Resistance Soldering Bus Strips to 18650 Cells

[spinningmagnets]

That's one of the reasons I am looking at resistance soldering. If you are using copper bus-strips (especially if they are thicker than the common 0.20mm nickel) then...heat soldering would require huge BTUs to get the joint hot (plus the physical act of heating with an iron is quite variable and inconsistent), spot-welding requires tons of energy to get the joint hot (using an expensive and un-reliable machine)...but

If you pass current through a thick copper 7mm X 0.30mm ribbon onto a stainless steel cap of an 18650 cell positive nipple, only the solder and cathode will get hot, and as soon as the joint is formed (two seconds?), the copper will heat-sink the BTUs away from the cell.

7mm X 0.30mm ribbon is equal to 2.1mm squared, and converting to round stranded copper wire, that is 14-ga AWG, 20A continuous (30A peak?) per cell?...(of course we can use thicker copper ribbon, too). Nickel has about 1/4th the conductivity of copper, so...

0.20mm nickel, no...0.30mm thick copper, yes!

Here is a youtube demonstration about resistance-soldering a thick copper part. Copper has such an ability to wick-away heat, that a huge iron would be needed to solder the parts with a conventional soldering iron. For mere connectivity (with no need for electrical conductivity), resistance soldering can be considered a "soft" form of brazing, by passing current through two parts, instead of using a flame-torch...

https://www.youtube.com/watch?v=cpAbL88_o5I

 

[Hillhater]

I think you should understand why these tin/copper solders are in use.
It is not because they have some mystery electrical or technical avantage, but purely to enable the use of a low temperature soldering process, using solder with NO lead content.
Healt and Safety regulations prohibit the use of lead based solders for commercial use , and have done for many years, so industries (particularly the electronics) had to source and develop alternative solders to replace the favoured Lead /tin one previously used. Tin/Copper ++ is the solder which now best meets those commercial needs.
The Traditional 63/37 Lead /Tin still has the lowest melting point of 183C and would still be the likely choice for non commercial cell assembly, though some higher tin content solders may provide better "wetting" to the cell, depending on the exact material of the cell components.

 

[spinningmagnets]

Yes, I can get powdered solder that has the old 63/37 mix that melts at a lower temp and flows well. I can mix-in copper powder to try that out. However, right now I want to try the SAC305. Its only about 32C / 60F more degrees of heat, and if I can get that to work, then anyone who wants to use the 63/37 blend (plus copper powder) should easily have an good result.

I'm sure I'll try both (with varying percentages of copper), just so see how it goes...

__C_____F____melt temps
185C / 365F___Tin/Lead, 60/40 solder
217C / 422F___SAC305, lead-free solder

1450, 2650___Nickel spot-welding

 

[sn0wchyld]

i agree... just think what heats up faster... the tip of your iron or the ceramic resistor inside it? this method discussed is effectively making the copper strip that resistor...

i suspect the internals of the commercial machine also closely resemble a MOT, or any other step down transformer. If you (speaking generally here, not specifically you spinningmag.) want to use a MOT, you can control its power output via the number of turns (which lowers the output voltage, thus lowering current and power) or by using a variac on the input (same principal, just safer/finer control, and much higher overall efficiency too). Variac's can be had at ~$500 too, so might be a more versatile option than the commercial unit.

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)

 

[kdog]

When I was putting together my first mot based spot welder, I wasn't getting much luck with the welds and tried spot soldering. It did work, but I didn't have the set up for it. I suspect a slightly lower amplitude and longer pulse would help. The excess solder was ejected at a good speed from the high current. Kind of like a mini bleve on your battery! I got a few cold solders and lost interest after my reworked spotwelder delivered the goods.
I think it has merit as a technique...love to see some high current testing done on some can/tab joins.
As cell tech gets better, nickel is goin to be evermore incapable of delivering the high current that batteries can provide. Specially in strip format.

 

[fechter]

With the right current, that copper strip will heat up very quickly. Remember the old Weller soldering guns? The tip was essentially a piece of 12ga copper wire and used a 1/2 turn secondary on the transformer.

Ideally, you'd want to somehow measure the temperature of the work and limit how hot the strip gets. You also have to hold it at that temperature long enough for the cell end to get up to soldering temperature and have good flow. Some kind of infrared thermometer feedback could possibly work. With a programmed, controlled profile, you could get very consistent results. The other alternative is the human eye and visually observing when the solder has flowed on the cell end.

