[ndg]

Hi,

I've acquired an older LiFePo4 pack that I would like to repurpose for my own build (it was very cheap!). It may only get used for testing motors and controllers, but that is dependant on if I can reconfigure it for my needs.

I need to reconfigure the pack to up the voltage from 217 to 352 to get the most out of the motor I have. So I'm going from 68s40p to 55s20p and dropping a couple of the 'modules' (not really modules just banks of cells). I've worked out a basic layout to get this configuration. The pack as it stands is rated at 200A. The cells are capable of delivering much more than this at 1200A in 40p, 600A in 20p. So as well as looking at contactors/fuses I want to review the busbars (which will need changing anyway to suit the new layout) to see what the current ones are capable of and how much bigger I would need to make them. Can anyone point me at the calculation I need to work this out? Google results are swamped by HV 3 phase busbars for factories - if these are okay great, but if there are EV specific tools about that would be very helpful.

:TLDR:
Does anyone have a link to an EV busbar sizing tool?


Thanks

Nathan

 

[MattsAwesomeStuff]

Does anyone have a link to an EV busbar sizing tool?

Not EV, but...

http://www.powerstream.com/Wire_Size.htm

Cross section of a busbar will depend on it's intended use, as you can see from the above chart, depending on whether wiring is bundled or not, there are wildly different standards. You can generally, greatly exceed either of those systems for an EV, but it gives you a ballpark.

One issue is voltage drop, if you care (generally matters for longer distance transmission).

The main issue is heat generated.

P = IV
V = IR

Substituting IR into V, we get

P = I^2 * R

You know I (current), you know R (by looking it up on the chart I linked), so you can calculate your power loss, and thus, how much heat you're generating and have to dissipate.

For example, let's say:
- you're pulling the full 600 amps that your batteries are rated for.
- by the time you account for routing, it's 20 feet of wire.
- you're using oh, 4 gauge wire.

Let's see how much power you're bleeding into the busbars.

P = 600amps^2 * 0.2485ohms/1000 feet * 20 feet
P = 360,000 * 0.005
P = 1789 watts.

That's a as much as an electric space heater on max... without delicate little wires and a fat to cool them.

That's as much as a spot welder left on constantly.

That's 2 horsepower just going to cable heat.

So, you can probably conclude, hrm, 4 gauge wire is too thin. Which is why I picked it for the example, because it's comically thin for EV wiring and the data shows that.

Maybe ballpark an amount of power you're happy with, and solve for R, which will tell you how thick of wire to use.

If you doubled up your 4 gauge, you'd only be using 900 watts. If you quadroupled it, 450 watts. And so on. It gives you a rough reference point for how much cross-section you need, as the relationship is inverse linear (double the thickness, half as much power loss).

 

[Emyr]

That's just the resistive properties of the wire though; EVs don't present a smooth load to the batteries, so the fact that wires are twisted means you need to consider the inductive properties of the link (or use laminated copper bars).

Whichever way you look at it, the shorter the better.

 

[kennybobby]

Consider the duty cycle also--smaller gauge wires can be greatly overdriven for short time periods with no problem. How often would you even be close to drawing 600 Amps from your pack? If drag racing it could be the entire 10 to 15 seconds, but on a road car it is not likely you would need to run it that way (except for showing off during joy rides).

edit: to remove incorrect data.

Since you have extra cells available you can make such a test on your own with just a few cells, some stove eyes, wire and a big-ass knife switch. Set it up with the resistance necessary to limit the current to 600A or whatever level you want, and an typical length of the AWG you want to test. Use a thermocouple, IR laser gun, a drop of water, or your finger on the insulation to determine how hot it gets over the desired time period. We would be interested to hear the findings.

 

[Hollie Maea]

This is based on testing where i put 40,000 Amps thru AWG 4 copper for up to 3 seconds with no wire damage.

This is not right. AWG 4 will fuse in fewer than 32 milliseconds with 40,000 Amps.

 

[kennybobby]

You are right and i have removed the reference. i should have looked it up before posting. Test was 3 cycles of 0.9 Hz AC at 4500 Amp peak current.

 

[ndg]

Thanks for the thoughts.

I like the idea of calculating the power loss, I think this is where factory and EV battery applications differ. I only want full current for a short time, max 10s, and my links will be short, so voltage drop and power loss much less significant.

Any idea where I can get 600A resistors ?

 

[kennybobby]

used stove eyes from the thrift store, put several in parallel. They usually sell for a $1 since nobody wants them.

So you have 20p that claims to be good for 600A, likely for only a few seconds if that much, since each cell would have to contribute 30 Amps.

But assume that is real, then you could just test 1 cell at 30 amps to see if they canl lsurvive.

 

[major]

...

Any idea where I can get 600A resistors ?

This is my load bank:

24 locomotive braking resistors. Each rated continuous at about 1.5-2 kw. Can configure for up to 1kA and ~600V. Have run at around 80kW for 10-15 minutes with strong fans. Big IGBT buck converter for control in Gray box.

For cheap, get a section of water pipe with two ground clamps so you can adjust attachment of 4/0 cables. If your tests are long enough to get the pipe hot, run water thru it. Obviously not safe for high voltage, but below 50, maybe 75Vdc, no worries. Just support on concrete or wooden blocks.

major