Spot welder for batteries?

Does anyone here have information on what equipment to buy in order to spot-weld battery packs? Yes, I know I can get batteries with spot-welded tabs, and I know I can solder them without too much damage to the cell.




I hope you mean soldering to the welded tabs - you should never solder directly to a storage cell.

We have a store nearby called Interstate Batteries, they have a little shop in the store where they can assemble battery packs to order on the spot. They let me watch while they did the spot welding with a specially made (I like the British term "purpose built") spot welder. They gave me a handful of the tabs because I told them I wanted to experiment with doing spot welds on old batteries using a small arc welder that I have as a current source. However, I didn't get any details such as the current they are using and the time the current is applied (these both seemed to be adjustable on the special welder). The machine was bolted to the table and had a special head which incorporated the electrodes that contacted the workpiece. The electrodes appeared to be cast in thick copper. One electrode was of a small diameter that came down in the middle at the point where the spot weld would occur. The other electrode was shaped like a yoke that surrounded the small electrode, it made contact with the workpiece in a larger area. I suppose this arrangement is so that the current density will be lower where the larger electrode touches the workpiece, and much higher where the smaller electrode touches, so the heat will be generated in a confined area. The operator used a non-conductive block under the cell to raise it to a certain height. Then he pressed a lever which lowered the head down onto the top of the cell (with the tab in between the cell and the head). This allowed him to make sure he had the head positioned properly, and then he applied a little more force to press the larger electrode down to make good contact. The final step happens quickly so I'm not certain how much is controlled by the operator's hand and how much by the machine: more pressure (or pressure on another lever) pushed the small central electrode down onto the workpiece, the current was applied (here I'm not sure if the current started because of the electrode making contact or if a timing circuit is in control), then the operator released the pressure pushing down the central electrode. I have ruminated over this (haven't tried to do it yet) and I think a critical step is removing the central electrode as soon as possible so it doesn't get joined in the weld. Also, that the other, larger electrode is firmly clamping the tab onto the top of the cell both before and after the application of the current. BTW, at the beginning of the whole operation, the operator cleaned the bottom surface of the electrodes with a piece of emery cloth.

I would like to know what alloy the tabs are made of (somehow I think it is mainly nickle), also the metal used in the cell can and positive terminal (looks like stainless steel).



The strips are indeed a nickel alloy; soft and flexible-- and the cans of batteries are usually steel. The interface between the two, whether welded as you are talking about, or in the case as when steel is plated by nickel, is very stable and corrosion-resistant.

The spot (or "resistance") welder is not all that special (the one at work is made by a company called Unitek); but the algorithms for the pulse intensity and duration, etc. are pre-programmed in the power-supply/control-head unit.

When spot-welding nickel strip to steel button-batteries, the operator will use a "lighter" setting on the machine than, for instance when trying to tack a steel mounting tab to a steel "D" cell. The spot welder can do both.

The operator knows what schedule to use on the machine, and how to adjust also the triggering-pressure, which is important, since too much pressure will result in material bending and deformation, but too little will cause excessive sparking and melting only the surface rather than the "interface" (a little spark-spitting is not terribly critical).

As said, the welder is a commercially-available stock item, but it isn't really cheap. You buy different electrode configurations as desired for particular jobs. Ours has two pointed copper electrodes spaced about 1/16 of an inch apart at the tips, the end of a very sharp "vee" of maybe 12 to 15 degrees. It is quite suitable for rebuilding packs and adding tabs to batteries. I've seen the horseshoe electrode arrangement, but never used it myself, so I don't know what advantages it may or may not have. You have to have a file and/or emery paper around to clean the surfaces and the electrode tips, which get fouled fairly easily.

I'd love to have a spot-welder myself, but the one my boss bought was around $10K-- so I've often wondered if there was a cheap model out there that would do a reasonable job for low-volume battery pack manufacture.

