Welding -- what to do about HUGE 2" arcs

[Ignoramus24136]

Tried to weld with my "modded" CyberTig. The power supply controller
was set at 130 or so amps, 90V OCV.

I used 1/8" 6013 stick.

It was kind of nice and easy to lay beads, all was great, but one
thing freaked me out.

If I start an arc and start pulling the welding rod away from work,
the arc continues until it gets about 2" long. I was quite shocking to
me to see just how big the arc was.

Finally I pulled the arc so far enough that the controller failed, I
suppose due to phase imbalance due to overloaded third phase (from a
phase converter).

It is not a deal breaker type of problem. I can still weld just
fine. I just need to remember to pull the electrode away quickly.

I think, though, that it is weird enough that something needs to be
done about it.

So... Any ideas what kind of simple solution I can have to interrupt
welding if the arc is too long.

I am thinking of some logic thingy that would detect a condition
such as:

- Current above 3 Amp
- Voltage above 60 volts

When that condition is met (should be possible to do with some
comparators or whatever), a NC relay would be opened and the
controller shut down for a moment.

Does it make sense? Anything simple that I could buy rather than build?

I am kind of thinking that perhaps the right way would be to plop some
money to get a programmable logic controller, but I have no experience
with them. I am a computer programmer though, perhaps it could help.

i

 

[Ignoramus24136]

Finally I pulled the arc so far enough that the controller failed, I
suppose due to phase imbalance due to overloaded third phase (from a
phase converter).

When I say that it failed, it did not fail as in "died", it shut
itself down with a FAULT LED indicator. Which is a good thing. It has
a built in detector of phase imbalance, which could happen if power
drawn exceeds the capacity of my phase converter.

i

 

[Tim Williams]

I am thinking of some logic thingy that would detect a condition
such as:

- Current above 3 Amp
- Voltage above 60 volts

When that condition is met (should be possible to do with some
comparators or whatever), a NC relay would be opened and the
controller shut down for a moment.

Does it make sense? Anything simple that I could buy rather than build?

A couple HV diodes, resistors and capacitors to make a low pass filter (you
only want the welding buzz (or DC), not the hash the arc makes) and envelope
detector. Toss on a voltage divider (to put it in terms a comp can handle),
add a comparator with pot to vary trigger voltage and send the now detected
signal to your shutdown thingie.

Oh, and some clamping after a resistor, so you don't blow caps or other
stuff. Voltage spikes are fun.

I'm guessing this doesn't have to be very fast at all, as long as 100ms
could be considered acceptable? Should be pretty easy then.

But really, you're COMPLAINING about 2" arcs? I've got to see this.

Tim

 

[Ignoramus24136]

A couple HV diodes, resistors and capacitors to make a low pass filter (you
only want the welding buzz (or DC), not the hash the arc makes) and envelope
detector. Toss on a voltage divider (to put it in terms a comp can handle),
add a comparator with pot to vary trigger voltage and send the now detected
signal to your shutdown thingie.

I think that I understand about LP filter, but what do you mean by
envelope detector.

Oh, and some clamping after a resistor, so you don't blow caps or other
stuff. Voltage spikes are fun.

Absolutely. I could as well throw a zener in there.

I'm guessing this doesn't have to be very fast at all, as long as 100ms
could be considered acceptable? Should be pretty easy then.

Sure. Maybe 50 ms, but there is no need to be fester than that.

I played some more with welding tonight. If I withdraw the electrode
quickly, there is no problem at all. It is when I pull it away slowly,
then I see the unbelievable 2" arcs.

But really, you're COMPLAINING about 2" arcs? I've got to see this.

It is unreal, I have never seen anything like that. Not that I can
claim great life experience.

i

 

[Gunner]

When I say that it failed, it did not fail as in "died", it shut
itself down with a FAULT LED indicator. Which is a good thing. It has
a built in detector of phase imbalance, which could happen if power
drawn exceeds the capacity of my phase converter.

i

How about a foot pedal? Like standard tig. Btw..if I set my control
on Local rather than remote..when tigging I can get the HV to run at
least 1.5-2" when I draw away from the work. I had to turn down my HV
a fair amount when doing a weldment that I coulnt use the foot pedal
on (no torch mounted controller)

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

 

[Ignoramus24136]

How about a foot pedal? Like standard tig. Btw..if I set my control
on Local rather than remote..when tigging I can get the HV to run at
least 1.5-2" when I draw away from the work. I had to turn down my HV
a fair amount when doing a weldment that I coulnt use the foot pedal
on (no torch mounted controller)

I think that you have a good point, that with a remote control it will
not be as much of an issue (not that it is a show stopper now). I will
definitely do remote control soon.

i

 

[Tim Williams]

I think that I understand about LP filter, but what do you mean by
envelope detector.

