The "thermal management of LiPo" thread!

jonescg said:
Passing coolant through the busbars has been explored before, but the issue is when you have more than one series cell in your battery - you can't have a continuous connection between conductors at different potentials. Even using plastic tube as an insulating break between busbars is fraught, because the direct current will cause electrolysis, especially when an aqueous coolant is used.
Click to expand...

Heat pipes aren't hollow busbars per se.

They are sealed units containing small quantities of liquid at low pressure. You connect one end to a heat sink and the other to your heat source, and they use the evaporation/condensation cycle to transfer the heat from one end to the other. They can have an equivalent thermal conductivity 100s of times higher than a simple copper bar or tube the same mass and length.

You would use 1 per parallel string, with one end of each thermally connected to the external case or other heat sink.

Other liquids besides water -- ammonia, alcohol etc. -- are possible for different heat ranges, but usually with aluminium rather than copper.
 
i have 3 questions...

1. Why are you running your LIPO packs so hard that you need to think about cooling them?
2. Don't you know that running a LIPO pack hard shortens it's life significantly?
3. Ever hear of using larger LIPO packs so that over heating is never a problem?

I guess if you need high current output, low weight and will use the pack for 5 minutes, then making a small pack that gets slammed is the way to go. For everybody else, probably this is not the way to go. Maybe even for folks that need a tiny race pack, the added complexity for cooling could be put into a larger pack instead that doesn't have heating issues? The physical size difference between a 10ah cell and a 12ah cell is not very much.

I'm typically running 2 or 3C and my packs never get hot or hardly even warm. They also last a long time and I get lots of range too.
I do run at 40-80 amps continuous on most of my EV's. If thermal issues in a LIPO pack are an issue, build a BIGGER pack.

A few years ago I messed with putting 2mm aluminum sheets between each LIPO cell. The tops and bottoms of each sheet connected to a 6mm thick aluminum heat sink. The idea was to mitigate heating inside the pack. What I found is the pack got MUCH larger and heavier and while it did stay cooler, the penalties far outweighed the benefits. In the end, I realized it was a dumb idea as I gained far less than I originally thought I would get. I went back to the KISS principal and used larger cells instead. This got me a similar sized pack as the silly one with the heat sinking in it that had far more capacity and amperage capabilities without getting warm.
 
ElectricGod said:
i have 3 questions...

1. Why are you running your LIPO packs so hard that you need to think about cooling them?
2. Don't you know that running a LIPO pack hard shortens it's life significantly?
3. Ever hear of using larger LIPO packs so that over heating is never a problem?
Click to expand...

From the first page...

jonescg said:
I still believe you should design your pack according to the demands of the system, such that you aren't hammering them too hard. Quite hard to do in a flight situation where weight and sustained power demands are critical, but in a road going EV it's certainly possible to overspec the battery.
Click to expand...

Cooling any battery will make it last longer. The more time spent over 20'C the faster they will degrade - Arrhenius' law.
The race bike is getting hit at 28 C = 280 amps. It only gets hot by the last session of the track day. No thermal management other than convection. But this is because the cells are very low DCIR - about 1 mOhm. If you want higher specific energy cells, like over 200 Wh/kg, you have to accept a hit on the DCIR - I've got some cells which are 20 mOhm. These WILL need thermal management as they'll get warm after a 2C discharge.

Anyway, I have since devised a solution:
12s10p_finished.jpg

Liquid cooling plate on the bottom face of the battery pack. Results will be posted shortly.
 
I recently noticed that high-amp cells have much wider tabs, and some even have tabs on opposite ends in order to allow the tabs to be as physically large as possible.

I also recently stumbled across some patents on cooling the tabs. High-amp cells typically end up using thick copper bus-bars, and I think a useful experiment might be to add a temp probe in the centermost part of a stack of cells (assuming that's the hottest part), and measure the temp difference between the center of the pack and the thick copper bus-bars.

I think thick bus-bars have been used for their high-current capabilities, but I suspect we haven't appreciated the significance of their role in adding some cooling as a heat-sink...
 
One of my recently completed LiPo modules with the liquid cooling plate on the base. The thermal epoxy has a thermal conductivity of 1.2 W/m.K which is nothing flash, but it's only about 1-2 mm thick, and it's oozed up in between the cells. That bottom surface represents about 20% of the available surface area, and it's in the plane of the highest thermal conductivity - so I have a good feeling about it's ability to maintain something close to ambient. I still haven't run the test yet (mainly just trying to build the thing in time!) but hopefully this week.

 
So I did a discharge test and measured the efficacy of the cooling loop with one of these modules. It's not great; better than nothing, but I'd prefer more heat transfer.



This was a massively artificial discharge test though - 4.4 C for 8 minutes from a cell rated to 2 C continuous. Still, the delta T on the coolant, given the high flow rate suggested barely 120 W was being removed as the cells generated nearly a kilowatt of waste heat.