Funny-- ours says it is Dual-Pulse, but you can't tell there is more than one event per trigger. They must be fairly quick. Also, any nickel strips you have laying around on the desk close to the heavy-guage cables that attach the electrodes to the power-supply/control unit, will jump up and dance in response to the pulse.

Would not like to test the pulse on a ring or, for that matter, on a ring-finger. I was told my predecessor at the shop managed to shock himself, but don't really have the details... I have hit the trigger with the electrodes on a wooden block, but the machine dumped its pulse internally and displayed an error message rather than arc into the wood-- so maybe the safety features are what costs so much. Certainly it is safe enough for a klutzy goof like me to use it regularly for a couple of years and not get into trouble....


I might be mistaken about the horseshoe shaped electrode. I looked at the cells they welded for me and the spot welds look like they are always in pairs, about 3-4 mm apart. Do you have any idea approximately what current and time would be correct for the thin battery tabs?


Ellis, I'll look at it at work sometime this week. I have never really studied the process fully; so voltage and current and pulse times are not in my head. But it should be (semi) obvious if I browse through the schedules menu.

I do know that the whole of the issue comes down to localized, quick heating, and then cooling. This is the reason soldering is NOT a recommended method of joining cells; you can't localize the heat required nor dissipate it quick enough to avoid damage to the cells, and still solder well.

But a spot weld is so quick and small, the majority of the activity caused by the heat is in melting the two metal pieces together, and all around that small spot is cooler metal, helping to solidify the joint. A second after the weld, it is cool to the touch. You'd have to "bump" it several times in a row to get the weld area hot enough to burn your finger afterward; or you'd be welding it "wrong" (i.e., not enough pressure, too much current, too long a pulse).

The coolness is part of the reason, too, why the copper electrodes don't stick to the joint. If the joint isn't cooling fast enough (in milliseconds, really), then they DO stick. That's an indication of too much heat (back off pulse time or current slightly). In automated systems, it often happens that many welds are made one after another, in which case they usually cool the electrodes with internally-piped water systems, so the electrodes remain cool-- otherwise they stick and hold to the joint.

I'm guessing that the pressure, squeezing the pieces together, is important mainly in that it places the spot where the two pieces are going to be welded into a tiny footprint where the resistance will be less than anywhere else, and so the current will all go into that tiny spot. So there is an optimal pressure for the gap between electrodes and the thickness of material to be "penetrated" by the current.

Well, here I am ruminating on the process, and don't have all the facts of it... I'll see what I can get for you off the machine, and maybe (though this is a long-shot) if my boss has a data-sheet on it, I'll ask him to hunt it out for me.

It would be *fun* perhaps to make some hobbyist version (modified stun-gun? camera flash unit? hehe) that maybe could be put together very cheaply, and offer a duty-cycle of only a few welds per minute-- not right for industry, but good enough to put together the odd battery-pack or two for calcs and shavers and cordless tools. Use it a few times a year 'round the home and it might well pay to have done the project....


glynn, that's what I'm hoping to do. I have a small arc welder with output adjustable from 30 to 100 Amps. The open circuit output voltage is about 35V AC. To control the time precisely (if necessary), I plan to use a solid state relay or my own Triac circuit on the 115V primary side. If I build my own circuit I will also be able to control the current down to lower levels with phase control. Once I have the heavy duty electrodes built, I can sense when the electrodes contact the workpiece with a continuity test and use that signal to turn on the timer for the current pulse for any number of AC cycles.

If the actual current required is considerably less, it might be interesting to try the camera flash approach: a large capacitor and a heavy duty relay to turn on the current. Then the current-time product can be controlled by just charging the capacitor to a certain voltage. I've nearly welded screwdrivers to capacitor terminals plenty of times!



Took a cursory look at the machine today. It says it is a "Unitek/Miyachi 250 Dual-Pulse Stored Energy Power Supply".

By "250" and "stored energy" I sort of assume that the number is a wattage rating, ie. volts X amps in DC, and some sort of a capacitive-discharge circuit...

And I don't know what voltage, it doesn't say, but it doesn't seem to be high. Like maybe 3-5 volts tops. But 250 divided by 5 would be 50 amps; divided by 3 would be 83.3; that would make a pulse certainly strong enough to melt a spot in light steel or nickel.

The stored schedules in the machine have the following format: Schedule (#__), Pulse 1= (xx)%; Pulse 2=(xx)%; Polarity=(+/-); Width=(Short/Med/Long).

I wish it had more detail onscreen but was Really surprised that it had only those parameters. That the polarity on all the settings I scrolled down was Positive, and the Width setting always "Short" was another surprise; I suppose that since this particular machine was set up for light jobs like battery-packs etc., many of the settings will be only subtly different from each other.

Schedule 1: Pulse 1= 3%; Pulse 2= 9%.

Schedule 2: Pulse 1= 5%; Pulse 2= 15%.

Schedule 3: Pulse 1= 7%; Pulse 2= 22%.

Schedule 4: Pulse 1= 8%; Pulse 2= 26%.

Schedule 5: Pulse 1= 9%; Pulse 2= 28%.

Schedule 6: Pulse 1= 10%; Pulse 2= 11%.

Schedule 7: Pulse 1= 10%; Pulse 2= 30%.

Schedule 8: Pulse 1= 10%; Pulse 2= 35%.

Schedule 9: Pulse 1= 12%; Pulse 2= 36%.

Schedule 10: Pulse 1= 15%; Pulse 2= 35%.

Schedule 11: Pulse 1= 15%; Pulse 2= 40%.

Schedule 12: Pulse 1= 15%; Pulse 2= 45%.

Schedule 13: Pulse 1= 20%; Pulse 2= 45%.

Schedule 14: Pulse 1= 30%; Pulse 2= 50%.

Schedule 15: Pulse 1= 30%; Pulse 2= 65%.

Schedule 16: Pulse 1= 50%; Pulse 2= 77%.

Schedule 17: Pulse 1= 2%; Pulse 2= 4%.

Schedule 18: Pulse 1= 1%; Pulse 2= 5%.

Schedule 19: Pulse 1= .6%; Pulse 2= 0%

And I think there are some dozen more settings on the machine, but that's all I wrote down today.

Say the machine puts out 50A at 100%. That's only .5A per percent, right? So the schedule I use to stick nickel strips onto (AA,N,C,D) batteries with, Schedule 3, kicks out a pulse of 3.5A, then follows with a pulse of 11A. It does so fast enough that you only hear it as one small "thunk".

Well, I still haven't learned much, I know. I will try to follow up with more info-- I am certain that somewhere buried in my bosses office there is an owner's manual. Maybe I can find it...

Ellis, having played with a photo-strobe power-pack, I know it ramped up from a 12v battery to supply some 430v or so, but not many milli-amps-- that's good because while I was measuring it, it bit me. Ouch.

But it should be not terribly hard to do the opposite: a bank of capacitors storing low volts but together punching forth high amps.

Probably the tricky part is to make the only part of the circuit that has, as its highest RESISTANCE to flow through, a junction of pressed-together metals (the intended weld-spot)-- if the legs on the capacitors or the controlling relay have higher resistance than that, it is they that will melt!!

TTYL.-- grh.


that's a lot of information! DC, low voltage, double pulse, and those ratios - especially the ones you use for a similar application. The double pulse reminds me of the pre-heating required before infra-red soldering of printed circuit boards. It is needed because all the moisture must evaporate from the solder paste before it melts, otherwise there is a lot of spattering of tiny solder balls all over the work. I notice that most of the schedules have the first pulse at a lower level.

And the manufacturer's name. I found a website, ( ), it is "A Unitek Miyachi International Company". I haven't found any "stored energy" power supplies there, but they have some DC supplies under "inverter weld controls". I downloaded an accessories catalog that has pictures and specifications of electrodes. Under "Benchtop System Solutions", the different families are divided by the amount of mechanical force they apply to the workpiece, from 70 to 700 pounds. These must be for bigger welds that batteries.

