Strange SCR power supply (welder) issue

What is the maximum current?

about 200A.

i

Leakage.

Cold Fusion

A blessing from Our Lady of Perpetual Motion

Gunner

"If thy pride is sorely vexed when others disparage your offering, be as lamb's wool is to cold rain and the Gore-tex of Odin's raiment is to gullshit in the gale, for thy angst shall vex them not at all. Yea, they shall scorn thee all the more. Rejoice in sharing what you have to share without expectation of adoration, knowing that sharing your treasure does not diminish your treasure but enriches it."

- Onni 1:33

That's close to my list of hypotheses.

Forgot to say, I use a LEM current transducer for current feedback. It has + and -15V inputs. I thought for a moment that perhaps if the + and - inputs are not precise opposites, it would exhibit the observed behavior (since there is not ground input, the zero is the middle point between + and - input), but I know that inputs are very close, like 15.03v and 15.01v. (according to my multimeter)

i

My first thought is you have an offset on your ammeter, that offset being - tada! - about 12 amps.

GWE

That's a good one, but, that's not the case. There is real current. The needle is at zero if I have leads disconnected. When I touch them, there are tiny sparks, and then current starts flowing.

i

What is the problem here, aside no troublshooting skills?

you have 12 amps coming from somehwere, find out where. What's between the power source and where you're reading 12 amps?

electricity flows in a circuit. You can can figure out where it's flowing.

1 gauge wire, with work lead shorted to electrode holder. i

It's not going to 0 volts. It doesn't take much voltage to get a current flow into that low impedance. Recheck with a good DVM. I would guess you have an offset voltage problem in your control circuit.

The board or the scr firing ckts may need to be balanced. IF you have current it is coming through the scrs. The other case may be you have a bad scr. You could put a scope on the primary side of the firing ckt but make sure you don't get on the scr side of the firing transformer unless you have one of the battery operated scopes designed for 600 volt insulation or better. Even that sometimes doesn't help though, my 221 Tek scope blew up the test lead when an internal metal shield in the scope broke free andwent across the scope lead input line inside the scope. It was scary. It took out the 30 kw drive too.

John

Good, now unless that wire is picking up some serious magnetic fields it won't be generating any current.

next step. check your power source. Maybe the current is coming from there.

Circuit = complete loop. check all the parts that complete it. Electricity doesn't come from nowhere.

Do you mean something in control circuit not referenced to the same ground?

i

Good thinking. I can try the following, supply AC power to SCRs, with the SCR firing board not even turned on (ie disconnected from 110v).

If that 12A currrent persists, I will know that it is an SCR that is at fault.

i

Yea, that's where it is coming from, the power source!

I am sure that it is coming from the SCR rectifier, there is nowhere else it could come from, the question is why.

So far I see the following possibilities:

1) SCR firing board is getting wrong control signal inputs (ie, wrong desired voltage signal). Maybe due to incorrect grounding.

2) SCR firing board gets bad feedback signals from the current sensor

3) SCRs are bad and "leak electricity"

Case 1 can be tested by forcibly connecting the "desired current" input of the SCR firing board to the ground (as well as the negative end of where the potentiometer is to be wired to the SCR firing board). If current disappears, I know that I have bad input.

Case 2 can be tested by supplying 4 volts to the current sensor input of the SCR firing board (that means that the board is led to believe that the current is too high, as my max current setting is much lower). So, if it is any good, it should go as low as possible.

Case 3 is easy to test for -- there should be no current when 3 phase AC is applied to the input of SCR rectifier and SCR firing board is turned off (ie it is not supplied with 110V power to run at all). If current persists, we know that I have one or more bad SCRs.

Any more thoughts?

i

Most SCRs have some sort of snubber connected across them to limit dV/dT turn-on. Usually, they draw only 100 mA or less. If you have an output transformer with a high ratio, it would cause higher output current, but your 12 A would mean a ratio of 100:1 or so. It would be helpful to look at the output waveform to see if there may be any high frequency spikes, which could be caused by transients in the supply line. They would be conducted by the snubbers and cause possibly high currents.

Paul

You haven't mentioned that you have measured the SCRs Gate terminal voltages, and what part of the circuit you're using for the zero reference. As you most likely know, the SCRs shouldn't be conducting when there is no voltage potential on the Gate terminals.

I would suspect that the Gate terminal signals are referenced incorrectly, and not being reduced to zero volts. If your DVM isn't indicating that signals are present at the Gate terminals, you would need to investigate a safe method of utilizing an oscilloscope.

Many electrical devices utilize several separate, different circuit ground points. Digital circuits ground is separate from power devices ground, and the separate grounds will be isolated from each other, and often isolated from earth ground.

I would anticipate a sophisticated welder might possibly have 4 or more grounds.. digital/timing, analog/control, power/output, and earth/safety ground.

You mention that the Voltage pot when set to zero, yields zero volts output. I don't think you were implying that you're measuring 12 amps at zero volts, but your explanation wasn't clear about that.

WB ................

Ignoramus17359 wrote:

If the voltage pot is set above zero and the problem is that you are not getting zero when the current pot is now set to zero, then I would suggest an offset in the current sensor output. If it has an offset control then adjust it. If not, then another op amp, twiddle pot and couple of resistors are called for to give you that adjustable offset. All this assuming that you aren't getting current when you disconnect the SCR gate leads :-)

Mark Rand RTFM

Not the case. Grounding desired current input did not change anything.

did not have time to test

also not the case

I am now tending to think that perhaps this is simply a deficiency of my SCR firing system, and that perhaps I should just "learn to live with it".

i