For soldering a fuse wire, it will be hard to beat a regular soldering iron. It seems like this would require much less heating of the cell than using a beefy strip.

 

[spinningmagnets]

Soldering a 7mm X 0.30mm ribbon, and soldering a fuse-wire...are definitely two completely different ballgames. This may take a month or two, but I believe in this, and I want real hands-on data...

 

[spinningmagnets]

Thanks for your insightful comments. I was somewhat shocked to realize that the positive "button" on our beloved 18650 cells is either stainless steel, or nickel-plated common steel (along with the entire shell "anode"). I assumed it was nickel-plated brass (which is stiffer than copper, but less conductivity)...

If copper has a conductivity of 100/100 IACS, then...these materials are 10/100. If you were an electrical design engineer and you recommended steel wire, you would be fired!.

You make a good point that trying to improve the conductivity of the free-flowing portion of the circuit is somewhat "obsessive", but...I cannot change the materials in the popular 18650 cells, and we are forced to use what we have been given.

A characteristic that has not been mentioned very often is that...copper busses will draw away heat from the steel cans on our 18650's, which has the benefit of making the weak-link in the system run cooler...

Not that I deserve to be listened to, but...my current best bang-for-your-buck high-performance DIY pack construction is...resistance-soldering a fuse-wire onto the positive button of an 18650 / 21700 cell...which is then connected to thick copper series-bus flat-bars. The broad and flat negative anode has a 10mm X 0.30mm copper ribbon (or thicker) held onto it by a button magnet.

Here's a PDF on making your own DIY cheap resistance-soldering machine. This guy uses a 12V car battery charger and a dimmer switch to vary the 120VAC input.

http://www.nmia.com/~vrbass/models/solderer.pdf

edit: I think I have found the gold-mine of DIY resistance soldering. Here is a hobbyist who details his opinions, and then details how to build a DIY RS unit. Unless you want to pay $450 for a turn-key product, the best candidate seems to be a large 120 VAC charger for a 6V battery, and you use a router variable speed controller to vary the input voltage. The output voltage will stay at approx 6V, but the output amps will vary, and it is all very safe to do.

http://technitoys.com/diy-resistance-soldering-outfit/

If you are familiar with the MOT spot-welders, you acquire a large and heavy transformer. You keep the coil with the larger diameter wires, and then in the space evacuated by removing the fine wire-coil...you insert 1-1/2 wraps of super-fat wire. By inputting 120 VAC to the smaller-wire coil, the output becomes roughly 3V and very high amps. Since the spot-welder probes touch the work, higher voltage is not needed, and the "lower the volts, the higher the amps" in the conversion.

An RS unit needs a large transformer to access the common 120VAC wall-socket power. You "could" rewind a salvaged large "Microwave Oven Transformer" (MOT), and add a router speed controller to it, but also...large and heavy 6V chargers exist, for a few dollars more.

Here is an example of a 100A 6V charger for $81 that might be a viable candidate "Schumacher SE-1510MA" (100A might be overkill).

The ideal candidate will have an output of 3V-12V...any voltage above that will likely cause sparks, and for a given voltage...the higher the volts, the lower the amps (with the power coming from the wall socket at a fixed 1200W (120V X 10A, with 15A breaker), so...1200W / 6V = 200A as a theoretical maximum). It is the amps we want for RS, and if you go down as far as 3V for the design to access higher amps, it may be harder to find an acceptable cheap used transformer).

This concept is still under research, do not make any purchases based on info here, until it is verified.

additional edit:
craigsj mentioned that he is getting good results resistance-soldering very thick copper 24-ga bus-ribbon to 18650's with a specific unit that provides 100W / 2.6V = 38A

copper sheet thickness

https://www.riogrande.com/search/go?w=copper+sheet

34-ga is 0.15mm (0.006")
32-ga is 0.20mm (0.008")
30-ga is 0.25mm (0.010")
28-ga is 0.33mm (0.013")
24-ga is 0.51mm (0.020")

https://endless-sphere.com/forums/viewtopic.php?f=14&t=88965#p1297238

Do NOT remove the probes from the work-piece until after you have de-energized them. The high-amp arc would be like connecting the two posts of a car battery. Remove all rings, watches, and necklaces, and of course place everything on a non-conductive mat.

