DIY spot-welder, transformers from microwave ovens?

spinningmagnets

Moderator
Staff member
Joined
Dec 21, 2007
Messages
12,952
Location
Ft Riley, NE Kansas
There are hundreds of used transformers on ebay, $20-$40...but which one to get, what specs? (sometimes listed as MOT/Microwave Oven Transformer)

I have been watching the DIY spot-welding threads, and I am now very drawn to the "Spot Welding Copper Strips to 18650 Battery Cells" thread ( https://endless-sphere.com/forums/viewtopic.php?f=14&t=84680 ). Copper is cheap and available, nickel is expensive and will become worse over time. Nickel has been used because it is the common material in factory battery packs, but...those packs are usually low amp. I contend that copper is the future for high amp DIY battery packs.

That being said, due to it being uncommon, there are no mass-produced copper spot-welders, so the few models specifically made for that are expensive. Copper requires more energy than spot-welding nickel.

edit: the two magnetic shunts shown in these videos MUST be removed for what we want, they would limit the amount of current. We want to start out with "too much" current and then adjust it down to where the device works for copper or brass spot-welding.

Transformer1.png


I have seen videos of using a transformer that was scavenged from a trash-day microwave oven. The one in the pic above was a wall-mounted built-in unit, and it was large and 220V. I suspect that the more common smaller microwave ovens that use 120V might be very inadequate. The spot-welding of copper requires very high current, but doesn't need high voltage. Copper absorbs heat fast, and conducts current easily. Nickel has enough resistance that it welds easily, since the welding current causes localized heat, but...of course, we don't want resistance...Using thicker and wider nickel strips to help battery current only helps up to a point, resistance is bad...

Using the resistance of nickel to get enough localized heat to create a spot-weld makes life easier for the manufacturer, but its a bad design for performance. Since copper doesn't have that resistance to help it create localized heat to create a weld, we must use actual high currents (the high-temp tungsten probes even glow).

This video shows how to remove one of the two stock windings of the salvaged transformer, and he then uses only TWO wraps of very thick cable (The fattest that would fit) and doing that seems to work well for producing low voltage/high current at the probes...

modifying a salvaged transformer (5 minutes):
https://www.youtube.com/watch?v=d5pGN6pqkyY

Constructing a spot-welder (9 minutes):
https://www.youtube.com/watch?v=vrlvqib94xQ

I think it would be easy to make a simple timing switch on the 120V AC side of the input. That way, we would hit the foot-switch, and we would get a pulse. Once the pulse is timed properly, no further adjustment would be necessary, if we only weld the same thickness of copper strips.
 
Before you mess with the transformer, i suggest that you do some science: get salty water, soak a wooden board, hammer in the two electrodes, and stand well back while you pump 6000v of microwave power and do lichtenberg lightning figures in the wood. If you are getting good lichtenstein patterns, i'd say that your transformer is good. don't step in any salt water.

https://youtu.be/b__z2apE7dA?t=1230
 
I'm certain a precise timing device would be useful, but the main benefit (IMHO) in being able to adjust the timing, and also adjust the current, is that it allows us to weld a variety of materials and thicknesses.

I only want to weld copper, and I only want to spot-weld parallel connections that are thick enough to run cool at 30A per cell. I may have to change the winding in the transformer, which of course is much more troublesome than just spinning a dial, but...I only have to figure out what works once. Then, it can be replicated easily and cheaply.
 
I understand the current must be much higher to decent weld copper, so one turn would seem to work.

Maybe if you source two MOT, and build one with two turns for welding Nickel and one with one turn to weld copper...
Then you only need to switch out the controller, instead of rewiring the MOT.
Then you can weld a variety of materials, if you have a adjustable timing controller like the Arduino based one.

Now if it doesnt work with one 120VAC MOT, you can always connect a second one in parallel?
 
No reason you can't have two separate secondary windings and swap connections to put them in either series or parallel for welding copper or nickel?
 
Great idea! I just looked again, and here's a youtube of a guy who made a single MOT spot-welder. He wasn't satisfied with the current, and stacked a second one to increase the amps on the output:

https://youtu.be/neR_RFTblqg?t=368

This may make the common, cheap, and small MOTs a viable option. Although...the smaller the physical size of the MOT, the size of the secondary cable becomes limited.

