Voltage sags too much!

[rg12]

Hey Guys,

Finished this snow bike build with a smart controller that shows all the info via bluetooth on my smartphone.
The pack is 16s 26Ah with a 50A continues BMS and the cells are LG M26 rated for 3.9C
The problem is that the power cuts off and comes back when I release the throttle because the voltage sags below the 53v limit that I put in.
It sags from 57v to about 47v (when I lower the cut off voltage (it's a bluetooth programmable)) while pulling 50A (the BMS doesn't cut exactly at 50A you can even pull 55A constant and it doesn't cut off).
Do you have any idea how is it possible to have a 26Ah pack with 3.9C cells drop 10v under 50A?
Something weird happens when I move the 3 speed switch from first speed to 3rd, it still pulls 50A but it drops the voltage to about 15v down!
What is wrong with my pack? I have built smaller 16s packs with the same BMS and same cells for much stronger rigs and never had a problem.

 

[redilast]

Do you have a clamp meter or some way to measure actual current draw?

What kind of cell conductors did you use to make the pack? Thickness and type of nickel?

Looking at a chart of a similar 10A rated LG cell on http://lygte-info.dk/review/batteries2012/Common18650comparator.php I can see that within a minute or two the cell would sag from 4.20v to about 3.85v. So for a 50A draw on those cells I would expect a 5-6v drop + any additional voltage drops due to the resistance of the other conductors and BMS.

Oh and you did mention for a snow bike. Temperature is also a huge consideration. Colder temperatures + Li-ions do very poorly until they are heated up. So cold temperatures + expected pack voltage drop and you could easily lose 10v-15v depending on temperatures and other pack and bms voltage drops due to high resistance connections.

 

[rg12]

Wow thats alot, so I guess it's within spec.
It's not cold here so we can cross that off the list.
It didn't sag in a minute or two but right away.
The welding material is nickel plated 8mm wide 0.1mm thick.
How critical is it to have pure nickel?

 

[rborger73]

You can still pull 50A with undersized power wire. What gauge wire are your main leads? Are there any bottlenecks in your wiring?

 

[rg12]

Using 12g wires, very short into an EC5 connector soldered like a pro.
I limited the current to about 38A and at half charge the pack drops from 60v to 48v while pulling 38A.
It seems like a huge drop for me.

 

[redilast]

Using 12g wires, very short into an EC5 connector soldered like a pro.
I limited the current to about 38A and at half charge the pack drops from 60v to 48v while pulling 38A.
It seems like a huge drop for me.

Does it drop from 60v to 48v and stay there? Because it might not be able to limit inrush current which could be easily 2-3x operating current for a fraction of a second.

Also any pictures of your pack so we can see the way the nickel connections are made. The 8mm * 0.1mm nickel plated steel I'd say is not an ideal choice above about 5A-6A or so. Is there a nickel tab going from one cell to the next for every cell in series or did you skip some cells for the series connections?

 

[kdog]

How many series connects?...0.1mm NPS is a Poor conductor... Not good for a 50amp draw.

 

[kdog]

I just checked nobuo's guide in the repository... 0.1x8mm NPS comes in at </=2.5amps optimal, >3amps equals heat. So hopefully you've got about 20 strips for your series connects for 50amps.
I'd be looking at unpacking your battery and observing/testing it under load.

 

[999zip999]

How about pics of your naked battery ?

 

[liveforphysics]

Cold cells sag a lot. The voltage sag amount will double for about every ~10degC cooler you go.

 

[rg12]

I don't have photos of the naked pack.
The series connections are done with 2 tabs.
I built a pack for a guy that is a 16S 20.8Ah from the same cells and the pack is connected with 4 tabs in the series connections (because his pack is split to two) and he just sent me a video of his amp meter while riding showing that the voltage sags from 74v to 64v almost constant at 50A draw.

* The series connections in the first stages of the pack would have much lower load than those in the end of the pack...

It's weirs saying that I need so many tabs in the series connections...why? take two pcs and roll them tight if you can, they get as thick as a 12-14 awg wire conductor.
Using 20pcs?! that will equal the thickness of a huge wire which seems not logical.

 

[Punx0r]

* The series connections in the first stages of the pack would have much lower load than those in the end of the pack...

It's weirs saying that I need so many tabs in the series connections...why? take two pcs and roll them tight if you can, they get as thick as a 12-14 awg wire conductor.
Using 20pcs?! that will equal the thickness of a huge wire which seems not logical.

Current is the same through all parts of a series circuit - it has to be.

Steel is about 10-time less conductive/more resistive than copper.

 

[Ohbse]

If you're attempting to pull 50a through two strips of 0.1mm nickel plated steel then you will probably end up on fire. Those links will be getting *hot*.

Steel is a terrible conductor relative to nickel which is terrible relative to copper, I suggest you read substantially more of the discussion around pack assembly and the capacity of various materials. As a general rule I aim for 1.5mm2 of nickel equivalent conductor per cell in terms of series connections. You have effectively got 0.10mm2 of nickel per cell.

You should rebuild your pack, or at least enhance considerably with additional pure nickel conductors. That will improve things considerably.

 

[kdog]

NPS is a shit conductor that's why. At 25 amps each strip I'm surprised it's not on fire already- and I mean that. Recall that one you built for a mate.
I recently did a test on 10mm x.2mm PURE nickel strip and it was too hot to touch at 30 amps. Yours would be nearly red hot at that current ( the nickel was just starting to glow at about 60 amps)
That nickel strip is prob able to handle 6 time the current of your 8x0.1mmNPS.
Think of each P group as a single cell connected by a series of wires. Anywhere along those wires there is the same current flowing, you just measure it at one end. Break it in the middle 50amps, other end 50amps etc
So the series connects take the flow from all the P cells and pass it on to the other.
Your pack is thermally rated at 5amps continuous.
I would unpack it and fix it before it destroys itself.
K

 

[rg12]

Just unpacked it and will add more strips on top of everything.
The question is, It's a triangle pack meaning that due to cell order I can't add one tab for each cell in series because the cells look like:
Example 4 in parallel (* = negative and @ = positive)
* * * * @ @ @ @
so the connection is between the last * and the first @
the question is, can I weld a few tabs on top of each other in the area where the series connection is? or should I weld one long tab on top of all those 8 cells in the example (each of the two 4 cell rows are welded from both sides already).

 

[kdog]

Well in your configuration I'd be tempted to solder multiple copper wires from cell to cell ( well between them). Failing that I would firstly ditch the NPS and get pure nickel. Then I'd make sure I had at bare minimum 8 (1.5mm^2) strips of this for the series connects but I'd be happier with about 12... But 16 (1 for each cell) would be best. These need to be evenly distributed amongst the cells for equal current sharing.
How you achieve this you'll need to figure out from group to group. Pics could help with some opinions.
I'll try again to upload my vid of this test I did on a poor piece of nickel- my phone doesn't like YouTube for some reason.
I'm really glad you have investigated this

 

[rg12]

Where can I find copper strip? looked around and found only copper tape and copper sheets that are very wide.

16pcs for 10 cells in parallel just to connect them in series? that's more than one tab per cell.

Just ordered 0.15mm 8mm pure nickel (so they say, will test like said in google when it arrives).

I think that for the side by side connections I will put a two strips over all the cells from the previous example.

EDIT: Nobody mentioned anything about voltage through the parallel connections because heat must change (because of higher or lower wattage) due to different voltages.
For example 50A at 10v is not as bad as 50A at 100v
kdog: at what voltage did you get that tab red hot?

 

[Hillhater]

Hey Guys,.......
The pack is 16s 26Ah with a 50A continues BMS and the cells are LG M26 rated for 3.9C
Do you have any idea how is it possible to have a 26Ah pack with 3.9C cells drop 10v under 50A?
.

Just to say......there really is no defined linkage between "C". rating, and voltage sag..!

 

[kdog]

Sorry I thought it was 16s. 10 then!

 

[Hillhater]

EDIT: Nobody mentioned anything about voltage through the parallel connections because heat must change (because of higher or lower wattage) due to different voltages.
For example 50A at 10v is not as bad as 50A at 100v
?

...No.....amperage is the defining factor.
...if you apply more voltage to the same resistance, it will increase the amps. !

 

[rg12]

Sorry I thought it was 16s. 10 then!

It is 16s but it has 10 cells in parallel for each cell in series to create the 26Ah capacity.

When you did your test, with what voltage and for how long was it until you produced the heat that you talked about?

 

[rg12]

EDIT: Nobody mentioned anything about voltage through the parallel connections because heat must change (because of higher or lower wattage) due to different voltages.
For example 50A at 10v is not as bad as 50A at 100v
?

...No.....amperage is the defining factor.
...if you apply more voltage to the same resistance, it will increase the amps. !

So you say that 10v at 50A and 100v at 50A will create the same amount of heat?

About the c rate / voltage sag relation, I think there is a big relation since I see 20Ah packs with 4C sagging 8v easy while my 10Ah 25C LiPo pack at the same voltage sags about 3v

 

[kdog]

Err sorry it's getting late yes you're right 10P...
Heat is a function of i^2R voltage is there to drive current through R so it's not directly related to heat production. The voltage I needed to get my nickel tab just starting to glow was 22v at 26v it was about to melt (86amps) power was controlled by a ( funnily enough) nichrome wire array to radiate heat. At 2000w it was red hot and toasty!

 

[rg12]

Err sorry it's getting late yes you're right 10P...
Heat is a function of i^2R voltage is there to drive current through R so it's not directly related to heat production. The voltage I needed to get my nickel tab just starting to glow was 22v at 26v it was about to melt (86amps) power was controlled by a ( funnily enough) nichrome wire array to radiate heat. At 2000w it was red hot and toasty!

Well I know that you can produce less heat by dividing the amps and doubling the volts (also equals more speed on ebikes).
There is a big difference between 40-50A to your 86A

 

[kdog]

True but at 30 I couldn't hold it, and it is pure nickel with twice the cross section.
It took about 5secs to glow at 22v and maybe 2secs at 26v.
In you're example R would have to be different, and yes the heat production would be different 500w vs 5000, but it's not a fair comparison because R is different. This would require a different amount/type of material you're testing.
Going to hit the hay :wink: