Spot welder from scrap, part 1

One of the things I enjoy is taking what otherwise would be called trash and turning it into useful device. The best kind of “trash” is, of course, old / broken appliances which you can get hold of by advertising that you’re a collector of such things among your colleagues and friends. It is also not uncommon to stumble upon them while walking in the city if you train your eye for such hardware fishing. Don’t get me wrong, I do have resources to buy another thing that I need, but it is strangely satisfying to start designing your manufacturing process around the items that just are there around you without relying too much on global supply chains. There is a little bit of Fallout kind of romance in it, but brought to the real life.

My holy Pile of scrap was, surely, the most criticised asset of the workshop, but it also was the starting point of quite a few nice projects, including this one. This time the Pile presented me with a couple of 1000VA UPS units. Those are heavy beasts and their weight is split approximately in half between the two bulkiest components — the lead battery and the transformer. The transformer looks more or less like this:

It consists of a stack of steel core elements and two windings, typically of aluminum wire.

There are several useful DIY applications for high power transformers in a hobby workshop. I was interested in creating a spot welding machine this time. Such a machine requires a transformer with output voltage of about 2 Volts and short-circuit current in a range of a few kA (kilo-Amperes). Obviously, the parameters of the low-voltage winding of the UPS transformer are way off from those figures, but the overall power capacity is somewhere in the lower end of the range and the high voltage winding may be used to directly connect it to 220V. However, as it is, the magnetic core is barely good for 1V per turn and this means that the output winding should consist of 2 turns to achieve the 2V goal. And it must be a thick winding, I must stress, as it passes several kA. It is practically cumbersome to wind 2 turns of wires of those cross-section on the bobbin and would be much easier to go with only 1 turn, which would not require inter-turn isolation. I did try to go with the 2 turn proof-of-concept and it did work, kind of, but there was nothing to be satisfied with in that design:

It is possible to get the needed 2V with 1 turn of winding if you double the cross-section of the magnetic core. Fortunately, it is not that hard to do: two identical cores might be bolted together, doubling the thickness of the core stack. After that the primary (high voltage) winding is wound around them, interwinding isolation is set in place, and finally, the secondary 1 thick turn is made. Here is what you get:

Even the original plastic bobbins were reused, only one sticking part had to be cropped from each one in order for them to fit flush with each other. The wire of the primary winding was reused as well. The non-isolated wire for the secondary winding was taken from a piece of stranded aluminum cable — the one you can pick up on the street when they are fixing power lines:

Both those wires needed to be straightened first and for that I’ve assembled a quick and primitive contraption:

These two pieces are brought together in a vice with a wire passing between them. Then the wire is pulled through for a couple of times and voila, you have a straight wire.

It would be almost impossible to densely pack lots of individual wire pieces into the available gap, so another trick was used: a multi-turn winding was first tightly wound around the bobbins and a temporary wooden rest, sticking outside from the bobbins, then the core stack was assembled, and only after that, all the turns were cut on the line of their “aphelion”.

Now the sticking ends of the aluminum wire must be crimped into cable terminals. This is usually done with a hand-held hydraulic press like this one:

But what I had at hand was a stationary (shop-built, of course) 20T hydraulic press and not a small amount of enthusiasm. This enthusiasm + milling machine + couple of metal scrap pieces have translated into the properly sized crimping matrix:

And with the press and matrix at hand it was only a matter of cranking the lever to get the terminals installed:

In order to bring the terminals together and pass the current to the electrodes I designed a pair of copper clamps. With a little bit of machining, layout and more machining, here is what I’ve got:

At this point my only wish was to spend several hours refining the chamfers and polishing the planes, but I was already way too far in the “Yak shaving” land.

As the dear reader might see, this post is already quite lengthy, while the story has barely begun. And if you have liked it so far, I invite you to continue the reading to the spot welding from scrap part 2.