So I wonder what the better option might be. I could easily put 9 L-shaped sheets of 0.2 mm aluminium sheet between each cell and use a silver-based goo to make contact with the cooling baseplate. This way any heat from the cell face will be conducted away by the sheet. I reckon it has the potential to pull more heat out, but the effective area for heat conduction ends up being 12 cell blocks x 9 per block x 50 mm wide x 0.2 mm thick = 1080 mm2 worth of conductor to the baseplate. The link from cell to baseplate is about 5 mm long, at worst.

Since Q = (A x dT x k)/L
and A = 0.001080 m2, k=203 W/m.K, dT is about 20 Kelvin, and L is about 0.005 m (its a short trip to the baseplate)

877 W for the whole module.

Now, that's pretty ambitious, but even half of that would be a big improvement on just the potting compound alone... :?
 
Stuff looks amazing, but there's $17,000 worth of graphite for a $24,000 battery.
I reckon aluminium might be a winner hey...
 
So this is the current thinking - L=shaped cooling plates which link the flat surface of each cell to the baseplate.

Cooling fins.jpg

I'm doubtful it will be that much better than the current system with potting compound, but I'll have to come up with a prototype and test it.
 
I have seen this exact configuration in some commercial packs, terminated onto a water-cooled baseplate. Good reason to think tab cooling on pouch cells would be more effective though (Couple of reseach papers attached)!
 

Attachments

  • LionTabVSSurfaceCooling.pdf
    1,007.7 KB · Views: 34
jonescg said:
Stuff looks amazing, but there's $17,000 worth of graphite for a $24,000 battery.
I reckon aluminium might be a winner hey...
Click to expand...

Too bad it's so expensive.

The article on tab cooling was interesting. Seems more heat is generated on the tab end because the current density is higher.
If the tabs were connected by clamps that had cooling water tubes, it might not be too hard or expensive. Each tab clamp would need a pair of tube ends sticking out and they get joined by rubber tubing sections. Using distilled water, the voltage differential between tabs won't be an issue. If the clamps were brass or copper, you could just solder a piece of small copper tubing to each one.

An even more simplistic approach (but probably impractical for some reason) would be to make a "box" around all the tab connections that's sealed around the edges and just circulate water around the whole area. You'd have to seal between the cells and around the edges. Self-leveling silicone glue would be good for this. The water would need to be distilled and probably replaced pretty often to maintain low conductivity.

I have some big lasers at work that use direct water cooling around the high voltage parts. Like up to 20kV. Pure distilled water is a very good insulator. It just has to stay pure.
 
For simplicity I would be looking at aluminium bus bars with a cooling passage passing up the centre and keep the flow rate up high enough to ensure the first and last cell served by the coolant aren't at greatly different temperatures.

You could maybe use an off the shelf extrusion or hollow section but the thin wall thickness would mean you'd need either a screwless method of clamp the tabs to it, or if you drill & tap, the screws would likely poke into the coolant gallery and need sealing.

A custom bus made from solid bar with a small hole bored up the centre would have enough meat to drill & tap blind holes, eliminating the possibility of leaks.

The coolant gallery would only need to be small to remove plenty of heat. Maybe ~1/4".
 
fechter said:
jonescg said:
Stuff looks amazing, but there's $17,000 worth of graphite for a $24,000 battery.
I reckon aluminium might be a winner hey...
Click to expand...

Too bad it's so expensive.

The article on tab cooling was interesting. Seems more heat is generated on the tab end because the current density is higher.
If the tabs were connected by clamps that had cooling water tubes, it might not be too hard or expensive. Each tab clamp would need a pair of tube ends sticking out and they get joined by rubber tubing sections. Using distilled water, the voltage differential between tabs won't be an issue. If the clamps were brass or copper, you could just solder a piece of small copper tubing to each one.

An even more simplistic approach (but probably impractical for some reason) would be to make a "box" around all the tab connections that's sealed around the edges and just circulate water around the whole area. You'd have to seal between the cells and around the edges. Self-leveling silicone glue would be good for this. The water would need to be distilled and probably replaced pretty often to maintain low conductivity.

I have some big lasers at work that use direct water cooling around the high voltage parts. Like up to 20kV. Pure distilled water is a very good insulator. It just has to stay pure.
Click to expand...

The DI water trick works if the fluid never wets something that's an available ion source (including atmospheric CO2). So for a well cleaned laser glass tube setup in a sealed loop it's pretty ideal. For cooling a battery, de-ionized water becomes ionized water in a few seconds after touching the cells (or being exposed to atmosphere).

Similarly, immersion cooling on pouches has the same thermal problem of clamping the pouch between cold-plates. This is because the decay mechanic is related to non-uniform temperatures between layers. If you cool the tabs or the edges of the pouch only, the Rth path stays uniform between layers, and hence the utilization of the active material becomes proportionately uniform. If you take the same pouch cell and and use immersion cooling, now the top and bottom foils in the pouch always have a lower temperature than the inner foil layers, and hence in use the inner layer active material gets cycled to decay/failure while the outer foils stay fresh from minimal use.