The voltage is the inverse of the amps, and since the spot-welding probes have physical contact when welding, the actual voltage is almost irrelevant (meaning, low volts should be OK for what we want to do). One of the benefits of buying a used MOT is not only the cheap price, but the primary winding is already made for us, and the new secondary (for high-amp welding) is a 2-turn thick cable, which is very easy to do.

Stacking MOTs in series/parallel to increase amps is awesome. But...it is immediately obvious to me that a physically larger MOT will result in higher amps, if only because it allows a fatter cable on the new secondary loop. That being said, I wonder if it is better to look at 220V input MOTs, or 120V MOTs?

The microwave generator operates off of very high voltage, so as a result, the stock secondary coil has many turns of thin wire, but...if "X" amount of output voltage is required, Does that mean that a 220V MOT would have a primary coil with more turns of thinner wire (compared to a 120V)? If yes, then a physically large MOT that has a stock 220V primary, would result in the modified version having a greater difference in the number of turns between the primary/secondary.

The greater the difference in turns between the primary/secondary coils, the more amps we would get from 2-turns of very fat cable. In one of the videos above, several bundles of stranded cable was stripped of its insulation and re-bundled into a thicker version, using only electricians tape as insulation. This one tip alone makes it easy for us to find cheap stranded copper wire, and "make" a low turn-count thick cable, without buying expensive welding cable. In fact, thick insulation limits the copper cross-section of the secondary coil.

I am drawn to the idea of using a MOT that is large, and winding the secondary coil for "too many" amps, because we can easily increase the number of turns on the secondary by swapping to smaller diameter cable, in order to "fine tune" the amps (more secondary turns = lower amps, would also result in higher volts but the exact amount of volts is almost irrelevant). If I start with a MOT that is smaller, I am limited on how high I can take the amps. Falling short would mean I have to start all over with a new MOT.

Here's a question...is there any benefit to using insulated magnet wire in the secondary coil (each strand is sealed in a clear epoxy)? Or is it acceptable to use bare strands found in common wire (which results in the strands acting electrically like a thick solid wire, but more easily bendable)?
 
A 220V MOT of the same wattage will indeed have twice as many turns on the primary but at half the conductor strand size. However, the current available with a rewound secondary is going to be the same: like different motor windings the power output is the same, as its the same amount of copper just arranged differently. The amp-turns will be the same for both the 110 and 220V primaries: the 220V winding will be half the current for twice as many turns. That creates the same amount of flux to induce current in the secondary.
 
Using a 240v 1000w MOT I found that each wind on the secondary gave about one volt. 1.5 wraps was all I could fit with fat cables=1.5v on output. This was ok if you were using copper electrodes but won't push current through anything else. I tried using smaller gauge and doing 2.5 wraps but didn't seem to help (halved current). 1.5 wraps was better for copper bc resistance is so low. I went to a bigger ~5000w ups transformer that I rewound to put out 6v. Incedentally , the mot draws about 750w no load, and at about 800w, draws about 1000w.
 
MOTs do draw excessive current at no-load due to their (cheap) design. IIRC the lack of iron results in saturation of the primary at low load, but this is normally OK as they're only run at full load in an oven.
 
spinningmagnets said:
Here's a question...is there any benefit to using insulated magnet wire in the secondary coil (each strand is sealed in a clear epoxy)? Or is it acceptable to use bare strands found in common wire (which results in the strands acting electrically like a thick solid wire, but more easily bendable)?

No, I don't think so since it will be running at 60Hz. At much higher frequencies, there would be an advantage due to skin effect. At 60Hz, you just want as much copper as you can stuff in there.

There will be sort of a sweet spot for the windings count. Too few, and you will have plenty of amps but the voltage drop in the probes, connections, weld area, will not let you reach the potential. More volts and you will overcome the resistances better. Too many turns and you have plenty of voltage but now limited by the resistance of the windings.
 
I think I'll just get the biggest MOT I can easily find, and then try 2T, 4T, and 6T on the secondary coil...test and see results.

I am now committed to trying this out, so...make any suggestions now before I buy. The reason I suddenly lost any reservations is because...even if it proves to be problematic for spot-welding the copper paralleling strips onto the nickel-plated can ends of the cells (which remains to be seen), I now believe this is going to prove to be the ideal way to connect the parallel strips to the series bars (as opposed to soldering with my 100W iron).

One of the videos linked above shows the guy spot-welding two large nails right where they cross when held together in an "X" shape. The sudden high current makes the joint glow orange, but the tips of the nails are being held with bare fingers, because the weld is over-with quickly, and he has plenty of time to set the nail down before the heat travels far enough to be an issue.


SpotWelder1.png


The red arrow shows the single nickel series-strip connecting one 4P module to the next (weak design, high resistance, only good for low amp pack output). The yellow arrows show the thick copper series flat bars I am suggesting (one bar per every two cells), and where the copper series bars cross the nickel strips is where I would definitely use the high-current MOT to weld them together. If welding the thick copper flat bars to the paralleling strips (whether the thinner strips are nickel or copper), If the heat is an issue (copper will react differently than the high resistance steel nails in the pic above), then this part of the operation can be done in a jig, before the parallel strips are spot-welded onto the cells.

BatteryBusBars (1).jpg
 
Best vid I've found on MOT as of yet https://www.youtube.com/watch?v=KRoPHKpCYmg

From my understanding, the size of a MOT will mean more power handeling capacity.
The biggest MOTs I can think of are rated for 1500 VA (volt x amps).
In theory, 1500 VA (or 1.5 kVA) would be equivalent to 1500 W of power.
In practice, MOTs are the cheapest big transformers out there and they have low efficiency and get warm/hot quickly (we usually dont power microwave oven for so long so nobody cares about them getting hot. Plus theres a fan to cool them down)
So if you have 80 % efficiency that's typical of MOTs, 1500 VA is really only 1200 watts of potentially usable power at the secondary coil.
The gauge of the secondary coil wire (thin lacquered-insulated so you can pack in more copper wire turns) also limits the power output, since the gauge size will dictate the maximum ampacity through it. Exceed that ampacity and you risk the insulation of the coil melting down and the coil will short out onto itself. Result : from too few "effective" coil turns in the primary mean you will also short your mains from the now lowered inductance, and (hopefully) trip your mains fuse. So in short, dont expect a one turn primary coils made out of 10 AWG to deliver 1200 A at 1V, even if the transformer has a big 1.5kVA rated iron core. That wire is too small and will just melt.

In theory, if your primary coil is 120 turns (often is in MOTs) and you run 120 Volts, then using that potential maximum 1500W will require 12.5 full amps through your mains, which should be okay since you probably have a 15 amp fuse or braker. The same MOT in europe would require only 6.25 amps on your mains at 240 volts.
Anyways at this transformer size scale range, the voltage will not be a bottleneck be it 120 volts or 240 volts. Even with low volts (i.e. : 120 volts), at 15 amps, you could run 1800 W through the primary without tripping the fuse or braker. Upgrade for a 20 amp fuse circuit and we are talking 2400 W.

But 2400 is more than you need since most MOT are between 800 VA and 1500 VA.

Of course if you run 120V in a primary that is 120 turns you will get 1volts in a 1 turn secondary. And if you can get 1200W out of a huge size 1.5kVA MOt. Wow ! 1200W at 1V means 1200 amps ! Well it would if your one turn secondary coil has a huge AWG gauge wire to support such a high current.

Why not use 24 volts at 50 amps instead... (24 turns of still quite big gauge, but you wont loose precious space from the wire insulation if you use varnish coated transformer wire instead of traditional wire).

I'm not so sure voltage drop the tip of the electrodes would be such an issue at 24volts (still safe to handle 24volts, even at 60amps). Even at 1volt, the drop is only as significant as the lenght of the electodes. Make the electrodes shorter and part of your problem is solved since voltage drop depends on the length of the conductor, not just its resistance.

I really encourage you to look at this vid if you want to play with MOTs : https://www.youtube.com/watch?v=KRoPHKpCYmg

Dont forget. When you open that old microwave to salvage a MOT, dont touch that capacitor... they often are still holding a charge at around 3000 to 4000 volts.... at those volts it only take a few micro-amperes through your skin to put your heart into ventricular fibrilation (malignant arythmia) and KILL YOU. Short out the capacitor to dischrage it but use adequate protective equipement to isolate your body from the potentialy lethal electric shock the capacitor could deliver. This is no joke ! 120 volts is a walk in the park compared to the voltage those capacitor holds. Same thing with unmodified MOT.... their secondary will put out 2200 to 2400 volts AC. Watever you do don't plug a MOT in your mains if they still have their original high voltage secondary coil. You could easily kill yourself with these 2.2 kilovolts. Finally remember that electrecity at these high voltages CAN JUMP TROUGH AIR and shock you even if you dont really touch the electrical contacts. This pnenomenon is called high voltage electrical arcing. The higher the volts, the longer the electrical arcs can be.
 