Peak detector- think AM demodulation (hence envelope detector). Just a
diode into a capacitor, with another resistor and capacitor to add the RC
delay, and an RC in front of the diode to slow transients.

Hmm, you have HF start right? A notch filter might not hurt, even with the
slow front end of this circuit.

Tim

 

[Ignoramus18798]

Peak detector- think AM demodulation (hence envelope detector). Just a
diode into a capacitor, with another resistor and capacitor to add the RC
delay, and an RC in front of the diode to slow transients.

Oh, I see. Thanks. I had some very educated opinions that what I have
(big arcs) is not a bad thing as it could let me weld over dirty spots
and such -- and that all that I should do about it is pull electrode
away quickly when I am done with a bead.

Hmm, you have HF start right? A notch filter might not hurt, even with the
slow front end of this circuit.

Yes, I have HF start indeed.

I have not yet wired it up.

i

 

[[email protected]]

2" arc at how many amps? When I borrow my friends tig, we usually use
HF start and it arcs at about 3 CM out at a maybe 100 mA, then as you
get close to the work the arc thickens into the normal welding arc.
Maybe you have it close to being right?

Steve Roberts

 

[Ignoramus18798]

2" arc at how many amps? When I borrow my friends tig, we usually use
HF start and it arcs at about 3 CM out at a maybe 100 mA, then as you
get close to the work the arc thickens into the normal welding arc.
Maybe you have it close to being right?

Not sure how many amps, but if I recall correctly (I did look at
dials, but in haste and mostly at the voltage dial), it was at about
100 amps.

Maybe, indeed, it is close to being right.

i

 

[Rene Tschaggelar]

Tried to weld with my "modded" CyberTig. The power supply controller
was set at 130 or so amps, 90V OCV.

I used 1/8" 6013 stick.

It was kind of nice and easy to lay beads, all was great, but one
thing freaked me out.

If I start an arc and start pulling the welding rod away from work,
the arc continues until it gets about 2" long. I was quite shocking to
me to see just how big the arc was.

It is a DC welder ? Yes, DC makes arcs this long.

Rene

 

[Ignoramus18798]

It is a DC welder ? Yes, DC makes arcs this long.

Yes, it is a DC welder.

i

 

[colin]

Tried to weld with my "modded" CyberTig. The power supply controller
was set at 130 or so amps, 90V OCV.

I used 1/8" 6013 stick.

It was kind of nice and easy to lay beads, all was great, but one
thing freaked me out.

If I start an arc and start pulling the welding rod away from work,
the arc continues until it gets about 2" long. I was quite shocking to
me to see just how big the arc was.

Finally I pulled the arc so far enough that the controller failed, I
suppose due to phase imbalance due to overloaded third phase (from a
phase converter).

It is not a deal breaker type of problem. I can still weld just
fine. I just need to remember to pull the electrode away quickly.

I think, though, that it is weird enough that something needs to be
done about it.

So... Any ideas what kind of simple solution I can have to interrupt
welding if the arc is too long.

I am thinking of some logic thingy that would detect a condition
such as:

- Current above 3 Amp
- Voltage above 60 volts

When that condition is met (should be possible to do with some
comparators or whatever), a NC relay would be opened and the
controller shut down for a moment.

Does it make sense? Anything simple that I could buy rather than build?

I am kind of thinking that perhaps the right way would be to plop some
money to get a programmable logic controller, but I have no experience
with them. I am a computer programmer though, perhaps it could help.

i

take a picture of it

Colin =^.^=

 

[Glen Walpert]

Yes, it is a DC welder.

But a 2" arc is not at all normal for a DC stick welder.

You could try implementing DROOP so that the current drops to a low
value at a reasonable voltage.