Another website, ( ), has laser welders and "fine spot welding" equipment including two "capacitor discharge" power supplies: MC80B and MC160B. They are single pulse supplies but unfortunately I can't read the specs in the .pdf file because of a missing font! This site has a page of techical information ( ). One of the articles linked from that page ( ) has an interesting picture of the waveforms of different kinds of welding power supplies. Here is what another article linked from that page says about Capacitor Discharge welding:

2) Capacitive Discharge Welding Power Supplies
By discharging the energy accumulated in the capacitors. They are suitable for welding materials with high thermal conductivity like aluminum and copper. Because the system stores the energy in the capacitors, it does not require a big capacity for the primary power source.
On the other hand, the current rise is very steep and can not be controlled, and it may cause a expulsion. In this case, higher squeeze force must be applied. Also, Peltier effect which must be controlled somehow because it is direct current.

By "expulsion" I guess they mean expulsion of molten metal from the weld site, because of pressure due to the increased temperature. The "Peltier effect" I assume is related to Peltier devices which are the semiconductor devices that move heat from one side to the other when current passes through. They are used to cool CPU's when overclocking and also for the little six-pack coolers that run on 12 volts DC. "Controlling the Peltier effect" might be what the polarity control accomplishes, I know with Peltier devices, what direction the heat moves depends on what direction the current flows.

I have already started collecting parts for a fixture to remove short circuits from NiCads, with a large capacitor and a heavy duty contactor type relay to quickly connect it to the cell to be "cleaned". The same type of relay (rated 30 Amps continuous) might work for welding if I use copper bars to make the connections. I guess SCR's would be faster, bounce free, and maintenance free, and 30 or 50 Amps isn't a terribly big SCR. Here's an idea I had to generate a double pulse: two capacitors and two relays (or SCR's) triggered one after the other.

One fine point: "watts" are a measure of power which I think would be a unit of the rate of energy consumption and watt-hours or watt-seconds - the integral of watts - would indicate the amount of energy delivered to the spot. With a capacitor, the product of the capacitance and the voltage is the number of Coulombs stored, one Coulomb being a huge number of electrons - the same number moved by a current of one Amp flowing for one second. By quickly connecting the capacitor to a short circuit, all those electrons will flow out in a very short time. But the total number of electrons involved in the discharge is determined by the voltage to which the capacitor was charged in the first place. In the case of a camera flash, the capacitor is charged to several hundred volts by a switching power supply. The capacitor is connected to the ends of the flashtube all the time but it is normally non-conducting. When the camera gives the signal, an SCR (if I am not mistaken) connects the primary of a step-up transformer to the capacitor (or to a separate capacitor similarly charged). The secondary of the transformer quickly develops a pulse of several thousand volts which is applied to the third, trigger electrode of the flashtube, which causes the gas inside to start conducting, then all the charge in the capacitor flows through the flashtube at once. Actually, the circuit must have some way to stop the flashtube from conducting before the capacitor is discharged, because when I set my flash to automatic, it recovers and is ready to flash again very quickly, so the capacitor must not have been completely discharged.

I can see that the critical thing will be to have all the conductors very large so that the spot-weld joint is the highest resistance in the circuit, as you point out. In the Miyachi catalog, all the conductors are made of wide strips of copper. As I recall, the spot welder at the local battery store had wires similar to those used for a car battery connecting the head to the power supply.


While you can blow shorts out of Nicad batteries, it is really not a good idea. Once a battery starts growing shorts, it is on its last legs. They always grow back in a short period of time. Zapping them only makes sense as a short term emergency situation fix where you can't get to the battery store and just have to have something running.

Possibly Related Threads…
Thread Author Replies Views Last Post
  HP-67 Shiny Spot on Face Colin Verrilli 8 2,639 08-09-2012, 05:20 PM
Last Post: Colin Verrilli
  For the Collectors (Spot the 35 & 48) Les Bell 1 1,048 10-11-2008, 02:55 AM
Last Post: Walter B

Forum Jump: