Resistance Soldering Bus Strips to 18650 Cells

Joined
Jul 21, 2013
Messages
2,010
[moderator edit] this thread has been split-off from another thread (seen in titles below)

This thread is about the general subject of resistance soldering bus-strips onto cells, regardless of whether the busses are copper, nickel, or nickel-plated copper. Resistance soldering refers to passing electricity through two parts to get them hot enough to melt solder. There is a wide variety of formats to use for RS, but 3V and 40A is our initial target that has proven to be workable. Voltages between 2V and 12V are all feasible, but the heat of 40A is the most important factor. A key reason RS is of interest is that...the thicker you make the copper bus, The ability to make a connection with RS does not get worse, it remains the same (using copper is problematic for spot-welding, and the thicker the copper, the worse it is).

"DIY MOT-based [home] Resistance Soldering Unit (RSU)"
https://endless-sphere.com/forums/viewtopic.php?f=2&t=89289

"DIY portable soldering iron or RSU from 18650's"
https://endless-sphere.com/forums/viewtopic.php?f=2&t=89379

1891 patent for "Electrical soldering"
http://www.google.com.gi/patents/US449258

The orginal post by ridethelightning is seen here below:

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

could you please post some pictures of the welder and packs you have built with this technique?
 
It's not a bad idea, not sure if it's really beneficial- I just haven't looked into it. What makes you think it can take 10x more current, I'd be a bit worried about dodgy joints and fatigue related issues. Or a poor joint heating, failing and a bit of nickel coming loose. I tried it once, got a few weak joins and gave it up.
 
craigsj said:
kdog said:
It's not a bad idea, not sure if it's really beneficial- I just haven't looked into it. What makes you think it can take 10x more current, I'd be a bit worried about dodgy joints and fatigue related issues. Or a poor joint heating, failing and a bit of nickel coming loose. I tried it once, got a few weak joins and gave it up.
It's trivially easy to get a 10g equivalent copper connections between cells, that's why. Copper is 5 time more conductive than nickel and it's no effort to get much more material in the connection. You don't have to limit yourself to paper-thin, resistive materials to get the spot weld to work. To make spot welds work you have to choose materials not well suited to the longer term task, quite the irony.

As for dodgy joints and fatigue related issues, it does take a bit of practice, but why aren't you worried about that with spot welding? How are the solutions different? Is 7mm 0.5 copper more fatigue sensitive than 7mm 0.15 nickel? I've had spot welds fail, too, and cheap spot welders can be unreliable and produce bad joints. I've also blown a few holes in cells with spot welders. Good tools and technique will result in reliable packs either way.

As far as "a bit of nickel coming loose" that seems more a problem with nickel strips than with copper that is 3 times thicker. I would say that is, without a doubt, the last thing to worry about.

Here is the unit I use: https://americanbeautytools.com/Resistance-Tweezer-Systems/98/features

It's the lowest power unit and is completely adequate for pack assembly. For lighter copper, say 28g, half power works. For 24g 7mm it takes full power. Yeah it's kind of expensive but produces a good result and you don't have to devise clever schemes to use copper far thinner this way.

I do not have pics of any packs that aren't currently wrapped. I have a project I will start soon and will take a few pics then. Using resistance solder is as fast as, or a little faster than, spot welding.

could you put some pictures or video how you do resistance soldering? So far what I have googled to me this seems just a bit more precise than normal soldering in terms of heat delivery but still much more Joules are to be pumped into the cell comparing to spot welding.

(BTW I think pulse arc welding is the way to go for copper but it's no small undertaking for DIY community. I have no problem spot welding .2mm today but concerned with material strength and amount of energy needed for the weld. Pulse arc welding would enable welding thicker and stronger copper)
 
wow. please post a video showing how it's done and the details about how much solder you use and what solder or flux you use and all the other practical details if you can.

whats wrong with soldering copper strip the old fashioned way?
 
Sounds interesting but can't find a single anything on resistance soldering 18650 cells. Since I'm from the Show-Me state can you help a brother out?

Tom
 
craigsj said:
vex_zg said:
...but still much more Joules are to be pumped into the cell comparing to spot welding.
This is not true though.

Spot welding 0.2mm nickel takes considerable power. I've seen numbers around 200 Ws posted here and used that much with mixed success. I don't have a particularly good welder. My 0.2mm are not especially reliable and don't work at all with the handheld electrodes because the unit isn't strong enough.

If you do 3 welds back to back with 0.2mm nickel, that's 600 Ws and it wouldn't be uncommon to do that. Add to that the layering of nickel that is often done for higher current.

