS master can inhibit both Charging and Discharging
It can report to the user the state of the system

So -
For R&D
I am fine with modules being connected together with #0000 at the primaries
With the balance taps not being connected together
With a single contactor isolating this part of the system from the Traction Controller and Charger

(NOT an official stance but it will work)

Officially one is suppose to pack up modules in Parallel First
Then... Wire those modules in Series
But... I totally get it... we are using what we can find here

...

I have found massively interconnected batteries (balance taps linked) introduce more risk and more failure modes than non-cell-level-linked hacked together systems.

Too many wires to pinch
Too many possible bad crimps
Too many rub points
Too many ways for primary current to try and run thru balance jumpers
Too much work

- so... he is just trying to get this up and running -
I am supportive of whatever he comes up with!

...

In the future he will be able to spend about $50-$150/module to install monitoring and balancing at the module level.
He will be able to spend about $50-$150 on a BMS Master that controls his contactors
(and handles Isolation Testing)
(and handles datalogging)
(and handles precharge)
(and handles other duties and tasks... like communication with High Rate DC Fast Chargers...)

So
Deep subject

Looking out 5 years you will really want (NEED) to be compliant with protocol
Safety
Charge Communication
Stuff like that

...

I know that Tesla Modules come with Monitoring and Balancing built in
They come in what - 60V chunks I think?
They communicate in an isolated manner
They can be stacked in Series

You COULD stack them parallel
My opinion is that we want to GET AWAY from high AH packs and toward High Voltage packs
(thats a 5 page writeup)

Suffice to say

You can usually accomplish what you are after by using higher voltage (on the same KWH) pack as you can by building up capacity at a lower voltage. I totally get that you have to hit the voltage which is compatible with the hardware you have

... as time goes on... this hardware will become ubiquitous... so it will be no big deal.

Good luck and have fun

Search Terms:
CCS1 & CCS2
Chademo
Tesla
Level 1 charging
Level 2 charging
Level 3 charging
SAE Combo
J1772
On-board AC
On-board DC
Interlocks

-methods

Watched the video

Yep - as I suspected

Due to proprietary (non-publicly documented) interface definitions Steveo is forced to bypass the included (and very frocking useful) balancing modules.

Sigh
Capitalism at work!

Well
From the video it stated that he had a controller with a maximum voltage to work with
That this controller was "current control" (meaning that it really does not care what the voltage is)
That he needed to hit a compatible voltage

He stated that he wanted to continue to use the cooling (good idea)
and... this sort of kills the idea of completely rebuilding the pack
Forcing him to build the modules up in parallel
Leaving him with the choice to try and parallel up the cells - or just tie at the tops

By running a lower voltage he avoids the issues around High Voltage ...
Where normally high voltage is anything over 60V
But... meh... I have been hit with 120V AC and DC hundreds of times
(its more about the capacitance and other LRC in the system than the RMS voltage - different subject)

No idea how he intends to monitor the cells
The cells... sigh... REALLY SHOULD BE MONITORED TO BE SAFE
Yea - monitoring on a cycle works... but who actually does it?

If he does not wire the ballance taps in parallel... he would need 3pcs of monitoring system
If he does wire the ballance taps in parallel... he could use Hobby King dongles to at least monitor on some level

...

As far as gauge for the balance taps

...

I would learn toward a gauge big enough to not fail in shock and vibration
Small enough that it will just fuse out and POP should things go real bad
I would consider fusing them or using resettable poly fuses (think hard about this...)

...

If I were doing it right now in my garage?

...

For certain I would just rig those up paralleled at the primaries
I would not try and rig all the cells together (because I am lazy)
I would check it like crazy for a few months and until I technology (and the market) catches up

...

So many choices

Watching

-methods

Final note:

Key pivot in the risk assessment and decision making is:

"Can he decode the information coming to/from the built in modules"?

If so -
There are 3 or 4 very clear and straightforward paths to take which will qualify as "totally legit"

if not -
He is inhibited and forced down a couple of alternate paths

Since the timeline looks slow...
He may consider...
Sitting down with a CAN analyzer and trying to see if the protocol is simple enough to decode

Going back to the time arguement
You can manage A LOT OF RISK
by effectively running a PID loop

What was the state of the pack?
What is the state of the pack?
What do we project the state of the pack being?

With that information we know exactly how and when to
Precharge
Limit or terminate power draw
Limit or terminate high rate charge
Emergency blow out contactors

...

And dont forget the temp sensors - data which is usually picked up by the modules
Thermal is a SUPER USEFUL way to make statements about the safety of your pack

-methods

Well thought out, if a somewhat difficult to read post.

Do you see any evidence that better control units are coming out? I'd prefer to manage as you describe too - Every cell gets its own protection, and they all talk back to a "Bus Master" (If I understand you correctly, some kind of control unit).

The current BMS on my Vectrix is built that way (The LiFePo4 main, not the LiPo IR reducing pack). In theory it's elegant. In practice... I had issues with it.

Wish I could get something like that for my LTO pack. Instead, I'm using per cell, not-fit-for-purpose super-capacitor protection boards that don't talk to each other (or anyone else for that matter.

Sunder said:

Flippy... As helpful as he can be, sometimes makes assumptions that don't hold true for a project. In this case, he seems to be thinking you're using 14AWG everywhere.

Flippy - By using 00AWG on the whole battery and 14AWG for the inter-cell balancing, only balancing current will be going over the 14AWG wires. If all cells are perfectly balanced at all time, then no current at all will go over the wires. You can load up 600A across all three packs, and if you clamped an ammeter over the 14AWGs, it'd show zero current.

Click to expand...

yes and no. in a perfect world this woiuld be true.

but in my experience you cannot get 3 packs that are identical in IR or wear. so you so get quite a bit of current on the balance wires when hammering it due to things like temperature changing the IR of the hottest and coldest cells just for starters, wear and uneven capacity loss also add up with each cell group of every single battery. those currents need a place to go. 14AWG is simply too dinky, especially when a cell or even a whole group pops, then you get some serious current on the balance wires when it naturally tries to balance it out.
reminder: we are not talking about a ebike battery, this is a full EV pack where the currents and risks are pretty much scaled up consideriably compared to the DIY downtube batteries people here usually deal with.

the orginial question was what the SAFE way to do this was. and that is using properly sized wires for a worst case scenario.

in this case there is not really a reason to cheap out on 10ft of 6AWG versus 24AWG. the cost and effort difference is not exactly huge.

Sunder said:

Well thought out, if a somewhat difficult to read post.

Click to expand...

Sorry about the riddle form.

Even close to a decade after signing NDA's (with people who literally threaten death by accusation of treason) and after going out and developing all of the core knowledge from (re)scratch under my own umbrella... I am still VERY CAREFUL about how I describe methods and ways which are known to work.

I am not even supposed to be on the forums :? but - I make sure to cruise by so as to make sure that you guys still have a free and open platform to discuss safety and reliability around what is obviously the future of global everything. To my thinking... so long as I still see "dangerous stupid shit" going on around Lithium Batteries... I am still "GREEN LIGHT" to stoke innovation at the public level.

Sigh...

Surf was up Yesterday. Kimberly got hammered by shore pound and she is still shaking sand out of her ears. I am going back out today to burn off nervous energy about interview week.

"Making it happen" is so much more complicated than it needs to be - when we have to tip toe through capitalist competitiveness.

Good thing what we do here is LAY OUT PRIOR ART LIKE ITS NOBODIES BUSINESS :x

-methods