Make sure to also see part two and three of this video, as there is usefull info on rewinding and wire sizing.

https://www.youtube.com/watch?annotation_id=annotation_163168&feature=iv&src_vid=KRoPHKpCYmg&v=YX3mYO-BIuQ
https://www.youtube.com/watch?v=-NLy-LL_TGQ
 
Not sure I follow you matador... The whole point is to have a fat wire at huge amps, not 24 v @ 50amps. Maybe I'm missing something, but we are talking about spotwelding. 50amps? That won't even warm your work peice!
 
kdog said:
Not sure I follow you matador... The whole point is to have a fat wire at huge amps, not 24 v @ 50amps. Maybe I'm missing something, but we are talking about spotwelding. 50amps? That won't even warm your work peice!

You're right. I get afterthoughts now... :oops:

I mean I thought the heat is just power.... As in P = V x I.... So my rational was increase V and reduce current (I), so you get the same amount of Power...

Well I now realise that my rational was not good : Yes you get the same amount of power, but it's a very smaller percentage of that power that is actually converted to heat, as the voltage drop is relativey smaller (in %) when you start from 24V than when you start at 1V. The heat used for welding is not P.
Rather it's the voltage drop (from the resistance of the thing you try to weld) that results in a small loss of power.
So the calculation should be dP = dV x I = Resistance x I^2. (dP is power drop aka dissipated heat and dV is voltage drop and Resistance is from the conductor where the voltage drop occurs)

So if the strip has a given constant resistance, dP (heat) will be propotionnal to the square of the current.

Your right, current increases the heat exponentially !....
Now I'm thinking when I have my mot, I should fill it up with a 1.5 turn of square section copper bussbars bent at right angles and heat shrinked... I could fit more copper in the MOT if the wire is square rather than round ! :D
 
Matador, the video you linked to was very helpful. The MOT we want is the big one from a 1500W microwave, very large. The used MOT's on ebay sometimes have model number, and in this way we can look up the size of the oven.

However, this cheap style of MOT is called E-I for the shape of the lamination stack sections. He says if you want to make a super transformer, take two MOT's that are identical, and only use the E-E sections facing each other. Take the two primary coils and connect them together (I think it is in parallel so turn-count does not change), and then you have more room for copper in the secondary. Would connecting the two identical primary coils in series be better, for what we are doing?

Transformer2.png

You can always cut back on the amount of copper in the secondary, but once it is full, you cannot put in more. If current is higher with this configuration, may be the pulse can be shorter? I think that would be good to help it run cooler.

A microwave of any size will have 120V/240V switches, plus it will have a fan that can be reused to cool the MOT. Enshroud the MOT, then...one fan pushing air in, another fan pulling air out?

Warning: I assume the 1500W limit on the size of the most common large microwave ovens is due to the size of the breaker on the average kitchen outlet 120V X 13A. If you want to mate two 1500W MOTs, you will need a 120V X 30A breaker (or more), but...this "E-E" style might be useful to mate two 700W MOTs, if its hard to find a single 1500W MOT.
 
spinningmagnets said:
Matador, the video you linked to was very helpful. The MOT we want is the big one from a 1500W microwave, very large. The used MOT's on ebay sometimes have model number, and in this way we can look up the size of the oven.

However, this cheap style of MOT is called E-I for the shape of the lamination stack sections. He says if you want to make a super transformer, take two MOT's that are identical, and only use the E-E sections facing each other. Take the two primary coils and connect them together (I think it is in parallel so turn-count does not change), and then you have more room for copper in the secondary. Would connecting the two identical primary coils in series be better, for what we are doing?



You can always cut back on the amount of copper in the secondary, but once it is full, you cannot put in more. If current is higher with this configuration, may be the pulse can be shorter? I think that would be good to help it run cooler.

A microwave of any size will have 120V/240V switches, plus it will have a fan that can be reused to cool the MOT. Enshroud the MOT, then...one fan pushing air in, another fan pulling air out?

Warning: I assume the 1500W limit on the size of the most common large microwave ovens is due to the size of the breaker on the average kitchen outlet 120V X 13A. If you want to mate two 1500W MOTs, you will need a 120V X 30A breaker (or more), but...this "E-E" style might be useful to mate two 700W MOTs, if its hard to find a single 1500W MOT.

