This welder, thanks to the limit switch, after lowering the upper electrode arm, automatically enables the welding process in a safe and proper way. First, it waits 1 second so that the user have time to clamp electrodes on welded material, and then turns on the weld current for time in the range of 0 – 4 seconds, which is set by the potentiometer. This allows both hands to be free, and there’s no need to use the foot switch.

Transformer:
Transformer comes from a Technics SU-V670 amplifier. Core cross-sectional area is 23.5 cm ². It seems small, but it has a considerable power for such size. Even before rewinding, short-circuit secondary winding, caused to knock out the B10 fuse. Now, the secondary winding measures almost 2 meters of LGY 35mm ² link (10mm thick with insulation). Window dimensions 46.5 mm x 9.5 mm allowed to fit four turns of wire. At 660mV at turn, 4 turns gave 2.6 V, with 950-1000A of short circuit current. This transformer in compare to a microwave oven transformer, has the advantage that it can be permanently supplied from mains, hence there was no need for installing a cooling fan, because all of it can warm up, is the heat from the electrodes and the secondary cable. Numerous holes in the casing provide passive cooling.

Construction:
The transformer and electronics circuit casing is made of the ATX power supply casing components, welded together. You could say that the welder weld itself together :) Of course at the beginning was no mounting arms or timer, so welding was done by hand and without the time control, but it came out pretty well. At the ends of the secondary cables, a 35mm ² copper ring connectors are soldered. The arms are aluminum 2cm x 2cm angle profiles, they play a very important role in the dissipation of heat from the electrodes – this allows you to perform more welds at a given time. For the electrodes, i have used DY 10mm ² copper wires (approximately 3.5 mm in diameter). They are squeezed between the arms and connectors with 10mm bolt, but run through a 3.5 mm hole made in this bolts. This direct connection provides a very good electrical contact and heat dissipation. Also, without any problem, the electrodes can be attached at any angle, at any length to get into various corners of welded components. Method of attachment prevents the electrodes shift even under strong pressure. Arms working length is 24cm.

Timer:
Limit switch is activated by a finger attached to the upper arm of the electrode, which starts the timer. Timer is a simple circuit based on 4 NAND gates contained in the structure of 4093 logic chip – LINK. The first timer is set at constant 1 second. After this time, unlocks the second timer, which has a potentiometer for adjusting the time of welding, in the RC circuit. Output of the last gate enables two NPN transistors (its output is only 0.5 mA) – why not one PNP? Because as soon as the button is pressed, the output of the last gate has a momentary voltage drop to 8V, and captured by the transistor, has been passed on further. Such phenomenon does not exist when NPN is used, and due to the fact that it inverts the signal, i had to apply one more NPN to invert it again to its original state. The rest also wasn’t bug free, the timer worked flawlessly in the CircuitMaker simulation, but in reality there have been problems and I had to make a small changes in the circuit, which in turn, did not work in the simulation. I insisted to make it without any uC and i did, also, uC can be very unstable in such harsh conditions (1kA inductance). CircuitMaker simulation file and a working circuit schematic are in the attachment. Further, the latest NPN transistor switches zero-cross detection optotriac MOC3041, and the triac BT138 which is closing the primary circuit. Timer is continuously fed from a small 10V transformer placed in the corner of the casing, with 14V after rectification. I thought that BT138 (12A) is too weak for this job, but i made a lot of testing welds and its still alive. But please use stronger one if you have.

Safety:
Please wear at least a safety glasses when working with such devices. Even if the current is switched on and off when the electrodes are pressed to the material, there is always a chance that a piece of molten metal will spatter in some direction, and it is only a matter of time when it will choose our eye as his direction. Another thing – in this case, the casing is made of a metal and has been grounded along with the transformer and electrode arms – so if any mains inside touch to the casing, home installation protection will go on and user wont be electrocuted. Triac with such power requires a small heatsink, and if the case is metal, you should carefully consider its mounting. In this case i have applied a double insulation. First, triac isolated from the heatsink with a silicon pad, then the heatsink is attached to the casing with a plastic spacers. Please remember to use only a grounded outlet for your own safety, as you never know when something bad can happen.

Results:
Welder easily handles with sheets 2 x 0.75 mm – weld can’t be ripped apart, instead, sheet is ripped right next to the weld. It can’t do 2 x 1mm, weld is braking when ripped, but it can be used for welding some light duty parts. Perhaps if the electrodes were thicker and pointed at the ends, it could weld thicker sheets. Current electrodes are deformed with a strong pressure and time above 3 seconds, and they heat up red with 2 x 1mm sheets with 4 seconds. If i get some thicker electrodes, about 5mm, i’ll enlarge the mountings hole, and these will be used. I didn’t installed any handgrip so far because it is unnecessary at the moment, and it will make impossible to reach some tight corners of the welded material. Aluminum arms quite well conduct heat away from the electrodes, so welding can be done without a lot of worries about overheating or melted wire insulation, and they get cold pretty fast so breaks in work are not too long – the transformer doesn’t heat itself.

Gallery:

Files:
DOWNLOAD – Timer schematic in eagle and pdf; CircuitMaker 6.2c simulation file