 

[cheapcookie]

very interested in this way of doing cell soldering

 

[cheapcookie]

A friend has this

should I try it ? it would be nice if it could pulse for two 2 seconds and then shut off

 

[bedesign]

A friend has this - 'img. of autotransformer'

should I try it ? ...

NO don't do it as it is, save your self and others

[pre]There are at least two main reasons -

1.
It's been warned about earlier in this thread, as these types
are AUTOTRANSFORMERS one of the output
leads and the whole battery pack worked upon will carry the
full peak voltage of the French/RU/EU 324 V RELATIVE TO
THE EARTH GROUND depending on how the wall plug is
turned at each plug in.
** TURNING DOWN TO 4V WON'T HELP **

Even if used in the US/UK/Canada?/Australia?
with fixed position of 'Neutral' and 'Hot Phase' in the wiring,
as implied by the plug and socket in the picture - there is
the risk of human mistakes in my experience in ALL
manufacturing and tests combined.
Add to this the odds of mistakes in the house internal wiring.

(just watch the recalls and error rates of typical 100/unit
in car manufacturing)
- too often the 'what if' design philosophy is ignored

2.
One has to trust oneself that the voltage adjustment knob has
not been turned up, or by someone else.


Though, putting an ISOLATION TRANSFORMER before or after
would make it safer (possibly cost motivated if the
autotransformer is for free) but this still won't
invalidate reason #2[/pre]

 

[fechter]

An autotransformer feeding a microwave oven transformer would supply the necessary isolation, so should work OK. A simple, heavy duty light dimmer might be just as effective and much smaller. All this will do is reduce the output voltage.

As far as I can tell, we want the work to heat up as rapidly as possible, so running full power and controlling the time is probably the way to go.

 

[cheapcookie]

thanks fechter and bedesign for the warning and explanation !

Would it be a good feature, for the the tool to shutoff after x seconds of resistance welding ?

Pulse for 2 seconds for example and then come to a halt, with maybe a reset button.

To minimize risk of mistake and applying heat for too long.

Hope this thread comes forward to change the way my battery packs are made

Edit: Reading the blog Galfish shared, the dude welds for 800ms on AA cells since 2009. Cool shit!

 

[fechter]

It might work to just use a simple timing circuit that controls a big solid state relay powering the transformer. Either a foot pedal or a push button would start the timing cycle. The amount of heat you get will be also dependent on the contact resistance, so might vary quite a bit depending on technique. At least with a timer you can limit the maximum. Ideally you'd want some kind of temperature measurement to control the shutoff.

 

[bedesign]

An autotransformer feeding a microwave oven transformer would supply the necessary isolation, so should work OK. A simple, heavy duty light dimmer might be just as effective and much smaller. All this will do is reduce the output voltage.

As far as I can tell, we want the work to heat up as rapidly as possible, so running full power and controlling the time is probably the way to go.

Would it be a good feature, for the the tool to shutoff after x seconds of resistance welding ?

[pre]We are possibly closer to having a working CONCEPT now,
but for the final scary detail to handle -

Most, if not all cheep light dimmers are switching the
power at any undefined times, which results in a train
of nasty very high voltage spikes
if used with inductive/transformer loads.

What is needed is the version of a TRIAC or anti parallel
tyristors with a drive circuit that always switches the
power (off) at the Zero crossings of the AC waveform for
a well behaved power/duty cycle control that is OVERLAYED
within the total ON-TIME setting, triggered by a foot
operated switch - yes, this gives full granular parameter
control as suggested by 'cheapcookie'

This could be made even more sofisticated if one gets too
carried away adding more complexity .. better stop there.

Then we are in bussines ... well if the capable members
here gets enthusiastic .. [/pre]

 

[spinningmagnets]

I just bought a cheap $20 tool-motor speed controller from "Harbor Freight" (and threw up a little in my mouth...I feel so dirty). It is labeled as a "router speed control", and the reason someone might want to slow down a high-RPM router is because you could better cut plastics and aluminum, etc. For wood? the stock higher RPM's are better.

Here is an indestructible showing the guts, and easy upgrades to cheaply fix its weaknesses.

http://www.instructables.com/id/Harbor-Freight-Router-Speed-Control-Mod/

I'm not saying one of these is a good idea to experiment with, so...proceed with caution. I am a verified and certified idiot when it comes to electronics, so...any help is appreciated...

