Fuse-wire for individual cells, DIY Tesla style

spinningmagnets

Moderator
Staff member
Joined
Dec 21, 2007
Messages
12,952
Location
Ft Riley, NE Kansas
Tesla uses ultrasonic vibrations to connect a fuse-wire from the nickel-plated aluminum-core bus plates to each of the positive tips of the cells.

file.php


I'm starting this thread to park information I find about using fuse-wire. This has to do with fire-safety, and is something that every battery builder should care about. The video below is a builder making a DIY Tesla-style power wall as a power backup for his home, using salvaged laptop cells. Laptop 18650's are all low-current and medium-capacity, they compete on lowest price only, not performance. However, if the pack is large enough low cell amps can still add up to high pack amps.

file.php


I don't like the idea about soldering onto the tips of the cells, and would prefer any one of the methods used in other threads around here (spot-welding, magnets, no-solder contact pressure from poron foam). Be that as it may, regardless of your preferred connection method...I have embraced the idea that the series connections (which carry the full pack current) should be thick copper (raw or nickel-plated), and the parallel connections can be very thin, so...it is intuitive that the parallel connections should be the fuse-wire.

also, I don't think the positive end and the negative end needs to be using an identical connection method. I think on the broad and flat negative, there is merit to using button magnets to hold flat copper ribbon onto the cell-end.

In this particular application, he used thicker wire on the negative cell-ends, and it is the positive ends that have the thin 0.25mm wire as a fuse. He is using magnet-wire, so the enamel has to be removed, and this builder merely heats the wire at the location of the connection point with his soldering iron (perhaps not the most professional method, but it seems to work for him). Your individual cells' current (when shorted) will determine the minimum diameter of fuse wire that is appropriate (which needs to carry the max amps of current for one cell without getting too warm...more research is needed to make an acceptable chart, to be posted here when finished), and this should be balanced against the diameter of wire needs to be so it melts rapidly when full pack current is used. Using 2X the minimum diameter might be a useful starting point (10A cont at the cell/ 80A peak for the entire pack, so...wire gets warm at 15A and melts rapidly at 20A?). This is just a suggestion, and it may be amended in the future as more information becomes available.

Fast forward to 15:31 to see the fuse-wire being installed (finished negative side shown first)

https://www.youtube.com/watch?v=xOVEnbbhbRY

[youtube]xOVEnbbhbRY[/youtube]

Fast forward to 7:30 to see the cell-shorting wire melt test

https://www.youtube.com/watch?v=5hiQCD8LPcs

[youtube]5hiQCD8LPcs[/youtube]

Fast Forward to 5:14, wire-melt test at 6:06 (actual melt at 8:16)

https://www.youtube.com/watch?v=raBWFsPlx7w

[youtube]raBWFsPlx7w[/youtube]

Damian Rene, FF to 0:20 for fuse melt test

https://www.youtube.com/watch?v=c7PGLFgkdf4

[youtube]c7PGLFgkdf4[/youtube]

The pic below is the battery pack from an Alta electric motorcycle

BatteryAlta.png
 
Very interesting, thanks ! How does the "fuse" works here, does it prevent total failure by shutting down only the shorted parallel group ? Never thought of this :D
 
There are several ways in which a battery pack might have thermal runaway event, leading to a fire. Having each cell fused, means that one type of cell failure could be eliminated as a possibity.
 
The wire looks fragile.
I don't know if it would ever make it on a home-grown bike.
He's doing packs that stay indoors and dont move or bounce or shake.
 
The wire doesn't have much mass so as long as both ends don't move in relation to each other, it should withstand lots of vibration. It works in a Tesla.
 
I've thought about this since learning about Tesla's application. It works for them, as they have the equipment to accomplish a weld without the heat inherent with soldering. I personally would not be confident in my ability to solder fast and well enough not to effect the electrolyte in some unquantifiable, yet permanent, way.

What I think we (DIYers) need is wide nickel strips that can be pulse welded, but have a thin fused section like these butterfly closure bandages:

11L757_AS01
 
That's a great suggestion Cal, that is what Damien Rene is using, a thin and narrow ribbon instead of round wire, which he found easier to connect [edit: 1.6mm wide X 0.1 thick nickel-plated steel ribbon].

The Tesla packs have a rigidly mounted bus plate, and the cells are rigidly held in place. I also suspect the slight curve of the fuse-wire (instead of being tight and straight) allows for a slight amount of flex (but not too much) which might result from vibration and also heat expansion / cold contraction cycles.

