spot welding fuses

Hi all,

SO...
I have built myself a spotwelder out of a golf-cart-relay (500A) an old car battery and an arduino (which over a mosfet times the closing of the relay very precisely) . So far I have great success welding nickel-strips . This works now nearly every time (except for when the cell is really dirty ormessy or not cleaned the old weld-bumps --> notflat... :angel: :dodgy: ).

Now I am aiming for the next step:
Welding fuses :exclamation: . For my project (probbably gonna be a 14s32p for starters) I want all cells individually fused. But if not strictly necessary,I 'd prefer notto solder directlyon to the cells. I'd rather prefer welding since it is convenient, fast, and nearly free of fumes and nearly doesn't heat up the cell at all if well configured.

I did some first tests with resistor / diode legs/leads (as they are similar to the glass-fuses I ordered... still waiting for those).
It was a fatal fail. It is on one hand nearly impossible to get the electrode tip to rest on the cell and on the other hand was the current to weak for the resistor wire to be welded. My welder had only one setup (hard-coded to the arduino) so I will implement an adjustment potentiometer next... but one other idea got to my mind which I would like to test:

I thought on tapping the connector wires of the resistors (respectively fuses)with a hammer to flatten them out. This leads to making the wirethinner/sheet-likeand give a larger area to position the electrodes on. I didn't test it yet but as soon as I have results, I will post images of the process and results.

Cheers, elkooo

Sounds like a plan. I think someone else decided to tap the legs to make them a little flatter where the spot weld was going to be at. Not sure who that was tho, been awhile. Should work pretty good.

Ive flattened one of the ends of the electrodes on mine to make it a little easier.
Using a K-weld with an eBay single handed and spring loaded handpiece with interchangeable tips, powered by AGM motorbike size battery and trickle charger.
By making one electrode a bit bigger makes it easier to get a good pressure on the fuse wire.

I thought spending 1/2 hour fiddling on one item was easier than spending a little more time fiddling on each weld.

I have no idea if this is a good/bad/terrible suggestion...
What if you made a small groove or channel in the spot welder electrode so the wire is held in place?
I think it would need to be very shallow, otherwise the electrode would touch the material you are trying to weld to, instead of only the wire.

I also spot weld my fuses to the cells.

My DIY spot welder is an MOT based unit with adjustable double pulse and power controls.

I haven't thought of flattening the leads but that sounds a great idea, even though it's does weld very well.

Having a groove in the welding pins is actually used on commercial welders BUT it can be tricky with thin wire like a fuse wire.

This is done to have larger surface between the welder and the object you welding so it doesnt stick/burn off.

OK, nice inputs, thanks!

Looks great @chuckp, thats how I hope my results will look as well!

I think I will start by modifying the welder.
Currently I can only weld with 50ms (this is how long the spool in the relay gets powered by the mosfet...mechanical delay for closing the relay not taken into consideration... should measure this with the oscilloscope sometime...). I will add a potentiometer that allows me to change the time between 50ms (which works great for nickel strips so far) and ~150ms for starters to see if this makes any sense... will adapt it, in order to get good results.
Second, I might start with a groove and also test tapped fuses.

I do not plan to directly weld the actual fusing wire, but use glass fuses. This gives me probbably a bit more of margin because the connector leads are thicker than actual fuse wire.
Will post as soon as I have news (not that easy with family and work distracting me from doing powerwall research )




btw.
if someone is interested in my welders design and Arduino code... could publish this if desired...

SO....
I had a few minutes today to optimize my spotwelder.
It's got a potentiometer now to set the welding time: It can now close the relay from 50ms to 150ms. With this new setup I did some first tests.

First of all I did the test with the flattened leads:



I put the lead on a flat anvil and hit it with a hammer. In a fraction of seconds you end up having a nice flat lead end. Worked great and I think even if I had to make hundreds it would be ok. Far better than pre-tinning cells in my opinion.

Then I went along to welding tests:



The first one (on the right) was cranked up far to high! It really only needs a hair more than 50ms (I think most of the 50ms are necessary to mechanically close the relay which leads to only a fraction of that time for the real welding current to flow). So with 60ms or so, I probably already have doubled the actual welding time.
Went on with lower times and got amazing results (on the left!). One weld was on a flattened lead and even the unflattened lead welded on just fine (only needed a bit more time to cleanly place the welding electrodes before hitting it with amps ).

