Cylindrical cell fusing and fuse calculation

[agniusm]

Perhaps someone has information or links on how to calculate fuses?
I would like to start designing a battery for myself and want to do it properly this time. I want to fuse each cell, then each string of paralleled cells and a fuse on the whole pack so no external fuses are needed
So how would one go about calculating material size for given discharge rate?
Lets assume that we use nickel sheet which is .20mm thick and we need 50A discharge on the pack. What size fuse i would need for:
a. single cell
b. string of cells
c. the whole pack
Thank you

 

[amberwolf]

To determine some of what you want (A) wouldn't you need to know how many paralleled cells you are using, so the current can be divided amongst them?

Then for B and C I'd think that the fuse would be the same size, if by "string of cells" you mean a parallel group that represents a whole "1s" group of the pack.

 

[agniusm]

Thanks. Well if you need to divide current between paralleled cells that would suffice for the answer. How about series connection? What about material sizing for given current?

 

[t3sla]

Relevant
http://www.electrochem.org/dl/ma/203/pdfs/0189.pdf

 

[agniusm]

Yes, but according to this http://www.ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20100037250_2010040765.pdf it doesn't help too much

 

[solar.g]

You stated earlier:
"Lets assume that we use nickel sheet which is .20mm thick and we need 50A discharge on the pack."
Are you proposing that you cut nickel sheet of that thickness to a particular width to be a fuse for each cell?
A fuse is actually a complex device, working on instantaneous and sustained currents, through a range of temperatures.
Fast-blow, slow-blow, vibration-insensitive, etc...
Many circuit breakers are rated for one current, but don't actually "break" until 30% or 50% above that rating for a sustained over-current (eg: 20A DC, breaks @ 26A after 30 seconds).
Why not just buy fuses or circuit breakers made by someone who has the production process worked out (and certified to work properly) and connect them up?
Reminds me of an old anecdotal story of people replacing their household (110V or 240V) fuses with rolled-up-foil from cigarette packets (guess how well it worked).
Yes, you could (theoretically) calculate what width of .20mm nickel sheet could be an appropriate fuse, but could you ever make high numbers of them well enough that they consistently did the job in a safe, predictable, way?
And when your home-made fuse "blows", does it's molten metal splatter all over the place and produce short-circuits on your other otherwise perfect cells?
Just buy commercial fuses and connect them in a way that's easily replaceable when/if they blow.
Just my 2cents worth....

 

[agniusm]

Well, i'm thinking laser cut tab terminals within .001mm tolerance. Should do fine, my guess. I would not know about going and fusing each cell
The way member Circuit had designed :http://endless-sphere.com/forums/viewtopic.php?f=14&t=55870&p=832320. Don't want to copy/paste, rather get behind all this fusing idea and perhaps come up with something of my own 2 cents

 

[Matt Gruber]

i use cells with a built in thermal fuse. IT WORKS! i tested them. Konions found in makita tool packs. LiMn

 

[agniusm]

Yes, i have considered those fuses but even if they costed .50ct it would add 100$, they were .80 -1$ when i looked, on top of you pack you'd still need tabs. Inefficient

 

[Hillhater]

Since you are thinking of "fuseable cell tabs" ,....... i assume this is a last line of defense against major pack problems... short circuits, etc ?
IE you dont necessarily need to know precisely what section of tab is needed for whatever current you decide is the limit,..say,.. 25 A...you just need a fuse to protect against overload current.
Whatever dimensions you calculate, or are advised, you will need to prove by testing anyway,
...so its probably simpler to just do a few tests with sample strips to establish the characteristics of your specific material .
I would start with a small (1.0mm ?) section and see how much current blows it. Then work your way up in size progressively.
Repeat each test a few times to confirm/average results and see how predictable it is.
For series strings ( at pack voltage) , i would go with proven commercial fuse's.

 

[davec]

i think this is a good idea
for safety all cells connected in parallel should have a fuse- without a fuse the other cells in the pack can overload the bad cell and cause a fire - if you have a good fuse this can blow reducing the effects
i would do as hill mentioned, cut the tabs thin and run current through to see how long it takes to react- making your own special tabs is the way to go- buying all those fuses and installing them is not worth it
http://mit-evt.blogspot.com/2009_06_01_archive.html

 

[friendly1uk]

Sizes of tinned copper wire for use in
semi-enclosed fuse carriers


(A) ............... Nominal diameter of wire (mm)
3 ................. 0.15
5 ................. 0.2
10 ................. 0.35
15 ................. 0.5
20 ................. 0.6
25 ................. 0.75
30 ................. 0.85
45 ................. 1.25
60 ................. 1.53
80 ................. 1.8
100 ................. 2.0


You would have to design around short circuit currents, as simple overload protection can't be done. These wires should be in a semi-enclosed ceramic holder due to the heat involved with normal use. If you want them to stay cool, they would have to be way oversized for overload protection, but could still operate if there was a short circuit current flowing. Even then, you just melted metal. It was hot.

All these wires come on the roll. Search 'fuse wire' and perhaps link your cells with it. I doubt it is wise though. Thermal protection will be more work than fuse sizing.

Edit: Having trouble with post layout. Looks fine till I post, looks fine in edit window, but I'm meant to have two columns and I just see one...
Edit2: Hmm...... Better with lost of dots, but don't you just hate forms that won't recognise spaces as requirements

 

[ridethelightning]

very interesting, just bookmarking this

 

[circuit]

Yes, but according to this http://www.ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20100037250_2010040765.pdf it doesn't help too much

Wow, page 13 really shows what I was nagging about all the time..

 

[spinningmagnets]

I really like Teklektiks electrical system, but I'm not going to pretend I understand all of it (He's an electrical engineer):

http://www.electricbike.com/custom-build-gallery-teklektiks-dual-motor-yuba-mundo/

"...used a 50A magnetic marine breaker for the main power disconnect switch (Blue Sea model #7230). These types of breakers are very rugged and are intended for high-humidity/salt-air use. They are designed for up to 65V-DC and can handle interruption currents up to 7500A making them virtually indestructible in this application. “They start to trip at 62A and with a couple of controller shorts, the CycleAnalyst never recorded a spike greater than 180A before the breaker switched off”...

 

[LSBW]

Check out pages 43 & 44 of the cylindrical cell battery pack design guide. Bottom of this page:

http://www.buya123batteries.com/ANR26650_Lithium_Ion_Cylindrical_Cell_p/300475-305.htm

In Figure 10, the 7 mil Nickel strap material is necked down to 3.6 mm, and clears at approximately 2100A in 0.1 sec.

 

[friendly1uk]

It would be a bit rough, but a typical AC MCB will see DC currant. A 32A 110v MCB might trip around 60 amps using about 70v DC, but that would offer some short circuit protection.

 

[agniusm]

Had couple of hours, so i went and made little jig to test some fuse ideas. Here is a quick arround:
[youtube]NPMvMWHDufg[/youtube]
The temp in my workshop was at 8 degrees C. I have mainly tested on old laptop cell:

Last test was done with this one:

Nickel tab i had at hand was .1mm and 3mm wide.
First test was with 1.5mm tab:
< Most videos capture at 1/8 slow motion. You can calculate the time takes for fuse to blow up by dividing time by 8 when frame rate is 15fps>
[youtube]4B4Ci6ROHW4[/youtube]
Second go was with 3mm nickel strip, done it twice:

1st attempt:
[youtube]TX6g9y_HPgs[/youtube]
2nd attempt:
[youtube]8yH10DBOkYo[/youtube]
Third was 1.6mm nickel strip:
[youtube]0kmL8e2-20Q[/youtube]
Then i figured out that i should short the cell:
First old laptop cell
[youtube]j_H0C7rpFRQ[/youtube]
2nd new LG 2600mah cell
[youtube]DTTMs3GeAb0[/youtube]
I have not noticed any significant temp rise. Most of the heat was concentrating at the top of the cell (near positive side). I was expecting cell to vent, but that did not happened. I will try next time to use a fuse and overcharge the cell significantly.
Any suggestions, shout!
Peace out