Sag from cells or from heat/resistance?

E-HP

10 GW
Joined
Nov 1, 2018
Messages
6,201
Location
USA
Something I plan to do this winter is open up my UPP hard cased triangle pack to check each parallel group voltage, then inspect the cells and construction. I have two UPP packs and both have worked well, but I've seen posts about the construction of some of their packs, so I'm curious to see what's going on inside.

Since some of the posts point to lack of sufficient nickel strips for the amount of current that needs to flow, my question is related to how that might affect voltage sag. I get a decent amount of sag when pulling current at or above the continuous rating of the cells when accelerating, and some sag below that level of discharge. I assume most of that comes from the cells, but wonder how much sag contribution the conductors add, if insufficient. In addition, resistance goes up as the conductors heat up, so maybe the impact is magnified under higher loads.

I've been researching spot welders for a future battery project, but if there's an ability to improve my existing pack performance, while getting some practice, I'd move that purchase up higher in the queue.
 
My thought is sag is about the cells as a rule. Wire and strip size aren't changing to cause the sag in available voltage. It's way more about the cell's ability to produce a constant rate. Some are clearly better at this than others.
 
AHicks said:
My thought is sag is about the cells as a rule. Wire and strip size aren't changing to cause the sag in available voltage. It's way more about the cell's ability to produce a constant rate. Some are clearly better at this than others.

Thanks. Ya, I found some data on the strips and it's just noise as far as resistance goes. They're 8A cells, but happier up to around 5A, so they sag 3 or 4 volts when I take off and pull ~60A from the pack. The pack is 8P, but works best staying under 40A (which is what the BMS is rated for anyway).
I'll still check to see if there are any obvious bottlenecks.
 
E-HP said:
but wonder how much sag contribution the conductors add

Why not solder or somehow connect a pair of tiny wires to either end of the suspect conductor and measure voltage drop under load? You'll have a definitive answer.
 
Comrade said:
E-HP said:
but wonder how much sag contribution the conductors add

Why not solder or somehow connect a pair of tiny wires to either end of the suspect conductor and measure voltage drop under load? You'll have a definitive answer.

I think it's just math. If you know the resistance of the nickel strip, and how many strips connect the parallel groups, then you just input the pack current to get the drop.

So even though I could find data on nickel strips, there wasn't anything I could find about how the resistance changes as the strips heat up (presumably when their current limit is exceeded), so I'm not completely convinced that beefing up the strips won't help much. I think a lot of folks may have done an experiment like this in science class growing up, as an example:
https://nationalmaglab.org/education/magnet-academy/watch-play/interactive/heat-resistance

I also think of examples using power tools around the house with extension cords. If you plug your drill into the outlet or into a 100ft extension cord, the no load speed is very similar. But when you start drawing a lot of amps, the one plugged into the outlet will drill right through a big block of wood with a spade bit, but with the extension cord, starts to bog down due to sag. I never really checked to see if the sag continues to increase as the extension cord heats up though.

If the resistance of the nickel strips remain close to the same as they heat up, then that could be an issue. If they don't, then based on the specs, they wouldn't contribute much to sag with my anticipated current output.
 
E-HP said:
So even though I could find data on nickel strips, there wasn't anything I could find about how the resistance changes as the strips heat up (presumably when their current limit is exceeded)

Well, they always heat up by definition if current is flowing through them. The question is just by how much. Current carrying derating due to temperature rise would never ever enter the picture in the context of nickel strips in a battery though. It would be incredibly negligible.

nickel.jpeg
 
Localized heating of the nickel interconnects can be more problematic than the total energy loss would suggest. For example, there's not much energy in the ember on a cigarette, but it's a problem if it contacts your skin.

Because the welds and nickel strips locally contact vulnerable parts of the cells, they can damage the cells disproportionately to their power loss, if they build up heat.
 
Comrade said:
Well, they always heat up by definition if current is flowing through them. The question is just by how much. Current carrying derating due to temperature rise would never ever enter the picture in the context of nickel strips in a battery though. It would be incredibly negligible.

nickel.jpeg

I think it depends on the configuration and construction of the pack. Using this simple 20S1P pack of Samsung 25R cells, pulling the 20A rated current as an example. I think I saw a table with various nickel strip resistances and that 7 mOhms seemed to be a good approximation, but hoping someone knows more:
Not confident in my math or specs, maybe an independent check would help

strips.jpg

I have a typo, should be 20S1P, not 1S20P
 
The resistance of the strip (using cross section to calculate) IMHO would be less important in this equation than the welds between the strip and battery terminals. And since spot welding is not standardized in any way, this is the wildcard that would make any mathematical estimation model pretty inaccurate.

If the battery is open on a bench, you could make a few measurements with a dummy load and isolate the resistance of the strip + weld, then multiply your real number by the number of welds. I doubt after it all you can do anything practical to improve total pack voltage by concentrating in the strips. Any possible improvement IMHO is negligible compared to internal resistance of cells themselves.
 
Any possible improvement IMHO is negligible compared to internal resistance of cells themselves.

The wires are in like picoohms. Nil. Negligible.

Unless they reach ' red hot" color. Then the resistance skyrockets.

For instance, I built a 20 cell pack, of 1mOh cells. The cells were all exactly 0.001 Ohms each... .

Measured 23mOh. The entire run, of four feet of 4 gauge, wire, and controller, connections, the distribution block, the ANL 450A fuse, through the shunt, and return lines. 3mOh.. AKA 0.003 Ohms. Pure copper ( 'cept the shunt,) with silver plated lugs... All around.

The pack sagged 5-7v on full load, 250A peak... 18kW. 23 horsepower.

The sag was from the cells, not from the wire in between. The 20 mOh of cells... not the 3mOh of wire copper inbetween.
 
DogDipstick said:
Unless they reach ' red hot" color. Then the resistance skyrockets.

But when does that occur? If a nickel strip of certain dimensions is rated at 7A for instance, and you run 25A through it, then what how would you know what the temperature rise would be. I think most conductor ratings are determined with temperature as one of the parameters.
 
E-HP said:
rated at 7A for instance, and you run 25A through it, then what how would you know what the temperature rise would be

Temperature rise depends on dissipation into the environment.

0.007ohm @ 7A = 0.343 watts

0.007ohm @ 25A = 4.375 watts

That is a big difference. First one should be perfectly fine inside a case with no air flow. The second would get pretty hot in a few minutes. You could stick a small temperature sensor on it.
 
Back
Top