18650 spot welding -how to- ULTIMATE REPOSITORY

I have both the spot welders Allex suggested and have been spot welding with a 3S5P lipo battery I made some years ago with old cells from an e-scooter and I still be able to make battery packs for local buddies.

flippy said:

jonescg said:

So I have ordered a Sunkko 709 AD

Click to expand...

please cancel that order as fast as humanly possible.
that thing is nothing but grief.

Click to expand...

Eggs Ackley!

Although my early version KWeld is a turd too. Jumped to soon and updates beyond my skill level make it a $100 visually interesting paperweight. My JP Welders rock on.... Listed for sale, the KWeld, without a single bite. In fairness there is an offer fro me to send back, $25, pay for upgrades, but I understand not all the way to the latest iteration, for additional fees. Look. It’s all good, but for a fella that just wants a welder that works... JP.

Whats the problem with the early K-weld?
I have a relatively recent one and although I've only done a few test welds so far it seems to perform exactly as advertised, ie works great!

Taswegian said:

Whats the problem with the early K-weld?
I have a relatively recent one and although I've only done a few test welds so far it seems to perform exactly as advertised, ie works great!

Click to expand...

Protection components need updating, et al.
https://endless-sphere.com/forums/viewtopic.php?f=14&t=89039&p=1349912&hilit=kweld#p1349748

Well my Sunkko arrived and it seems to have come with two hand-held probes. I guess after one heats up too much you can swap leads while the other cools down

OK guys so i've just built a B52 Bomber battery 22S10P using LG MJ1 3500 mAh cells
i've used 5mm wide 0.2mm thick solar panel bus tape - its tinned copper.

scotchbrite the tops/bottoms of the cells and tinn them with solder and 80W iron, takes less than a second using fairly heavy gauge solder.

result - far quicker than spot welding nickel (I have a 12V battery type spot welder)
resistance vs 8mm wide 0.15 Nickel is 3x the current carrying capacity.

Bomber pulls 75A max so 7.5A / cell draw which is well within the 10A continuous they are rated at.

After a 40km ride i've eaten through 30AH of the 35Theoretical per the Cycle analyst, battery warm to touch - motor well not so much and whilst I haven't melted the 5304 yet I think the possibility exists.

In all at $11 for 10m of the tinned copper bus bar strip I was just a little short and needed about another meter of it.

have done a botch and connected the end cells to the bike wiring via copper braid and still waiting for the BMS, i think its a viable solution for a home builder vs spot welding which using nickel IMHO is an inferior solution for current carrying.

sag at 6000W is only 3-5V depending on the pack voltage to start with.

Obviously soldering nickel is nigh on impossible so for both performance enthusiasts for the 3x current carrying capacity and home builders without spot welders alike I think the tinned copper is a great solution.
tables tell the story for me
https://www.endless-sphere.com/forums/download/file.php?id=248217&mode=view
Karl

I have been looking at man of these tables for current rating for nickel online and they all seem to more or less line up.

I see the odd mention that people think they are way over-provisioned, but I figured the groupthink was probably correct.

However, I've been looking at the batteryhookup fused nickel strips recently, and they say these carry 3A from each cell continuously with no voltage drop due to heat.

By this (relatively unscientific) measurement I made, the width of this 'fuse' wires appears to be roughly 1.1mm:



I can't actually find out how thick the strips are, but as they say nothing about needing a particularly capable welder I would have assumed they were .2mm or below. But even at .3mm a 1mm wide strip doesn't seem to match up with most of the tables I have seen online at all.

I was just wondering what the calculations behind these tables are and why the batteryhookup strip seems so out of whack with these calculations.

I assume, they where behind the 8A instant blow value and considering nickel, which is not the best fusing material, with its high melting temperature, all ended at this low cross-section.

Assuming:
Nickel 1mm2 R at 20C =edit68Ohm/1000m
Link cross-section =0.2mm2
Link length =20mm
Current =3A

We get:
V droop/link = 0.021V
Heat dissipation/link = 0.063W

Not ideal but acceptable, if you plan to stay below 3A

Thanks, that makes sense.

Is that 'acceptable' the same as the 'acceptable' in the OP, so for example 0.1mmx5mm = 3A acceptable?

No_Shorty said:

Thanks, that makes sense.

Is that 'acceptable' the same as the 'acceptable' in the OP, so for example 0.1mmx5mm = 3A acceptable?

Click to expand...

Not that much of acceptable, here we have a "nickel" fuse link that must heat up to 1500C to blow, OP is calculating nice cool conductor.
It is acceptable as fuse link (a would prefer Zink with low melting point<500C), remember, it is made to fuse high resistance modem cells with 1C discharge rate and max load current of<5A, 8A instant blow rating must be about dead short rating of those cells.
I would use them in application with 3A peak load per cell and trying to operate<2A/cell.

