Saturday, 21 November 2026

Well - does the TIG button circuit work or not? Almost but not quite....

Now that I've been shown the magic spell required to make the TIG200PACDC work with a remote amperage control, there's nothing stopping me from trying out the TIG button circuit.

Does it work, Fatty?

Hmm, not quite. 

The first issue is that the +5V and GND pins have been reversed. That is my dumb fault for letting The Stupid fat Bloke figure out the connections to the torch connector. So out with the scalpel and some fine wire. 

Of course, I never established what the available supply current was and it seems that my nifty new circuit exceeds what's available. When I took the lid off the welder, I could see that the interface circuit had its own floating PSU rail but I must admit I didn't see any series resistor for limiting the current. I'm guessing there was one but I missed it. I was rather loathe to spend ages on this while there was uncertainty over whether the remote amperage control even worked in the first place. Or that's my excuse at any rate.

The LED lights up (very) briefly, then dies out. The relay within the welder only makes a half hearted click. That's not right. It's obviously taking too much current when the SSR is triggered.

What dominates the power supply draw?

  • Primarily the SSR, which incorporates an LED on one side of the barrier and a photo transistor (in fact back to back MOSFETs) on the other. It probably wants about 7.5mA, depending on the Vf of the LED (about 1.15V).
  • The green LED that shows when the arc is to be powered up. This has no actual function and won't even be visible when the PCBA is installed in the handle. This would be about 4.6mA at 2.85V.
  • There are a few other elements that then dominate the remaining current draw but these are an order of magnitude or more less than the LED currents:
    • The opamp. There's perhaps 0.5mA or so that could be saved by changing to an alternative with a very low quiescent current.
    • The voltage reference (actually ratiometric, rather than a fixed voltage source) - increasing R1 and R20 could save the best part of a further 0.9mA or so.
So it would have been happily powered up until the registered setpoint triggered the SSR, at which point it would have crowbarred the power supply rail. Nice work!

But for now I'm not planning to respin the board, or at least not immediately, so how to knife and fork it so that it actually works?

The LED and SSR are visible in this section of the schematic:


And here's the mod to eliminate the LED currents:


Note that the "EN_COM" is actually a +ve voltage and the "EN_NO" (normally open) would arguably be called the ground by convention. I didn't know, or bother to check. So my bodge requires the "EN_NO" PCB finger to be linked to the GND net and the "EN_GND" to be linked to the top of the MOSFET.

The FET that currently switches the SSR LED can switch the welder's "arc enable" circuit directly. There isn't any galvanic connection between the 2 circuits, so it hopefully won't complain if I bridge them here. I measure just over 1V on the contacts and ~4mA when shorted out with the DVM. From memory, the switch is shorting out a transistor that otherwise drives the LED or an opto.

In the flesh, this is what needs to be done:


And for the R1/R20 mod:


Red = R1 (12k); Blue = R20 (22k). I only have 1206 SMT resistors, so they would swamp the original pads. This might work:


This should then reduce the total supply current:
~0.7mA from ~1.7mA (at idle)
~0.8mA from ~14mA (enabled)

Given that the supply only manages to source ~2.5mA when shorted out, it's not hard to understand why it didn't actually work when connected to the welder.

Measuring the voltage on the 10k pot, I see 4.15V. Assuming it's a 5V supply, that suggests a series resistor of around 2k. I need the output voltage to reach at least ~4V with the quiescent current happening, so I really need to get the current draw down to ~0.5mA or less, primarily by changing the opamp. I'll set Claude and Adom to work on that, along with the above bodges. 

This will be a v2 layout but I'll test out the other changes before thinking about ordering another set of PCBAs.

Time for some delicate soldering, once I locate suitable resistors.....

No comments:

Post a Comment

Well - does the TIG button circuit work or not? Almost but not quite....

Now that I've been shown the magic spell required to make the TIG200PACDC work with a remote amperage control, there's nothing stopp...