Tested my TIG inverter bridge under full power

Congratulations! Was that 199 A into a short or into an arc?

Thanks... It was into a short. To have a welding arc, I would need to put the bridge to the proper place on the commutator.

Congrat's. I'm wondering if you came across any schematics of commercial tig inverters during your hunt? Or are they drawn as black boxes? Thanks, Jim.

This isn't much of a schematic. All the circuit boards are just rectangles. Also, this is a DC only welder.

Jon

Certainly looks like an AC/DC welder. One good hint is the AC/DC selection switch in the lower-right corner (after rotating the page so it is readable).

The right-hand side shows two DC supplies, one + and one -, that are selectively connected to the "+" output via the transistors T21 or T22. They could be controled to give DC+, DC-, or AC.

It would be interesting to know what parts they are using for the IGBTs and the T21, T22 transistors.

I agree, except they appear to be bipolars, not FETs.

I think you are wrong. The tapped secondary is connected into two full-wave rectifiers for the two supplies (+ and -). Both halves of the secondary are always working, but only one supply (+ or -) will be connected at a time and supplying current.

Thanks. I may try to weld aluminum this weekend.

I have not found anything.

The only thing that my inverter does not have, is control of power when electrode is negative, and separate control when electrode is positive (I think that it is called AC balance). For my inverter, the absolute value of the voltageand current is the same for positive vs negative half cycle.

Basically, I think, commercial inverters are a lot like mine, they have a timing circuit, drive circuit, full bridge driven by FETs (MOSFET or IGBT), a snubber (maybe) and controls.

I am just a chump who put together my first electronic contraption, so I may not know as much as more learned members of these newsgroups.

Take my answer for what it is worth.

I have just scanned the schematic for a Hitachi inverter TIG and added it to the dropbox. The jpg is about 1mb.

http://metalwork>>

What is wrong behind my idea to switch the bridge asymetric? For example

Nick

It is a AC/DC welder. Look at FETs T21 and T22, you can see that they can be used to supply either +, or -, or +/- to the line labeled

in AC mode, I think, only one half of the secondary winding is working. I may be wrong, I slept only 3-4 hours for the last 2 days and am barely thinking, .

You can also see RCD snubbers near T21 and T22.

Could be. I thought that bipolar transistors are also field effect transistors, that's why I said so. I am an amateur.

Exactly what I was trying to say, not so well.

The schematic very clearly shows a bridge rectifier at the output, and the output terminals are marked "+" and "-".

Jon

OH, heck, you are absolutely right! I was viewing the thing in Mozilla, which just gives me a peephole view of the file. Yes, T21 and T22 are between the bridge rectifier and the output terminals. (Dumb labeling of those, they should be electrode and work, not + and -.)

No, only one half of the bridge rectifier is working. Presumably the work terminal is connected to the center tap, and the bridge always produces a + and a - voltage referenced to that. The two transistors select from the two polarities. Certainly a minimum solution to the problem.

Jon

Bipolar Junction Transistors come in NPN and PNP types, and do not use the field effect. They use base current to modulate (and amplify) collector current.

Field effect transistors use the electric field between the gate and source to modulate the resistance in the channel. Gate voltage essentially controls drain current through this method. Generally, gate current is zero.

Insulated Gate Bipolar Transistors are a combination of the two. They really are a bipolar junction transistor with a FET connected between the collector and base. They give many of the advantages of both types. With the FET, the voltage drop across the drain-source increases linearly with current. (It is a resistance, after all.) With the Bipolar and IGBT, the voltage drop is almost flat with changes in current, usually around 1.5 V. Regular bipolar junction power transistors require a base current about 1/10th of the collector current. With 100 A transistors, that amounts to 10 amps the control circuit needs to deliver to the base. Darlington transistors have at least 10 x higher gain, but the voltage drop will be double, a major disadvantage. The IGBT solves that problem. The gate current required is zero, once the gate capacitance has been charged.

One other pro/con thing about IGBTs is that, like other bipolars, the forward voltage drop goes DOWN as they get hot. This effectively prevents thermal runaway with single devices. But, it makes it a lot harder to balance current in paralleled devices, as the hottest one will have the lowest voltage drop, and thus get all the current. (By comparison, the resistance of FETs goes UP with temperature, making thermal runaway possible. But, it makes paralleling devices a breeze, as they will self-equalize their current.)

Jon

I can control the period of TIME when voltage is positive, or negative, as percentage of the cycle.

I cannot make the CURRENT or VOLTAGE on positive or negative different.

...I liked the site I found where a German engineer had modified a washing machine into a beer brewing machine....

Steve

That's what I meant.

That's clear. But if you average the current flow (better to say polarity) and do not have a 50:50 cylce, you will have a DC-offset. Imagine having a 100:0 cyle. It would give you DC with one polarity (of course).

You do have a H-bridge, I guess.

I don't know wether sophisticated AC-TIGs do it this way, but I would try it this way. ... if I could not get my hands onto a TIG and connect a scope to it for an easy verification.*) Before you ask: No, I don't have one. Good AC-TIGs are to expensive for me. So I don't have a TIG at all. But if your modifcation works, I'll have to rethink that.

Nick