18650 spot welding -how to- ULTIMATE REPOSITORY

well I reckon you've made a mistake...
186/3 =62. 62 amps through 6x0.15mm nickel would glow red hot for a few seconds then pop, arc and set fire to your pack/bike/groin
I put 80odd amps through a test strip of 10x0.2mm nickel to see how it would fair. it was bright red hot (but didn't melt).
there is no way this set up would survive... there are a few charcoaled battery photos around where arrangements like this have failed.
are you sure its the series connections?
 


This is how it looks like.
 
ridethelightning said:
quote [I started with something like that as goal but ended up building a 600ws 3F/20V CD welder myself, arduino controlled and configurable. I wanted something to weld .2mm copper and .3mm nickel, powerful reliable and customizable. Could not exactly find it so built it myself. Spent hundreds of hours, bought oscilloscope and learned a lot about transients, mosfets, arduino, and welding. Really happy with the result but I seriously overdid it :) quote]



I would be really interested in any details you might be able to share regarding building a welder like that :)

Not finished yet, I estimate to be at 75% completion - todo: charging & discharging control, temperature monitoring, error detection. Box. Electrodes for higher power welds.
The most difficult part was figuring out how to drive 25 mosfets in parallel, avoid ringing, deal with transients, minimal inductance layout. I blew up probably 50 mosfets over time. Lost count of how many days I spent prototyping and testing bad ideas until I figured it out. When I am finished I might offer this as a kit.
 
Started my first pack. Comments and critiques welcome. Batteries are samsung 30Q. I'm building 4 modules of 5s6p, for a total of 20s6p. Each cell is capable 20A for a total pack capability of 120A. Using 0.15x7mm strips, I would need 6 strips per cell, per series connection which gets me 21A of capacity. That's a lot of strips and a lot of nickel-to-nickel welding. My shitty welder couldn't handle that. It turned out a blessing in disguise and I ended up soldering 12 awg wire per series connection instead which is more than enough.

The balancing strip between the parallel cells is a single 0.15x7mm strip, which is capable of roughly 4.3A, which should be more than capable of balancing.

I believe in my informal testing, soldering is ok. I tried different insulators and spot welding against those insulators induced more burns/melt on the insulators than quick soldering. I know that's not necessarily conclusive - welding could have put less total watts of energy than soldering. I scratched each contact point, laid down blobs first. Each blob took 3 seconds max of heat contact. Then I heavily tinned the ends of each wire and finally soldered the ends to the blobs - again, no more than 3 seconds of heat contact needed - just enough to melt the tinned wire and the top of the solder blob to fuse.

Cell holder is custom designed/printed on 3d printer.

My lead wires are all equal length so no path is favored.

Pics:
2re2phy.jpg

aak4ft.jpg
 
riba2233 said:
What do you plan on welding with that beast? Too powerful for batteries, too weak for metalworking...


suprugu za kuhinju :)

my wife for the kitchen board :)

I wanted the ability to do copper welding at lower voltages to reduce blowouts. I can do .3mm copper now, probably even thicker, but still experimenting with weld timing, electrodes, pressure, voltages ... to reduce energy needed for weld and reduce chance of cell damage. A bit short on free time now - summer in Sweden just started and you better be out in the sun because you might not see it again till next year.
 
aethyr said:
Started my first pack. Comments and critiques welcome. Batteries are samsung 30Q. I'm building 4 modules of 5s6p, for a total of 20s6p. Each cell is capable 20A for a total pack capability of 120A. Using 0.15x7mm strips, I would need 6 strips per cell, per series connection which gets me 21A of capacity. That's a lot of strips and a lot of nickel-to-nickel welding. My shitty welder couldn't handle that. It turned out a blessing in disguise and I ended up soldering 12 awg wire per series connection instead which is more than enough.

The balancing strip between the parallel cells is a single 0.15x7mm strip, which is capable of roughly 4.3A, which should be more than capable of balancing.

I believe in my informal testing, soldering is ok. I tried different insulators and spot welding against those insulators induced more burns/melt on the insulators than quick soldering. I know that's not necessarily conclusive - welding could have put less total watts of energy than soldering. I scratched each contact point, laid down blobs first. Each blob took 3 seconds max of heat contact. Then I heavily tinned the ends of each wire and finally soldered the ends to the blobs - again, no more than 3 seconds of heat contact needed - just enough to melt the tinned wire and the top of the solder blob to fuse.

Cell holder is custom designed/printed on 3d printer.

My lead wires are all equal length so no path is favored.

Pics:

I personally would never be doing this kind of soldering directly to cells, even more because it is a thick AWG12 wire that takes quite some energy/heat to weld properly. I think this increases chance of cell damage/fire/issues later.

If you insist on soldering then you can presolder wires to tabs before welding them to cells - and solder wires in between cells, not on the actual cell top where a welding joint is to be.
 
vex_zg said:
I personally would never be doing this kind of soldering directly to cells, even more because it is a thick AWG12 wire that takes quite some energy/heat to weld properly. I think this increases chance of cell damage/fire/issues later.

If you insist on soldering then you can presolder wires to tabs before welding them to cells - and solder wires in between cells, not on the actual cell top where a welding joint is to be.

Yeah, I know I'm taking a bit of a gamble. But as I said, I did do some experiments to see how much heat is generated beneath the strips. There is a lot more instantaneous heat under a spot weld, enough to char wood, whereas soldering of 12 awg wire did not. I think a spot weld penetrates deeper than people expect. I also think that soldering on to the strip helps dissipate some of the heat sideways. Either way, I'll find out soon enough whether direct soldering is viable.
 
I also think that soldering in between the cells should do a better job. But then I question myself if you will still need like 6 nickel strips on top of each other, because one single one won´t do the job, I think...
 
I did another experiment.

I placed a small strip of 0.15x7mm nickel over a large piece of steel and proceeded to heat a blob of solder on top of the strip for 5 seconds, while pressing down on the strip with the soldering iron. I then immediately removed the strip and touched the steel underneath. The steel piece wasn't even warm! The nickel strip was quite hot, however, and I could only flick it with my fingertip.

I repeated with a smaller piece of steel, a 1 inch washer, spending about 10 seconds of heating time. This time the washer was warm to the touch, about 45-50C.

The larger the heat conductive material below the solder, the greater its thermal capacity. So the larger steel piece was able to absorb and distribute the heat throughout, negligibly increasing its total temperature. The metal washer was much smaller, with smaller thermal capacity, so its total temperature was much higher.

My theory:
While not exactly the case, good conductors of electricity tend to be good conductors of heat. I believe therefore that an 18650 battery itself is going to be a good heat conductor, with enough thermal capacity for soldering without damage. This means as I pump heat into the terminals from the soldering iron, it will efficiently distribute that heat throughout its entire volume. I think 5 seconds of solder heat will only increase the total temp of the cell by a small amount and easily under its max operating temperature. The washer whose total volume was far less, only heated to 50C after 10 seconds of solder heat, which is far more than the 3-4 seconds needed to solder.

This may be controversial, and still just my theory, but I wonder if spot welding, with its instantaneous 2000+ degree heat, might be worse because the battery isn't able to dissipate that much concentrated heat quickly throughout its volume.
 
The can end is a good conductor of heat-you're right. But it has very little heat absorbing capacity because it's very thin. So the heat goes through to the internals very quickly.
 
aethyr said:
I did another experiment.

My theory:
Theories are subjugated by laws, you experiments don't change the laws of physics, or what happens to a heated cell. Manufacturers experiments have been done under laboratory conditions by engineers and the standards have been set. There's no value to experiments with no standards that fly in the face of everything the manufacturers have learned and discovered using the LAWS of physics. Soldering cells is a bad idea. Period.

If there was a dime for every bloke that came here knowing better I could afford a whole new bike.
 
Does any of these 78X or similiar welders do 0.15mm nickel to 0.15mm nickel weld? Mine dont (220V version).
0.15mm nickel to battery is ok. So what silverish material is 18650 terminals made of? Attracts magnet so its not stainless steel. Not copper and nickel either.
 
tomjasz said:
Has anyone built this welder with an Arduino timer? Also a display option?
I have. Bought the whole kit, initially without display and later upgraded firmware and added the display.
For a long time I struggled getting it to work properly - the controller seemed to reset during welds which I believe was a "brownout". Fixed it by adding large electrolyte cap before voltage regulator. Just confirmed, I can weld 12x0,15mm nickel to nickel without problems. Actually this is with 150ms weld and 50ms pre-weld times; anything more blew holes through. Which means it could do more than that; I'll have to do some tests.
Used 50mm2 welding cable - stripped it from rubber insulation and used heat shrink tube instead (tip: don't heat it to shrink - it's much easier to work with like that). Welding arms are CNC milled copper with cross-section area of 200 mm2 (20x10). So yeah, overkill is just about enough.
Only "hitch" right now is, text on my display is turned 180 degrees, upside down. Must be some kind of firmware flag/option.
 
I wonder if you safe really that much of money if you built your own spot welder.. :roll:
 
tomjasz said:
aethyr said:
I did another experiment.