In contrast, I can resistance solder ⅛" 28g copper strip in 6 seconds with 50 W or 300 Ws total. This amount of copper will carry more current than 0.2mm nickel.

I usually use 3/16" 24g (0.5mm) which takes more like 500 Ws. I occasionally use ¼" (7mm) 24g when I want 10g wire equivalent. If you need even more for packs where 10 gauge wiring isn't enough, the unit I use is probably too small.

Copper sinks heat much more effectively than nickel so less of the total energy is "pumped into the cell". It is also easier to work than nickel and lays flat on cells more consistently.

The criticism of soldering is that it destroys cells because of heat. Spot welding heats cells too, in fact possibly more so that resistance soldering does. You have to be careful comparing the two from a cell heating perspective, a fair comparison requires that you get a similar current carrying result. We're conditioned to accept the meager capability of thin nickel as the standard, if we were to set copper as the standard our view would be different. Spot welding has production floor advantages, people should stop thinking its an ideal electrical solution. I'm not even convinced it's the best mechanical solution.

vex_zg said:
...this seems just a bit more precise than normal soldering in terms of heat delivery...
It is TREMENDOUSLY "more precise". They are not remotely comparable.

I built a CD welder with weld energy measurement, I weld .2mm copper in range of 180-240J per weld, .2mm nickel takes around 80J, maybe less, per weld. I can run tests/video if more precise values are requested.

6 seconds with 50W sounds reasonable, but in my experience (using regular soldering, not resistance soldering) 50W can't a good solder joint in 6S, so please put your video where your mouth is :)

Lot of us has been after copper but so far not ideal solution has been found. Your turn.
 
craigsj said:
vex_zg said:
I built a CD welder with weld energy measurement, I weld .2mm copper in range of 180-240J per weld, .2mm nickel takes around 80J, maybe less, per weld. I can run tests/video if more precise values are requested.

6 seconds with 50W sounds reasonable, but in my experience (using regular soldering, not resistance soldering) 50W can't a good solder joint in 6S, so please put your video where your mouth is :)

Lot of us has been after copper but so far not ideal solution has been found. Your turn.
My turn? Well, I commented here for a reason... ;)

Resistance soldering copper is far easier than welding it. The question is whether it can be done reliably and without damage to cells. I believe it can be. If you have a sound electrical and mechanical connection, you shouldn't care if it is welded or joined in another manner. Furthermore, welding alters the cell so doesn't lend itself to "do overs". In that respect, anything would be preferable if it worked as well.

I can't comment on your welder's power usage or your results nor can I say what my welder actually delivers to the joint, but your number is lower than I've seen. It's still only one weld and 0.2mm material. Obviously a huge improvement over nickel, but with resistance soldering you can go much thicker. 0.2mm copper would require very little power to resistance solder. I don't know the number since I don't have anything that thin.

Resistance soldering delivers power to the joint instantly, it cannot be compared to a soldering iron in the way it works. Solder melts inside the joint about a second after application of power, the rest of the time is needed to allow the electrode to heat sufficiently to avoid a cold joint. The power level is chosen based on how much copper is present since it pulls heat away so effectively. You can't compare the process to an iron which relies on thermal conduction only and needs a good thermal path to work at all. With resistance soldering much of the heat is generated inside the joint in the same way as a spot welder does it. A side benefit is you apply pressure with the tool so when you remove power you can hold the joint in place while it cools.

I'm building a pack today so I'll have a couple pics. Not accustomed to posting video online. I'll look into it.

Uploading on youtube is really easy, basically just dragndrop. Now would be a good time to start getting accustomed :) Take your time but hurry up :)
 
Awesome technique. I wonder about the advantages compared to standard iron soldering. Less heat than with a soldering iron ?

A rewired microwave oven transformer (MOT) is typically in the 600-1500 W range.
If we need to tune down the thing to 100 W.s for 5-6 seconds (meaning 16.7 to 20 W power), I wonder how we could do it...

I see 2 ways since P = V x I.
So either trying to reduce V (AC, not DC since a transformer only works on AC eletricity) in order to reduce wattage, or reduce current (more complicated right ?).

Any ideas ? 500$ is some money for a welder...
 
yea whats the benefit in paying 500$ to solder? less heat? faster? easier? And why don't I ever see people soldering the traditional way with copper anyway?
 
craigsj said:
Hummina Shadeeba said:
yea whats the benefit in paying 500$ to solder? less heat? faster? easier? And why don't I ever see people soldering the traditional way with copper anyway?
Far greater control over the heat and less heat as a result.