And for nickel it also works just fine:



The welds on the left were the first ones I did, on the right I also did a tearing-off test (the weld held, the nickel was torn apart...). The two weld spots on the left of the negative cell side were the last ones I did - in my opinion they look just perfect :heart:

So my battery will be welded on the cell side. I am still thinking about the bus-bar-side... but if I could also weld on the bus bar side, this would be just perfect... we'll see...

Copied from one of my other posts. Sorry could figure out how to do a link?

I didn't take any photos at the time of building it.





I had an old UPS for a PC that was no longer useful. I stripped it out just for the case as I thought it would make a good housing for the electronics.

I used a JST41-1200 Battery spot welding control board, 16 single chip microcomputer control, 1602 MCU LCD, encoder double pulse pcb to control the pulse duration. https://www.aliexpress.com/item/32810563725.html

Striped out an old 230V 850W microwave oven i had kicking around to salvage the transformer, cut out the secondary windings and removed the magnetic shims i replaced the windings with two turns of 16mm2 flexible battery cable with approx 1M of spare cable to exit through the case for each of the spark electrodes.

I mounted the MOT in the case along with the pcb and used a small 230V to 12V transformer for the pcb power supply. Cut the end of the case out and fitted an old PC fan on the end to keep everything cool, not that anything gets hot, I probably didn't need the fan. Pushed the battery cable out through a couple of holes drilled into the front of the case and cut holes for the display board and adjustment knob.

I soldered a couple of soldering iron tips into the ends of the battery cable for the electrodes and covered the ends with a few layers of heat shrink.

Used an old sewing machine peddle switch to operate the welder.

It works very well for welding fuses to cells with pulse power set to about 80% but I do need to have the pulse power set to 100% if I'm welding nickel strip otherwise the weld isn't strong enough.

I have now obtained a 1000W MOT that I'm going to fit in at some stage. I will take a few build photos when I get around to doing it.

I haven't measured the output amperage but I would imagine it's around the 450A mark, the voltage at the electrodes around 1.8V

I found that a two pulse weld works well on the fuses, P1 set to 65% for 70ms (to pre-warm the metal) a 25ms delay and P2 set to 80% for 90ms (for the weld).

The only purchases i made was for the JST41-1200, 230V-12V transformer and 2.5m of 16mm2 battery cable, the build cost was around 35 and a few hours of my time.

---

elkooo said:

So my battery will be welded on the cell side. I am still thinking about the bus-bar-side... but if I could also weld on the bus bar side, this would be just perfect... we'll see...

Click to expand...

I solder to busbars and spot weld to the cells.

Spot welding to the twisted copper strands was proving difficult, perhaps if I'd used a flat copper busbar it would be easier. Maybe an option on my next batch of packs.

Just in case you wonder what my setup is, here is the breadboard design of my arduino welder:




I had issues with thesolenoid oscillating. Not sure if the voltage drops too low for the solenoid spoolas soon as the welding current flows for a certain amount of time but I overcame the issue by only powering the secondary side of the relay by the car battery and having a separate power source for arduino & solenoid spool (primary).

In my case, I used a small DC/DC converter, operating at 7v output to power the arduino. Not sure if it could run from a 12v car batterydirectly (specs say 7 to 12v and the fully charged car battery is more like 13.6v or so..)

The electrodes are made from solid copper (I think it was really thick wire, salvaged from an old electrical installation). I cut threads on it and used thick copper ground wire for the connecting leads. They are mounted at an angle on a wooden handle where a small trigger switch is also attached to fire the welder.









The electronics and solenoid are mounted really messy on a piece of wood :angel:
But it works

If it is going to be parked and stored indoors, I would add fuses to each cell. It doesn't protect against all possible incidents, but...it mitigates the possibility of the entire pack catching fire due to one cell developing an internal short. If that happens, the one bad cell will still boil over, but...at least the rest of the pack has a chance to avoid that. It's not about saving the rest of the cells in the pack, it's about saving the garage and house.