Here is a laser-cut P/S bus-plate made of nickel, which has integral fuses on the right half. If only the plate shown below is used, the pack would be 2S / 6P, the positive ends plus the fuses are on the right side.

JT9M6Jd.png


A pic from "insideevs.com"

159-350x303.jpg
 
After looking different DIY fuses packs, I came to realize that this is only suited for low discharge cells for big parallel packs.
At the levels I want to discharge mines (30A continuous to 60A seconds of bursts per cell), are there any fuses with compact sizes able to stress 60A without breaking?

The fuse indeed does replaces the direct link between the strips and the cell, thus breaks under overcurrent condition. Even so, that'll be a sh*t load of power rampaging before it breaks. But maybe I'm missing something and I'll gladly listen to correct my mistake.

Maybe the fuse must be adapted to the discharge ratings cause the rampaging cells would actually be "normal" under the 30/60A levels and start stirring things over this level?

Would you still recommend it in my case? Because I hardly see the benefits over a single big fuse over a parallel group :?: Don't know what to think due to lack of knowledge here.
 
Vanarian said:
After looking different DIY fuses packs, I came to realize that this is only suited for low discharge cells for big parallel packs.
At the levels I want to discharge mines (30A continuous to 60A seconds of bursts per cell), are there any fuses with compact sizes able to stress 60A without breaking?

The fuse indeed does replaces the direct link between the strips and the cell, thus breaks under overcurrent condition. Even so, that'll be a sh*t load of power rampaging before it breaks. But maybe I'm missing something and I'll gladly listen to correct my mistake.

Maybe the fuse must be adapted to the discharge ratings cause the rampaging cells would actually be "normal" under the 30/60A levels and start stirring things over this level?

Would you still recommend it in my case? Because I hardly see the benefits over a single big fuse over a parallel group :?: Don't know what to think due to lack of knowledge here.


You would just need to set the proper thickness to your particular application. Using this Wire Fusing Current Estimation Table, you'd see the different thicknesses that you can use with varying materials. Of course, I would test the size I would use first.
 
Vanarian, you mention 30A and 60A, but...I assume that is for the entire pack? If your P-groups are 5 cells, then each cell contributes 6A continuous, and 10A peak.

If there is a short inside one cell that allows the full pack current to pass through it, then that cells' fuse-wire will pass 60A, when it normally passes 6A-10A.

It is easy to experiment, and I would suggest for your pack: a wire diameter and material that is warm to the touch at 12A continuous, and blows quickly at 20A?

I am still learning, and i am posting all of this here to help me learn, and for others to help teach us.

________________________________________

If you are truly running 60A peak "per cell", then...a 5P pack would provide 300A. In that case, maybe try running fuse-wire that blows quickly at 120A? [edit: I am certain a shorted cell will flow all of the unrestricted amps that a parallel group can put out, not the amps of the whole pack]
 
If he's talking about 18650 cells it doesn't really matter because no cell is going to deliver that without cooking itself...

Fuses rely on a good differential between normal operating current and fault current. If you're trying to draw normal operating current near a current source's short-circuit current then getting effective protection from a fuse becomes difficult or impossible.
 
Punx0r said:
If he's talking about 18650 cells it doesn't really matter because no cell is going to deliver that without cooking itself...
.

.i suspect he is talking individual cell amperage.......i guess you missed this thread..
https://endless-sphere.com/forums/viewtopic.php?f=14&t=87968
And there is always the A123 18650 cell ! :wink:
 
He's talking 'per cell' and clearly typed the words.

If he's abusing his cells or whether they can take it is another topic for a different discussion thread. But with his requirements, there is a correct wire gauge equivalent that will give him the fuse ratings that he is seeking. Some forethought and testing required.
 
Vanarian said:
....... But maybe I'm missing something and I'll gladly listen to correct my mistake.

Maybe the fuse must be adapted to the discharge ratings cause the rampaging cells would actually be "normal" under the 30/60A levels and start stirring things over this level?

Would you still recommend it in my case? Because I hardly see the benefits over a single big fuse over a parallel group :?: Don't know what to think due to lack of knowledge here.

Fusing each individual cell protects the rest of the parallel cells against the failure of a single cell...such as an internal short..which could overload the entire group internaly even if you had external pack fuses.
They will also protect the cells agains an external short or overload.
 
Ok thank you guys, and thanks for the link! Should try a lower size to experiment the fuse at high amps.
Just to see how it breaks.

On YouTube I saw a comment from a guy speaking about thermo-relay fuse, what is that? I mean what is the difference between a properly sized "cable" and this?

Another question, when the failed cell is shorted by the fuse, does the related parallel group switch "OFF" or a second general fuse remains advisable just in case?
 
when the failed cell is shorted by the fuse, does the related parallel group switch "OFF"

If the paralleled sub-pack is a 5P and one cell-fuse melts, that string will suddenly become a 4P sub-pack, but...the pack will still run, with each of the other four cells having 20% more amps pulled from them. I have a few ideas about how to flag a blown cell-fuse right away, but...the important thing is for an internally shorted cell to stop thermal-runaway immediately, so the entire pack doesn't go up in flames.

I have quite a few pack-building ideas I am keeping to myself until I can see if they are smart or stupid (the safe bet is on "stupid"), but...when it comes to fire-safety, I want to be open-source and if Tesla has a good idea, I want to shout it from the rooftops to everyone.
 
Individually fused cells is about the only way to protect against a cell that shorts internally. I don't think this kind of failure is very frequent but it does seem to happen. I think Tesla was more worried about pack damage from a crash. Bikes crash too, but the packs probably less prone to getting shorted from it.

Where using large format cells in a 1p configuration, I don't see much value in individual fuses. A single mid-pack fuse seems to be adequate.

I think you could test fairly easily by setting up a single crash test dummy cell with the fuse link and try shorting it to see how long it takes to blow (please video). Then test the link at the maximum expected current to make sure it doesn't blow or get hot.
 
I think a useful visualization exercise, is to see a 5P string in your mind, and then suddenly swap one of the cells for a fat copper wire (representing an internally shorted cell, from a wreck, or an overheat situation). All of those cells will now flow from one side to the other, so that fat copper wire is flowing FULL 5P cell amps, with no BMS to limit the max amps.

Now, visualize that fat copper wire being connected to the positive bus side by a short and thin wire. Once that thin wire melts?...the unregulated flow of all the 5 cells in that P-string is stopped. At best you'd have to replace that single bad cell...at worst you'd have to replace all five cells in that P-string (which I would recommend). But, the best reason to do this is...the pack doesn't go up in flames...

Even if you end up throwing the entire pack away...your choice is buying a new $600 pack, or having a lithium fire in your home.
 
Not only internal cell shorts.
In any situation where a short between cells in a pack (loose wire, abrasion between cans, dropped tool, iinsulation failure, accidental damage, etc etc) ...anytime that happens on the pack side of the bms/system fuse, then fusable cell connections will protect the pack from overload type failures.
It would still be a disaster needing a full pack rebuild, but should prevent the worst case thermal event.
 
Hello,

I built my kick scooter pack exactly like in the title.
6S4P Pana PF cells. Max. current draw (in line with cell capability) is 40A. Added benefit of the fuse wire is the lowest possible solder heat impact on the cells negative terminal.

kickscooter_battery2.jpg
kickscooter_battery1.jpg

T.
 
Takyka, do you feel that the negative cell-end is more sensitive to damage when soldering? I had assumed that both ends were equally fragile, but any evidence that the positive cell-end can take more heat without being damaged would be useful info to me.
 
Spinningmagnets,

I'm quite sure. At least in case of PFs. The positive terminal "button" is not solid one, but the contacting surface is connected to the main body by the means of three tabs equally spread along the circumfence. Very small mass to heat up and cool down during soldering, so it is a fast process. +the heat has to travel along a relatively long path.

I was reading a lot from here before I took this route. Somewhere I found the most destructive is the soldering on the midle of neg. terminal. (That's why I soldered to the edge there)


T.
 
spinningmagnets said:
Takyka, do you feel that the negative cell-end is more sensitive to damage when soldering? I had assumed that both ends were equally fragile, but any evidence that the positive cell-end can take more heat without being damaged would be useful info to me.

Have a look/read of this Sm...
Screenshot_12_1024x1024.jpg

http://www.lygte-info.dk/info/battery%20protection%20UK.html
lg_cylindrical.gif
 
The PFs has no extra protection built in.

Looking like this:
http://www.bazaargadgets.com/image/cache/catalog/products/lights/flashlight/4PCSNCR18650PF37V2900MAHRechargeableLithiumBattery-SKU4x105910-4975-800x800.jpg
You can see how the cap is made...

T.
 
Back
Top