This video here: https://www.youtube.com/watch?v=BAPHF3Sq2t8&t=519s

He tests them under a thermal camera, and at 3a they keep a pretty stable 55deg f, which seems to basically be ambient. It blows at around 8-9a and at 626f.

So even with a nice cool conductor, this 1mm wide strip seems to be perfectly acceptable at 3A, considering the lack of any temperature change at this current - I cant really see why it wouldn't even be classed as 'optimal'.

I would use them in application with 3A peak load per cell and trying to operate<2A/cell.

Click to expand...

Indeed, that absolutely seems to be their use, although a higher current version is coming. But if a 1mm nickel strip is actually perfectly adequate then recommending 5-7mm+ seems pretty heavily over-provisioned (nothing wrong with that, over-provisioning is a sensible safety precaution for any amateur build), and this seems likely to follow for higher current if the maths is the same going up the table.

No_Shorty said:

He tests them under a thermal camera, and at 3a they keep a pretty stable 55deg f, which seems to basically be ambient. It blows at around 8-9a and at 626f.

Click to expand...

Nice video.
My concerns rises a nickel fuse that is glowing white and liquefying at polycarbonate (plastic) melting temperature instead of seeing 2500f heat spots. V loss also seems totally wrong.

No_Shorty said:

this 1mm wide strip seems to be perfectly acceptable at 3A

Click to expand...

As fuse maybe. Do not forget, loses will grow exponentially, at 6A you will deal with 0.042V droop and 0.252W in loses. Usually packs are enclosed, links are not open air and only way to cool is to push heat in to the tab and cell tab adding to cell heating from internal resistance.
If you get some more data, please share.

Thanks for your thoughts, I agree that if the fuse blows it may damage the plastic - but if this happens the whole battery needs tearing down and cell replacing in any case.

Just had a reply from Battery Hookup saying the strip is 1.5mm.

- I'm designing a pack at the moment, and the extra safety of having a fuse at the cell level appeals to me. 3A is too low for me needs however so I may simply cop the 'Tesla' method and weld an actual fuse wire to each cell and then to a bus bar.

No_Shorty said:

Thanks for your thoughts, I agree that if the fuse blows it may damage the plastic - but if this happens the whole battery needs tearing down and cell replacing in any case.

Click to expand...

What I mean is, his thermal camera reading have no value if they show 626F nickel melting temperature.

No_Shorty said:

I may simply cop the 'Tesla' method and weld an actual fuse wire to each cell and then to a bus bar.

Click to expand...

Be careful with fuse wire material, Tesla uses some special fuse material, which has high electrical conductivity, low melting temp and low heat conductivity.
Because of those qualities and cost, Zinc is most popular commercial fuse material.
Copper is to heat conductive and high melting point, I could have a link hold 30A with good heat sinking on the ends and then
fuse same copper link at 3A with no sinking.

Ah is see,

I'm assuming you could just use a PCB fuse of your chosen current between the battery and bus bar?

KarlJ said:

OK guys so i've just built a B52 Bomber battery 22S10P using LG MJ1 3500 mAh cells
i've used 5mm wide 0.2mm thick solar panel bus tape - its tinned copper.

scotchbrite the tops/bottoms of the cells and tinn them with solder and 80W iron, takes less than a second using fairly heavy gauge solder.

Click to expand...

Interesting! Would you have a link to the kind of solar panel bus tape that you used? Thanks!

just a thought,has anyone tried using cut down stainless steel spokes as electrodes in a battery tab spot welder for welding nickel strips to 18650 cells?

Why? Copper is pretty easy to get..

Stainless steel would get hot very fast, it would be so hot it would glow.

Best electrode is fat copper that is welding thin nickel. For a high volume assembly line, it's possible to run non-conductive fluid through copper tubing, with a replaceable threaded copper electrode tip.

It's not technically difficult to do in your garage, but I'd rather just take an occasional break and have a sip of some refreshing beverage, to let my probes cool down...

Yeah speaking of hot tips, I partially solved it by switching to lower resistance wire (6AWG silicone) and running small fan pointing at the welders (JP Welder) mosfets. Now it only gets too hot to work with when ambient temps are high (40C+) or I do more than about 40 welds in rapid succession which is pretty rare.

Cheers

Where’s the best spot to weld on the cell? Center or the edge?

On the negative end, avoid welding to the very center.

On the positive end, the entire surface is very isolated from the interior and you can weld to any part.

how bad is it to have done this already?

If the negative tabs on the inside have not come loose, you'll be fine. You can individually cycle test them, and put them under load while looking at them with a cheap heat camera, to see if there's a hot spot.

With the kWeld currently unable to ship to the UK (confirmed by Frank this morning) what options do I have for building fairly high discharge 4S2P packs from INR21700-30T cells? I think I could probably get away with 0.2mm pure nickel but would be more comfortable with 0.3mm.

What's out there that can handle 0.2-0.3mm pure nickel, including stacked for the parallel/series connections? Is my only option to play the QC lottery with Sunkko units?