Dear Diary,
This is not really about RE but it may still interest the odd person on this board. Besides, it could be used for spotwelding batterypacks and those packs have use in RE...
Yesterday I finally finished the capacitor-discharge spotwelder of which construction started in early January (but the plans had been brewing in my head for nearly a decade). It is meant to be used for welding batterypacks. The nickel strips that can be seen in the packs of cordless drills, for example.
By a stroke of luck I managed to recently purchase capacitors for the bank at a very low price (with enough capacitors left to build another welding machine). About a week later I found the required large thyristor. And a few months earlier I had finally built a few powersupplies (Oztules' PC PSU modification) that could power the machine. In short, everything came together at the right time so there was no excuse to not build one anymore after years of procrastinating.
The principle is simple. A modified PC PSU with current limit at 10 A puts out a user-adjustable voltage of 2.5-24 V that slowly (in this case, just 3-4 s) charges the capacitor bank. The bank consists of 23 capacitors of 68.000 uF / 25 V, so in total 1.56 F @ 25V. The bank can thus store E = 1/2 * C * V^2 = 1/2 * 1.56 * 25^2 = 488 J (Ws) of energy. Since I reduced voltage of the supply to 24 V (so as to be sure to not exceed the capacitor voltage rating) the actual maximum energy that will be stored is 450 J.
When the bank is charged the user puts the electrodes on the workpiece and presses the footswitch to discharge the capacitors in the workpiece. The footswitch triggers the thyristor (via the control board). In less than a millisecond the energy stored in the capacitors is released into the weld zone. This produces one (or two, depending on how one welds) small spotwelds.
That's basically all there is to it. By adjusting charge voltage one can adjust the weld energy, thus preventing vaporizing small workpieces (yes, that happened a few times too...). When the welds are performed properly (clean electrodes, enough electrode pressure) there are only very few sparks (still, one should wear safety glasses). The welding proceeds so fast that the workpiece has no time to heat up; one can touch the weld zone immediately after the weld is made.
In the picture below can be seen some test welds. To my surprize the apparatus is able to not only weld the 0.005" nickel tabs for which it was designed, but up to 3 mm total thickness stainless steel as well. Very impressive.
All in all, a useful addition to the shop. And after roughly hundred testwelds, the machine seems to perform satisfactorily without a sign of problems.
Two minor details that anyone who considers building one should keep in mind: the relay to disconnect the PSU during welding failed pretty quickly; it got welded stuck in its resting position (NC) after just a few welds. The relay I replaced it with did too. Obviously, when the capacitor bank is discharged and the relay falls back to its NC position, the large inrush current (now supplied by the capacitors in the PC PSU) weld the contacts together. Lesson learned. The relay has been bypassed by a wire. This means one has to remove the probes from the workpiece to stop the thyristor from conducting, but it's no big deal. No arcing occurs, as I had feared (and what was the reason I added the relay)
Another possible issue is that I've used aluminium bar to make the capacitor bank, not copper. Aluminium has only slightly higher resistance than copper, but my main worry is that it might oxidize over time. We'll see how it keeps up, if it turns out to be a problem I'll replace the bars with copper bus bar.
Below are some more pictures of the machine in its various stages of completion, along with pictures of a few test welds.
Aluminium cooling fin contains the thyristor, the green board on top is the homebuilt controller. The board on the left contains a 24V @ 2A power supply.
Under construction. The modified PC PSU can be clearly seen here in the background.
Some testwelds.
More testwelds.
More high-resolution pictures can be found here, for anyone interested:
http://www.anotherpower.com/gallery/album92?page=1
One last interesting detail: apart from the surplus capacitors (1 E/pc) it was built entirely out of junk, i.e. parts/components/enclosures that were thrown away by others. Similar commercially manufactured machines (with only 200 Ws of weld energy) cost about 3000 US$.
All in all, a simple, fun and useful project.
Peter.