Assuming you mean soldering cells the traditional way, it's because it's hard to control the heat and soaking a cell with heat damages it.

That's the benefit of paying $500, you get a tool that works. It's only costly because not many are sold. It's a very simple device.

I'm sure we could make a device like this for less (maybe a fifth of the price or even less)... I've got 2 MOT lieing around unused, both are 1500W.
If only I could look what's inside of this resistance welder... :D
 
Here's a DIY MOT resistance soldering project: https://softsolder.com/?s=Resistance+soldering&submit=Search
It appears to use completely different electrodes, so the currents we need may be different, but you can still probably learn a lot from it.
Edit: fixed link
 
Interesting. You are using soldering paste instead of wire, good idea.
It really solders much faster than a "normal" soldering does.

Can you try forcefully ripping of the copper strip away from the solder joint? to test that it is a good joint. Sometimes when I soldered nickel strips onto cells everything would look perfect but it would sometimes be a weak joint.

This process could probably be improved by using MOT as power source to be able to deliver more power in shorter amount of time.
 
craigsj said:
vex_zg said:
This process could probably be improved by using MOT as power source to be able to deliver more power in shorter amount of time.
I've seen some DIY online where the process is done in 100s of milliseconds rather than 1000s like I'm doing. Don't know if it's better or worse, no reason not to consider it. Higher temp gradients will be better for cell health so perhaps faster would be better here. I'd think several hundred watts for a second might work really nicely. The more wattage used, the more variable the heat generated if you don't carefully control the pulse length.

I think my unit targets fine soldering tasks like jewelry. There's no reason to believe it's particularly optimal for this application.

once the power source becomes strong enough, you are stepping into spot welding territory. There is another thread here about a guy who was actually using a spot welder to do "spot soldering".

I think this is as good and safe that the soldering can get, with more power could be made even better. I will consider this in the future projects.
 
I have been reading about silver solder, and a high silver content is used for various reasons depending on the application.

There is a sub-set of metal bonding that uses various methods and materials for jewelry. In this application, a high silver content is only used in order for the end-result to have the "appearance" of silver, because you are joining two silver components in making a piece of jewelry.

That being said, I found this solder paste. It is a homogeneous product (fully and evenly mixed, like a smoothie instead of a stew), and it contains flux and the bonding solder.

"This is a $34 one-ounce syringe containing silver solder paste mixed with flux. Free Flowing Silver brazing alloy containing 56% silver, 22% copper, 17% zinc, 5% Tin with a strong flux. This is our most popular paste. Can be used with a torch or in a furnace. This Melts at 1205 F"

I don't know if its acceptable for resistance soldering, but...for many, the biggest complaint about soldering is that it creates a layer of resistance (poor conductivity). I think the 78% silver/copper mix should dramatically improve conductivity of the joint, over conventional 60/40 solder made from tin/lead.

The perfect spot-welding bus material (IMHO) is C40410, which nobody makes. It has a couple percent of iron mixed-in with 98% copper. The copper conducts very well, and the tiny amount of iron provides the resistance needed to allow spot-welding to work. Nobody stocks this, much less ever being mass-produced to get a good price.

Spot-welding works on nickel because nickel is such a poor conductor that it has high resistance, so there is significant local heating at the "spot" where current is passed through (without heating the entire part). It's difficult to spot-weld copper strips because copper is a great conductor and it needs huge amps to get hot enough to melt (1980F / 1084C), so spot-welding is tricky onto an 18650 cell. In this paste, the 22% mix of zinc/tin would provide the resistance that would allow resistance soldering to provide local heating at a precise spot.

https://www.amazon.com/Solder-Paste-Flux-Silver-Easy/dp/B000VQ9HX4

There may be a better paste for resistance soldering, but this might be a good start for the discussion?

edit: here is a useful short video on how solder paste is used in industry to screen-print solder onto the locations of a circuit board where the components are "surface mount". Fast forward to 2:50

https://www.sparkfun.com/products/12878

microscope pics of solder paste

http://www.evilmadscientist.com/2013/solder-paste/

From Indium corp (industrial circuit board makers)
Typical tin-bismuth alloys melt around 140°C, tin-lead alloys around 183°C, and tin-silver-copper (SAC) alloys around 217°C

SAC305 has a conductivity of 13/100 IACS, one point better than the old 63/37 tin/lead, and several points better than the steel in the cell cans...
https://www.avoutlet.com/images/product/additional/f/iacs_of_metals_n_solder.pdf
 
I would think you ought to be aiming to use the lowest temp solder you can get. (pure lead ?)
I prefer solder to welding, but are there not some cells that simply refuse to take a solder joint ?
Also, the conductivity of the solder is of little relevance if the actual solder film thickness is kept to a minimum ( < 0.1 mm)
 
For resistance-soldering we want to find a paste with a good balance between a low melting temp, but as high of electrical conductivity as possible. In order for resistance soldering to work, there needs to be something mixed-in that is resistant to current, so it gets hot when you pass current through it.

