Soldering/Welding Techniques

Thread Starter


Joined Jan 17, 2007



Front side

Back Side

Machine settings:
  • 8% power
  • 21ms time
  • 1s gas flow @ 0.15 Mpa
  • Material: stainless steel 22ga (0.029").

A few things worth noting:
  • The machine comes with a small motor and a diamond coated wheel for electrode sharpening. The wheel works fine, but to say that the motor sucks would be praising its greatest quality and a gross understatement. It works using an USB connector to a 5V source at the back of the machine. This source is fed by the small transformer that I had to reattach using zip ties. The thing is not only puny power-wise, but it's also of terrible quality. So bad that its shaft is made of plastic! not even metal. It sucks, sucks, sucks... So what I did was attach the wheel to my Dremel tool arbor shaft and it worked beautify when I sharpened the electrode.
  • The manual says that the electrode should be protruding 10 to 15 mm from the nozzle. But 15mm looked like way too far from the nozzle to me, so I adjusted to the required 10mm minimum. Hell, maybe next time I'll adjust it down to 5mm and see what happens. I'm under the impression that the closer the electrode is to the nozzle, the better, since a more uniform gas flow should hit the work piece.
  • The maximum power that can be adjusted at the controller is 10% with a duration of 30ms, so the test I did was close to the maximum attainable by this toy, I think. But no matter, the fact is that the type and size of welds I'll be working on are of the same size as the one I performed on the test, or smaller. So I'm not worried in that regard.
Image00004.jpg Image00005.jpg



Joined Jan 18, 2008
I agree. For TIG, I would consider 15 mm way too far for the electrode to extend past the nozzle. I usually use something closer to 5 mm or even less for 1/16" electrodes. You want a laminar, non-turbulent flow of shield gas. In really still air, you may get away with more stick out. 20 cubic feet per hour is my usual gas flow rate.
Last edited:
If those FETS were transistors, you generally would use emitter resistors to help with current sharing. It's so a few transistors don't hog all of the current. The parallel combination is to cut costs. To increase the cost, you could have a big heatsink and those expensive fined resistors with mounting holes.

I used something like these to mimic a light bulb when I replaced car lamps with LEDS. It screwed up the dimming and retained power circuit.

You open trunk. The lights automatically turn off because it was open too long. Without the mod (trunk light only), you would have to disturb something like open/close a door and the trunk light and interior lights, I think would be half-bright until you closed all of the doors and the trunk and opened the trunk again. The resistor sits under the rear deck near the trunk lamp. I didn't secure the resistor. It should never be on more than 15 minutes anyway.

They are expensive.

Next you can put the capacitors together with buss bars,

that was basically how the spot welder I used was put together. A very old version of one of these It may have had vacuum tubes in it.

Something like: UNITEK-125-Watt-Resistance-Micro-Spot-Welder-Power-Supply-With-80f-Weld-Head

Thread Starter


Joined Jan 17, 2007
Bingo! ... I've been experimenting with my new toy, and arrived at the conclusion that a setting of 60% @ 15 ms with a gas flow of 1 second is adequate for what I want. While 70% @ 21 ms and 1 second gas flow makes for a very strong weld (see attached pic). So much in fact, that it melts the material a bit more than I find acceptable. But whatever, this thing is working as expected and has completely solved my welding needs. At least for this application.


And BTW, I thought the machine could be adjusted from 1% to 10% only, but I was wrong. It can be adjusted from 10% to 100%, this because the two digit LED display is too primitive to show three digits. So 100% power is shown as a 10 in the display. This means there are 9 different levels of operation. It's too bad there are not 8, since it would've told me something about the presence of 8 nFets in its circuitry. So instead of using one additional nFet at a time per power setting, maybe a more complex PWM approach is being used... and come to think of it, that's probably the way I would've gone too. Using an incremental number of transistors per power setting would have the drawback of stressing the first transistors too much, leaving the upper ones unused most of the time. Using PWM would make all of the transistors carry the load simultaneously, regardless of the power setting.