DROOP lets you taper off the heat into the weld puddle by pulling the
rod away slowly, to prevent craters and crater cracking, without
resorting to manual current control via footpedal or torch
potentiometer as in TIG.

The right amount of DROOP is what you want, not constant current or
constant power. DROOP. Adjustable DROOP.

With enough DROOP, the arc will be substantially cooler at 1/4" and go
out at around 1/2" to 3/4" at the most, way longer than the 1/8" arc
you should be welding with.

The straight poop - DROOP DROOP DROOP.

If that dosen't work, try DROOP.

Did I mention DROOP yet?

If you don't like DROOP, you can call it DIG, BOOST, or SUPER DELUXE
EXTRA SPECIAL DC STICK WELDER AUTOMAGIC OPTIMAL CURRENT PROFILING.
Just make sure current increases when voltage drops and decreases when
voltage rises by enough for a controllable arc.

 

[Ignoramus18798]

Thanks Glen. I appreciate it. What do you think about implementing a
feedback based not on current, but on power (current x voltage), it
should not be too difficult with a couple of opamps and analog
multiplier like AD633.

i

 

[Rene Tschaggelar]

Yes, it is a DC welder.

Reduce the max voltage as soon as a current flows.
That would be a limit on the max pulsewidth then.

Rene

 

[Glen Walpert]

Thanks Glen. I appreciate it. What do you think about implementing a
feedback based not on current, but on power (current x voltage), it
should not be too difficult with a couple of opamps and analog
multiplier like AD633.

I think that would be a nice experiment, but that there is probably a
reason most welding machines use linear current vs voltage control, or
piecewise linear current vs voltage. This is actually easier to do,
only a few op-amps required, but it will require a bit of care so that
the droop control does not significantly change the welding current at
the normal arc length voltage, which would probably be in the 25 to 30
volt range. You could check this with a data logger or a second
person to watch the meter while you weld.

So you want to convert welder current and voltage signals to a range
within the signal capabilities of your op-amps, and do a bit of simple
analog math as described in for example
<http://www.analog.com/library/analogDialogue/archives/philbrick/computing_amplifiers.html>
to compute the current set point for your controller.
I'm sure a search will turn up a ton of info on analog computers, but
George Philbricks "Computing Amplifiers" is a real classic.

You can implement all sorts of I/V curve modifications like this
including power control, but I still recommend starting with droop.

Sorry about the DROOP rant .

 

[Glen Walpert]

Reduce the max voltage as soon as a current flows.
That would be a limit on the max pulsewidth then.

Rene

That is an excellent idea. In this case he is using a standard
controller which wants a DC voltage limit signal, which can be reduced
when current is above some threshold, in the same analog computer
which implements droop and whatever other effects he wants to try.

 

[Ignoramus8797]

That is an excellent idea. In this case he is using a standard
controller which wants a DC voltage limit signal, which can be reduced
when current is above some threshold, in the same analog computer
which implements droop and whatever other effects he wants to try.

Glen, do you think that a microcontroller based control solution would
let me kill many birds with one stone? to do droop, tig timing, plasma
cutting timing, etc etc?

If so, any suggestions for something simple? I could spend some bucks
(not too much) if it saves me time and wiring efforts.

i

 

[Ignoramus8797]

I think that would be a nice experiment, but that there is probably a
reason most welding machines use linear current vs voltage control, or
piecewise linear current vs voltage. This is actually easier to do,
only a few op-amps required, but it will require a bit of care so that
the droop control does not significantly change the welding current at
the normal arc length voltage, which would probably be in the 25 to 30
volt range. You could check this with a data logger or a second
person to watch the meter while you weld.

So you want to convert welder current and voltage signals to a range
within the signal capabilities of your op-amps, and do a bit of simple
analog math as described in for example
<http://www.analog.com/library/analogDialogue/archives/philbrick/computing_amplifiers.html>
to compute the current set point for your controller.
I'm sure a search will turn up a ton of info on analog computers, but
George Philbricks "Computing Amplifiers" is a real classic.

You can implement all sorts of I/V curve modifications like this
including power control, but I still recommend starting with droop.

Sorry about the DROOP rant .

Sounds good Glen. You never have to say sorry to me, I appreciate your
guidance. Laid some 6013 beads last night, it was functioning very
solidly. Very good relaxation.

i