(BiSn solder melts at very low temps, but is brittle, and in frequent hot/cold cycles it develops cracks. By adding 1% silver it becomes dramatically more malleable)

C___F___Melting temps

140 284 Bismuth/Tin/Silver 57/42/1% (Indalloy #282)
185 365 Tin/Lead 63/37%
217 422 SAC305 lead-free solder (96% tin)
232 450 Tin
327 621 Lead
419 787 Zinc
921 1690 Brass, Yellow
961 1760 Silver
1000 1832 Brass, Red
1080 1980 Copper
1370 2500 Steel, common
1400 2550 Stainless steel, 316
1453 2647 Nickel

Old style lead solder has a melting point near 365F (185C), lead-free is roughly 428F (220C)

Pure tin is 450F (232C), and pure lead is 621F (327C). It is odd that tin/lead solder melts at a MUCH lower temp than each of the components...

As a newbie solderer I found the temp-controlled iron was a big help with lead-free solder.
I had a big reel of lead solder handed down from my dad. No idea how old it was but its lasted me a long time and gave great results even with a cheap iron. The difference I found is that the lead solder is more forgiving. The lead-free solder needs extra flux and everything immaculately clean. I find rubbing the solder pads with the pcb rubbers/cleaners, even if they look clean, essential.

I hate lead-free solder, but I have to use it because of work. I find that for through-hole tin/lead I set between 700F to 750F. Tin/lead SMT I set 600F to 650F (a little high)

lead free:
750 SMT, 800 TH

now before everyone flips that my temps are high, remember that I solder fast. These are momentary touches to component leads

The most commonly used lead-free alloy, Sn96.5 Ag3.0 Cu0.5 (Tin, Silver, Copper), commonly referred to as SAC305, has a melting point of 422°F (218°C)

There is also solder-paste that is copper-based, brass-based, and as stated above, silver-based (made for jewelry manufacture).

https://www.interweave.com/article/...icro-torch-soldering-copper-brass-and-nickel/
 
The Nickel-strips need to get up to 1450F at the point of contact for the spot-welding to melt it enough to make a good bond.

If I can make a resistance-soldering bond in two seconds at 428F, the "pulse" is longer, but the temps are much lower. Maybe this will end up worse than the existing methods, but I feel compelled to investigate further to see where this might go...

The evidence shows that copper-strips can be resistance-soldered onto 18650's quite easily. A bad cell can be un-soldered, and the device should be much less expensive than the cheapest spot-welders, plus extremely reliable because it only passes current through two probes. As an example, my 100W soldering iron passes current through steel tip, and it is about 15 years old, with no signs of dying.

edit: the stock terminal caps/ends on an 18650 cell are likely to be stainless steel. No matter what else we do, there will always be an SS link in the current path. Having a copper bus can also help pull heat away from the SS 18650 tips.
 
..There is also solder-paste that is copper-based, brass-based, and as stated above, silver-based (made for jewerly manufacture).
sure, but those materials are for much higher temp processes, generally requiring an open flame torch for sufficient temperature.
( Silver soldering, Brazing, etc) ..There is also a similar Aluminium based process.
maybe a resistance type process could be developed, but the temperatures will still be much higher than lead/tin soldering.
 
There was a process used by the Inca's to make platinum jewelry. This was a shock to modern researchers, because platinum melts a higher temperature than steel. They looked for evidence of small blast-furnaces and found none. A microscopic examination revealed that the platinum had never been melted.

It "appears as though" they ground platinum into a dust, and then mixed it with something that melts at a lower temp. I think the article mentioned silver, which was very common for pre-Columbian jewelry, and well-understood by the craftsmen. Then, an acid was used to dissolve away the top layer of the silver, leaving only protruding granules of platinum. The jewelry was then polished, resulting in the appearance of a solid platinum part that had been cast.

If a copper solder paste is 20% solder, we only need to melt the solder. The microscopic copper dust granules can remain un-melted. This may be a waste of time, but I am interested enough to perform a test.

Although this could be used for a common 7mm wide ribbon/strip, I think using a fuse wire would require putting even less heat into the cell-tip, and could be very easily and affordably do-able for the DIY garage enthusiast.

The price of silver is about 80 times more expensive than copper, but it is only about 8% better electrical conductivity, so...copper paste it is.

edit: just as a start to the research, I immediately found 1.6-ounce (14.7 gram) syringe of copper solder paste for $12, and it states that it melts and flows evenly at 430F (221C), so the heat is equal to "lead free" solder. https://www.amazon.com/Copper-Bearing-Solder-14-7-Grams/dp/B0058ED0QE
 
With a good milliohmeter it should be possible to test all kinds of solder joints for resistance. If the solder layer is kept to near-zero thickness by keeping the copper is pressed firmly against the cell end, I think the resistance of the connection will be way less than other parts of the circuit. The math can be done on all of this.

When spot welding nickel strips, the heated area is very small. When soldering, a much larger area needs to be heated. The total thermal impulse energy is what will determine the peak temperature.

I'm not an expert on cell chemistry, but as far as I know, heat damage would be primarily from either melting an insulator or from boiling the electrolyte. You don't want to do either of those. A partial melt on an insulator might be OK if it solidifies in its original shape/thickness.

Good 'ol 60/40 tin/lead solder has about the lowest melting point of anything you'd want to use. By quickly ramping up the temperature and cooling afterward, heat damage can be minimized (if there is any). Resistance soldering might be possible with copper strips and that MOT I have under my bench. Instead of trying to add resistance to the strip or solder, using enough current should be able to heat the strip. I'm trying to picture why this would be better than my big soldering iron. I guess you could possibly heat and cool faster and the electrodes would hold the strip down while cooling.
 
I may have it wrong, but...

I am convinced that resistance soldering (passing current throught two probes to make a certain spot hot) will work much better than using the physical heat of a 100W "fat tip" soldering iron. While the iron is heating the two parts that need to be bonded, that heat will spread, plus the consistency can be variable depending on the contact point of the soldering iron, along with a proper waiting period between solders to ensure the iron remains at a consistently high temp in order to have consistent joints.

I just have a gut feeling that this can work better, if a given builder was set on soldering a joint, instead of using a pressure contact or spot-welding.

edit: unless some new development presents itself, I plan to test resistance-soldering a fuse wire to the positive tips of 18650's, and as far as the bottom negatives, I am warming up to using button magnets over copper ribbon as the conductor (The magnets do not conduct current)
 
great detailed info guys. thanks fr doing the research.
this is a very interesting field
I have been interested in the idea of welding or soldering a fuse thickness wire to each terminal, if the wire is the correct material(say, tinned copper), and the correct thickness, minimal heat would need to be applied to get the connection required.
 
I agree!

There is a process for mass production of circuit board with surface-mount components (capacitors, resistors, etc), called "wave soldering" and it seems like they mask anything on the bottom of the circuit board that they don't want soldered and then run the board across a protrusion that has hot liquid solder flowing across it. The new "lead free" solder requirements means that there is a lot of industrial info on tin/copper based solder (SnCu) for wave soldering. Here's one example:

http://www.kester.com/Portals/0/Doc...per_Based_Solders-for_Electronic_Assembly.pdf

short video of wave soldering:

https://www.youtube.com/watch?v=VWH58QrprVc

short video on "dip" soldering

https://www.youtube.com/watch?v=TJ8IAw4jKvY

edit: the copper solder paste used by jewelers is only for appearances (they want the joint to be pretty), and the solder paste used by industry is made for ease of soldering and typically runs 5V on electronics at very low milliamps, so...we are probably on our own here. If we don't figure this out, it will remain unknown.

Powdered copper, zinc, and powdered solder exist and are very cheap and available (so hobbyists can mix their own blend of solder-paste). Solder paste has flux in it, and here is a short video on DIY flux.

https://www.youtube.com/watch?v=iv_2ULCmA28

Here's some copper-specific flux, from the jewelry manufacturing industry.

http://www.firemountaingems.com/itemdetails/h205453bs

SAC305 is a lead-free solder mix that is the most popular in industry for a variety of characteristics. Flows well at 217C / 422F, and is available as a powder. Mixing raw zinc/tin/lead powders will not have the same result as creating a solder metal, and then powderizing it, so...powdered solder should be our best bet.

I am now curious about a DIY solder-paste that is a mix of copper powder, SAC305 solder-powder, and a copper-friendly flux.

short training video with good graphics, about spot-welding

https://www.youtube.com/watch?v=kl8TAVdvE48
 
Back
Top