My theory:
Theories are subjugated by laws, you experiments don't change the laws of physics, or what happens to a heated cell. Manufacturers experiments have been done under laboratory conditions by engineers and the standards have been set. There's no value to experiments with no standards that fly in the face of everything the manufacturers have learned and discovered using the LAWS of physics. Soldering cells is a bad idea. Period.

If there was a dime for every bloke that came here knowing better I could afford a whole new bike.
Actually, laws of physics simply state the behavior and theory attempts to explain why something behaves as it does. For example, Newton's law of gravity states that anything with mass will attract to each other with a force dependent on their mass and distance between them. That law doesn't attempt to explain WHY particles attract to each other based on mass and distance, only that it does. Various theories exist to try to explain why. Theories and laws work together :D

Nothing I've done contradicts whatever the engineers have done. Specific heat and thermal capacity is physics/chemistry. As an engineer myself, I know what we publish to the public and what we engineer are often quite different, for many reasons, both legal and from an engineering standpoint. We often over engineer beyond the specs of the product and if some area hasn't been over engineered, we simply say don't use or do something to that area.

For example, it could very well be that soldering for under 5 seconds is perfectly fine, but 10 seconds becomes questionable, depending on temp and wattage of the iron. Its better from a legal and engineering standpoint to simply say: "don't solder", because you can't control HOW people solder and all the different soldering variables. And since battery safety mechanisms are based on heat and pressure response, its difficult to engineer the battery to be safe AND handle high, prolonged soldering temps. Therefore, a cell simply has no over engineered heat tolerance and so they have to disavow any heat application to the cell as part of its published specs.

Second, I'm not entirely convinced they've done a lot of experiments on soldering vs welding. If you look at the datasheets for sony or samsung, they publish various torture tests - overvolting, shorting, dropping, high temps, low temps in great detail. Do they publish that level of detail for soldering vs welding? No, they don't. They simply say don't solder.

I've found only one statement that only briefly explains why not to solder http://www.hurt.com.pl/prods/bat/_li_ion/ncr18650ga.pdf:
(11) Soldering
Do not directly solder to the battery.
Soldering directly to the battery could melt the separator or damage the gas release vent or other
safety mechanisms. This may cause the battery to generate heat, smoke, catch fire, or explode

Note, there is nothing about how soldering will damage to the cell chemistry and cause capacity loss, but rather that it could compromise the safety mechanisms that are heat sensitive. Now that's a valid issue, and I will try to test this. In fact, I plan to cut open a cell, apply soldering heat to the cap and measure the temperature and observe any damage to the gaskets and venting mechanisms.

But I thoroughly do not believe that the battery chemistry can be harmed from proper, quick soldering because of the very laws of physics which dictate how heat distributes and is stored in a material.

In the end, I'm simply showing my findings. If you don't ever want to solder, don't. As for me, I have high discharge cells that need more than a single strip of nickel to transmit that current, so I needed to solve this in some way. And I fully admit I could be wrong and my pack will blow up. :pancake:
 
All good with me. I'm listening and watching. Most importantly, to me at least, im always willing to be wrong.
 
Over the years I've seen quite a few people build soldered packs. I don't recall any reports of failures. I've done it myself on a pack and it's still going. I wouldn't recommend it based on manufacturer datasheets but empirically it seems to work ok.
 
soldering between the cells +1 when you put the cells in holders and spotweld the parallel tabs you have more space between the cells to solder the serie wires :idea: okey take,s some extra space on the total pack bud i think this wil wurk save and good not to much heating the cells.
 
We all know that 787 welder does not weld 0.15mm to 0.15mm nickel.
But what about 0.1mm nickel to 0.1mm nickel?
Recent Weld/build:
http://i.imgur.com/xiZBpTx.jpg
 
DVDRW said:
We all know that 787 welder does not weld 0.15mm to 0.15mm nickel.
But what about 0.1mm nickel to 0.1mm nickel?
Recent Weld/build:
http://i.imgur.com/xiZBpTx.jpg

I've switched to 0.1mm strips and my 788h with homemade pen setup welds fine, nickel-to-nickel.

Why are your + terminal connections to your copper bus so thin?
 
To protect house, cells and save money.
Thanks for the info.

Actually you can just easily weld test if nickel is pure or not.
Ni plated steel: Sparks flying everywhere when welding
Pure nickel: No sparks, little puff of smoke.

So no time consuming salt water test required.
 
.1mm cu bus welded to .15mm ni.
Easy to do this one bc I had ease of placing an electrode each side. Welds are the strength of the cu which is reasonably weak but adequate.
 

Attachments

  • image.jpeg
    image.jpeg
    53.3 KB · Views: 5,970
Back
Top