Another tear-down video...

https://www.youtube.com/watch?v=n_uK2G4qFG4&t=688s

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Here's a video of the small $150 80A DC welder from Harbor Freight, it's weak for a welder, but maybe perfect for RS? (lowest setting appears to be 40A)

edit: these appear to put out 900V of DC, so I cannot recommend them for RS

https://www.youtube.com/watch?v=GRxrU6MP8hM

Inverters were introduced into welding power supplies first to generate direct current (DC) and later to generate AC. In these power supplies, incoming 50- or 60-Hz AC power first is rectified to DC and filtered and is then fed into the inverter section of the power supply, where solid-state controls switch it on and off at frequencies as high as 20,000 Hz, effectively converting it back to high-frequency AC.

This pulsed, high-voltage, high-frequency AC then is fed to the main power transformer, where it is transformed into low-voltage, 20,000-Hz AC suitable for welding. Finally it is put through a filtering and rectifying circuit to obtain DC welding current. Output is controlled by solid-state controls that modulate the switching rate of the switching transistors.

Because the power transformer runs at 20 kilohertz, it is much more efficient than one that runs at 60 Hz. This means the transformer can be much smaller and lighter, so the power supply itself can be lightweight. Inverter power supplies also "chop up" the incoming AC very finely, resulting in a steady DC without the typical 60-Hz ripple, so it produces a stable welding arc...

One of the videos on this welder indicated that if you truly want to weld with it, the leads need to be twice as fat (and he adapted cheap car jumper cables to do this), so...don't feel bad about chopping-up the stock cables to make RS leads.

 

[bedesign]

I just bought a cheap $20 triac ...

Here is an indestructible showing the guts, and easy upgrades to cheaply fix its weaknesses.

http://www.instructables.com/id/Harbor-Freight-Router-Speed-Control-Mod/

Another tear-down video...
https://www.youtube.com/watch?v=n_uK2G4qFG4&t=688s

The amount of heat you get will be also dependent on the contact resistance, so might vary quite a bit depending on technique.

Can confirm. The handset and electrodes are important for this reason, it's not just the power source.

Lower power, longer runtime, and a simple foot pedal allows you to observe what's happening and adjust. Not that automating higher power is wrong, just that it increases the need for feedback like temp sensing.

[pre]I had a look at the link with the heat sink modification.
As it was now mounted externally it should be examined
for isolation with margin from the line voltage.

Attached is an easy to make test circuit to detect
unwanted eventual voltage spikes generated by rough
TRIAC drivers into tranformers, as this could be
irritating for older people if conducted between
the right and left hand - or as a Twitter would
say "BAD"



Another aspect is EMC -
A rough test for that could be to scan the Medium
and Long wave AM radio band before and after turning
on the driver, to detect any significant difference.
There are filter circuits specified if needed.

Good with the input from real life experience with
connection resistance dependence -
perhaps an integrating current sensor could replace
a temp sensor in this case if the work pieces are
consistent, as it's a matter of transferred energy.[/pre]

 

[parabellum]

Bla bla bla

Is there something SPECIFIC you could tell about?

P.S. I think, everyone here knows how triac works, there must be some care taken at inductive loads (more voltage tolerance related). So, what???

 

[Overclocker]

[youtube]6V-XCkeC8Xk[/youtube]

good job! i think this will become more relevant once we see more high-power/bigger metal can cells where we need higher amps per connection

 

[bedesign]

Here's a video of the small $150 80A DC welder from Harbor Freight, it's weak for a welder, but maybe perfect for RS? (lowest setting appears to be 40A)

https://www.youtube.com/watch?v=GRxrU6MP8hM

[pre]Things to check for first, as this might be how it behaves -
A welder internal impedance is designed to match the arch type of load.
To get the arch going and maintained the typical voltage is 50V
I can imagine some spectacular firework if a welder arch is
ignited on a LiIon with it's thin metal can.

Shorting for resistance heating would throw it off it's
designed envelope and make it inefficient with heating
internally instead, or a state of the art regulation may
turn down the current.

A modified MOT with low 2 to 3 volt output will match
the transformers internal impedance much better for
this kind of short circuit.

An adopted design with current limitation used for modern PC PSUs
could be interesting for heating applications, as they deliver
low voltage with 10s of Amps in a small volume.

Let us know the result when anyone have tested the
'Spike' detector circuit.[/pre]

Bla bla bla

Is there something SPECIFIC you could tell about?

P.S. I think, everyone here knows how triac works, there must be some care taken at inductive loads (more voltage tolerance related). So, what???