If you cool from the edges of the pouch or the tabs of the pouch (and keep a heat path off the faces of the cell), the foil layers all get cycled together.
 
fechter said:
Using distilled water, the voltage differential between tabs won't be an issue. If the clamps were brass or copper, you could just solder a piece of small copper tubing to each one.
...
I have some big lasers at work that use direct water cooling around the high voltage parts. Like up to 20kV. Pure distilled water is a very good insulator. It just has to stay pure.
Click to expand...

Yeah we tested 4 volts and aluminium conductors in pure, distilled, deionised water. It was eroded away within 24 hours. Direct water cooling of conductors is just not an option. Chilled air is the better option in this case.
 
liveforphysics said:
Similarly, immersion cooling on pouches has the same thermal problem of clamping the pouch between cold-plates. This is because the decay mechanic is related to non-uniform temperatures between layers. If you cool the tabs or the edges of the pouch only, the Rth path stays uniform between layers, and hence the utilization of the active material becomes proportionately uniform. If you take the same pouch cell and and use immersion cooling, now the top and bottom foils in the pouch always have a lower temperature than the inner foil layers, and hence in use the inner layer active material gets cycled to decay/failure while the outer foils stay fresh from minimal use.
Click to expand...

Good info :bigthumb:

Presumably the IR drops on the inner layers due to their greater temperature and this results in them being over-loaded compared to the cooler, outer layers?
 
Punx0r said:
liveforphysics said:
Similarly, immersion cooling on pouches has the same thermal problem of clamping the pouch between cold-plates. This is because the decay mechanic is related to non-uniform temperatures between layers. If you cool the tabs or the edges of the pouch only, the Rth path stays uniform between layers, and hence the utilization of the active material becomes proportionately uniform. If you take the same pouch cell and and use immersion cooling, now the top and bottom foils in the pouch always have a lower temperature than the inner foil layers, and hence in use the inner layer active material gets cycled to decay/failure while the outer foils stay fresh from minimal use.
Click to expand...

Good info :bigthumb:

Presumably the IR drops on the inner layers due to their greater temperature and this results in them being over-loaded compared to the cooler, outer layers?
Click to expand...


Yep, that's it.
 
Has anyone tried using 3M Novac to fully submerge batteries? There are some Datacenters and PC builds that use it to cool components. The fluids boiling point is 61°C, which is a bit high for batteries, but I'm pretty sure at least it has a lot of heat capture. https://www.youtube.com/watch?v=V7-QWmT2jfk here is an example of a ~900W being dissipated. There is also the benefit of ''thermal runaway protection'' https://www.youtube.com/watch?v=6hThYnL9zzE
 
multifrag said:
Has anyone tried using 3M Novac to fully submerge batteries?
Click to expand...

That's an interesting idea. I wonder how much it costs. Something cheaper, like mineral oil or the lube oil made for hybrid car AC systems might also work (but flammable). Keeping it from leaking out would be the main challenge. You'd also want to be sure it's compatible with the pouch materials.
 
fechter said:
multifrag said:
Has anyone tried using 3M Novac to fully submerge batteries?
Click to expand...

That's an interesting idea. I wonder how much it costs. Something cheaper, like mineral oil or the lube oil made for hybrid car AC systems might also work (but flammable). Keeping it from leaking out would be the main challenge. You'd also want to be sure it's compatible with the pouch materials.
Click to expand...

Transformer oil doesn't have the benefit of phase change(moving the heat away with out intricate routing), but it is readily available, while novec you have to order it as a company, with big batch. https://www.youtube.com/watch?v=CGTNg-uIuRI In this video he cools down a Laptop which uses pouch cells. No idea how they will interact with longer tests.
 
Several battery-powered device companies are using 3M Novec (or equivalent) to cool their battery packs, but none that I'm aware of cooling LiPo or pouch cells.

Xing are using it in their products: https://www.xingmobility.com/
Kreisel are using it in their products: http://www.kreiselelectric.com/en/

And yes, Novec is seriously expensive stuff. Like, $800 a litre expensive.
 
One thing you can actually use is paraffin wax in conjunction with aluminium heatsinks.
 
You must log in or register to reply here.

Similar threads

W
BMS compatibility with new controllers and battery build 16s9p
Replies
14
Views
258
WonderDrift
W
B
Battery pack shuts off after a while
Replies
2
Views
264
bogddy
B
L
Joe's E-bike build
2
Replies
48
Views
4,435
lightbulb010
L
H
Belt and chain custom frame build thread
Replies
20
Views
2,081
Hummina Shadeeba
H
S
Separate bettery packs in parallel
Replies
2
Views
536
Sattva Ram
S
Back
Top