Here, the 1st of July is the day that people move if they have to change appartement rental lease. So the first of July, you can find a lot of junked old microwaves in the trash. I went around and picked up two MOTs before the city garbage truck could have the chance to scavenge them. I would not buy one eBay... People there inflate MOT prices and shipping those heavy irons can get expensive. MOTs are quite ubiquitous.... To find a big MOT, i just look at the max power rating that's on the sticker on the back of the microwave. Don't forget to discharge te capacitor by shorting the leads before touching anything inside.

For our purpose, I think two MOTs wired in parrallel is good enough (make sure the output voltages of the two secondary do match close enough). No need to couple two E's in my opinion.
 
spinningmagnets said:
Could you please show a picture of two MOTs wired this way? I want to make certain that I am doing it correctly on my first try.

Mines are not rewired yet. If you want them in parallel, just wire bot primaries in parallele and both secondary in parallel.
On thing that is important : Make sure that both primaries turns are in the same direction. Also make sure that both secondaries turns are in the same direction.
This is important, because you need both altenative currents to be in phase with each other and add up : both postitive sinewaves peaks add up, both negative sinewave drop add up...
It they are out of phase (one coil in one direction and the other in the other direction), peaks and drops will be out of phase and will cancel out each other.
 
Since most of the time the duty cycle (% on time) will be very small, the wattage rating can probably be exceeded by quite a bit without excess heating. I think saturation of the iron will be the limiting factor for sizing the transformer.

Also, since these run off the AC line, the output will be a bunch of pulses at 2x the power line frequency. A triac or SS relay is used to switch it on and off but the minimum pulse duration will be on the order of one half cycle (8.3 ms). Really short pulses would be hard, but I guess not needed for most applications. With my Riba welder, I'm using about 15ms for nickel battery tabs. That would be like two pulses.
 
spinningmagnets said:
Could you please show a picture of two MOTs wired this way? I want to make certain that I am doing it correctly on my first try.

The polarity (phasing) is critical. If you get it reversed, you now have the two secondaries fighting each other (nasty short). It might work better to put the two secondaries in series. This way you can use a single piece of wire and don't need massively heavy connections. You also won't have the potential for a shorted secondary if the primary phasing is backward.

At some point, you'll start blowing the mains breaker.

I've seen some large solid state relay units that could be used to switch the primary. The good ones have built-in zero crossing switches that wait until the power line is at the zero crossing point before turning on. This reduces HV transients caused by sudden turn-on. These typically use an optocoupler on the input so the timing circuit can just drive this directly.
Here's an example (not sure if it has the zero crossing detector): http://www.ebay.com/itm/SSR-40DA-3-...366167?hash=item58bdcbb597:g:lP4AAOSw-W5Uyu-d
 
I have been planning on building a spot welder for a while and as I have a few drill batteries to repair I went ahead and bought a MOT to start my build. I was going to try and get hold of an old microwave bybut having worked on radar and knowing I didnt have a good safe place away from the kids to strip down the microwave I decided to just buy a MOT from ebay and found a cracker.

http://www.ebay.co.uk/itm/162265141842?_trksid=p2057872.m2749.l2648&ssPageName=STRK:MEBIDX:IT

its from a 1900w commercial microwave oven and is pretty big, big enough that I have concerns about completely removing the shunts as the input current could be big enough to blow a 13A fuse even if it is only flowing for a fraction of a second.
I will probably try removing half the shunt laminations and will use an unfused 16A comando plug and fit a comando socket in my workshop (not finished building the workshop yet but I do at least have the roof on now).

for the secondary I have some welding cable but also have some copper lightning conductor strip which must be about 20 or 25mm by 4 or 5mm with a thin flexible insulation so 100mm2. I will have to fold the strip over at 45deg at the start to get the 2nd turn on but I think if I fold it roung a 10mm tound rod it might go without damaging the insulation.

for the timing I have ordered this http://www.ebay.co.uk/itm/322032196773?_trksid=p2057872.m2749.l2649&ssPageName=STRK:MEBIDX:IT , it can produce a pulce down to 0.1s which im hopeing will be ok

I have also ordered 4 6 ohm 50w resistors which I am going to switch in, in series with the primary going from 2 in paralel for 3 ohms to 18 ohms which from my calculations should work out at about 25% output with 18ohms to 90% with 3ohms

I might have to wait till my workshop os finished to actually build the thing so dont hold your breath
 
Thanks for that timer link. It is from the UK, but...if it works well, we can find something similar in each of our own countries, once we see what does a good job.
 
Back
Top