[pre]Yes, why having a discussion at all if 'everybody' knows this?
The 'un specifics' in the thread for the moment is about
making informed choices for construction and purchases
and not getting killed too early.
When this state has matured, evolving into
specific design details could come naturally.

If you know all this, and don't want to tell us,
just skip over it
If one is to impatient to have the fundamentals
sorted out first the result will be "BAD"

Hint:
This company have ready made units for delivery -
Ready Made Heater

They will be happy to shut up and take the money
Hope this is specific enough ...[/pre]

 

[spinningmagnets]

Yes, a previous poster stated they achieved good results with an "American Beauty" unit at $450 that provided 2.7V @ 38A (2-second pulse, timed by eyeball). I can afford $450, but I am hopeful that I can find some way to achieve "acceptable" results for less than $200 (and then to publish results)...

 

[vex_zg]

Yes, a previous poster stated they achieved good results with an "American Beauty" unit at $450 that provided 2.7V @ 38A (2-second pulse, timed by eyeball). I can afford $450, but I am hopeful that I can find some way to achieve "acceptable" results for less than $200 (and then to publish results)...

can't we just use a MOT wired with bit thinner wire so not too high current is generated?

what happens if we wire MOT with, cca 2 turns of AWG12 wire?

 

[spinningmagnets]

The American Beauty website has a lot of instructional and educational information, and they stated that the output of their machines is AC, so a MOT with some way to adjust the input (dimmer switch?) seems like it would be a very viable option.

Over in the MOT thread, they are using them for spot-welding, and the struggle they have is to have an adjustable timer that is reliable and can actuate in the millisecond range (150 ms?). An interesting facet of that research was that in order to have the extra-high amps needed to get the acceptable amount of energy when a pulse only lasts 1/4-second, was to find a very large MOT, and 1500W industrial microwave ovens are desirable (although it is possible to series two smaller MOTs, or stack two small "E" sections to make one large MOT).

Of course the smaller microwave ovens are much more common and would likely be acceptable for a 3V / 40A pulse of 2 seconds. Another bonus is that such a device would put less strain on the homes circuitry, since the common household breaker (in the US) is a 120V AC X 15A = 1800W max range. The problem is not just the total 1500W power of a large MOT spot-welder, it is the sudden pulse that strains a home power system, and can trip the breaker.

3V of 40A is only 120W, and I have had good results with my ancient 100W soldering iron when assembling harnesses and connectors for 10-ga wire. I suspect a small MOT (very cheap and often free) would be a great candidate for a home Resistance Soldering Unit (RSU). I have been scrolling the web for info, and the term RSU is sometimes used in industry, and by hobbyists.

edit: If I assembled a MOT-based RSU, I'm sure the 120V AC input would turn on and off with a common tool foot-switch (sold for routers and table saws). Would a common dimmer switch on the input adjust the volts or the amps? (edit, youtube suggests that 'most' dimmers chop the voltage in some way) How would a dimmer on the MOT input affect the MOT output? The big draw for me when I got the small welder above, was that it is low volts, and clearly states that the dial adjusts the amps, right out of the box.

I have several simple DMM's to check volts, but what's a good device to check amps, one of those ring-meters that "clamp" over a cable?

edit: 1.5 wraps of wire on the output of a 120V AC MOT should provide 3V, and the thickness of the wire should affect the output amps, so swapping different sizes of wire could be used to make a simple RSU, once we have stumbled across the right gauge of wire that gets the amps you want. However, I like very thick wire on the output so the cables don't get hot, and I am wondering is there's a cheap and easy way to adjust the output amps from 20A-60A, since 40A seems to be good for an initial performance target.

 

[fechter]

I've used dimmers with mixed results. In theory, the MOT while heating would resemble a resistive load but the inductance of the windings will create some extra voltage transients. They make dimmers designed for motor speed controls that are more likely to survive. Since the soldering pulse is relatively short, triac heating shouldn't be a problem.

 

[vex_zg]

I've used dimmers with mixed results. In theory, the MOT while heating would resemble a resistive load but the inductance of the windings will create some extra voltage transients. They make dimmers designed for motor speed controls that are more likely to survive. Since the soldering pulse is relatively short, triac heating shouldn't be a problem.

MOSFET bridge on secondary coil to serve as a switch? Can mosfet in bridge mode be used to switch AC ?

(I never used mosfet in bridge configuration, only p-mosfet or n-mosfet)