Weldernator project

[Ian]

The windings will be protected by the diode bridge, but the diodes need protection from voltage spikes and oscillations. Ive heard of a delta-wound configuration which breaks up the phases just enough to ca cel oscillations, but I'm trying to work with existing alternator hardware.

 

[Ian]

Oh, 6" of carbon rod ought to do it for a resistor but it would kill the arc.

 

[Pistolero]

Need some resistance to dampen oscillations. How much, haven't calced, but not enough
to kill the arc. Just enough to dampen oscillations. You can run 100 amps thru at resistor
with modest voltage drop, just not a big number of ohms - like a chunk of rebar.

OK, I thought you were going without diodes - yes need to protect them.

Bill

 

[Bret4207]

Used to be a commercial rig like that, a box you hook the welder to.

There still is.

 

[Bret4207]

Ian, should be plenty of dead gensets out there with 8-10hp Briggs or Kohler horizontals in them for little of no money. They run at 3K and have governors. HF Predators are an easier but more expensive alternative.

 

[RBHarter]

As much mod as you have done it doesn't seem unlikely that you could add an idler pulley to pull the belt contact up on the smaller pulley from 100° or so (if I remember the GM A/C location correctly) to the better part of 180° . Harder on the belt flex but better for the fuel consumption and run load . These days 10-12 run hours on a savings of 1.5-2 gallons would buy a new belt .

 

[popper]

MOV probably won't last long, they tend to break down after a lot of transients. Inductor is to reduce output ripple but will increase spikes when the arc breaks. Extra alt. coil would work so you don't have to make one. Almost all auto elec. use shunt reg. to kill the voltage transients (starter motor kick) and they are so cheap they usually are the first thing to go - look at transorbs an similar stuff. Quantas had several 747 grounded due to the transient problem - until the guy that designed the stuff could implement the fix I made and fly there to replace stuff. Yea, little 10W part could generate 1500 volt spike. L-C snubber across the diodes would help too. IIRC, harmonics from the delta are less but it won't control interruptions any better. Navy used delta for a leg loss but converted to Y for operation.

 

[Ian]

You're right about the MOVs, I can't find one rated for more than 10k-100K transient cycles at 150 amps, more of a sacrificial part than I want. TVS diodes may be the way to go as a shunt to soak up reactor spikes when breaking the arc, will investigate, thanks for the tip! Still wondering about a heavy cap in parallel with the output to help with arc stabilization, maybe not needed after all and just another source of oscillation in the loop between reactor, arc, and cap plates? I'm not sure if the Wye system with all the little peaks at the same frequency is going to set up oscillations within the three phase generator circuits or not, am thinking most of the oscillations will be on the DC-output side (welding lead circuit) because that's where the regulator is connected. The alternator's original self-exciting internal circuit had two sets of diodes: a bridge rectifier and another diode which may act as a snubber to prevent loops between the field circuit and the rectified field supply circuit, plus there are diodes in the voltage regulator according to schematics. I bypassed all that because it isn't rated for more than automotive voltage anyway, may have to look at the alternator schematic again and see if I can't duplicate some of the circuitry with higher-voltage components. The auto system also has the advantage of a huge honking battery to soak up a lot of current fluctuations, which the whole reason I feel I will need a reactor in series with the welding lead.

 

[popper]

AC start cap may help, they aren't cheap. Cap in your point dist. is for a snubber, not for radio interference. I forget what the arc resistance is but the sputter of arc break generates some mighty big spikes and any inductance (yes, the cables add some too) will increase the spikes, trying to maintain the arc. A big fast reverse steering diode behind the inductance will also help.

 

[freebullet]

This sounds like an experimentation required deal, fire in the truck & all, inclined to prove out before in the truck attachment/ testing. The technical, definitely beyond me.

With the 200$ e start 6.5hp gearing it is all that is needed. Ran a 5.5hp on a regular alternator as a kid. Had it in a boat on a lake where we could only use electric motors. Motor about, charge it up, run with what I had. Cheap to do with go kart axle, sprockets, & pillow blocks. Had 3to1 reduction, gov@3600.

Could swap alternator & weldnator with similar, & have a kinda benchrest to test before marrying potential fire hazards to the truck. I do believe you can pull it off, just thinking risk, saftey, reward from my own position. Carry on.

 

[Ian]

Ah come on, where's yer "hold da beer while I fire this thing up" spirit?

I dug deep into SMAW technobabble and found some things I didn't know. Different electrodes require different inductance. 7024 rods pretty much prefer none, just straight DC. 6010/6011 like a LOT of inductance, like a couple of mH or more or they won't burn well. 7018 likes a little bit of inductance but hates DC. This is stuff you never realize when running commercial machines....just pick the rod for the job, set the machine up to run the rod via a switch and/or a couple knobs, and let the mystery magic inside the sheet metal box take care of the rest.

The problem with caps is they help start the arc, but right after that the voltage sags while the cap recharges and that could make the arc extinguish. Using too much cap is the issue there. I'm not sure I even need a cap for certain rods, only some sort of TVS diode to protect the rectifier bridge. I can always make an in-line reactor with a breakout or two for running different rods, but the easiest thing for now is to skip it and run rods that don't require much "L" in the circuit. If I fry the external rectifier in the process of figuring out how to protect it, I'm out about $30 so no big deal.

 

[Pistolero]

Think about a ringback diode with resistor to let the back EMF from arc shutdown go to ground. On my aircraft
electrical system, it is necessary to have this on the main contacter control input because when it shuts down 12VDC
the contactor coil (very much like an old Ford starter "solenoid" which was a contactor) blows back a 300+V spike
which will fry the avionics.

Bill

 

[Ian]

Until I build the inductor and actually scope it I don't really know the characteristic of the spike I'm trying to suppress. 6010 5P+ is my rod of choice for pipen so I'll need an inductor.....maybe. I've burned nearly 20 pounds of them so far with an AC machine which is pretty much as far from correct as it gets.

If the weather warms up a little this weekend I might get to try it out.

 

[Brad]

Warms up? Ours would play hell getting any colder.

 

[fiver]

just a question but what about running resistance on the grounded side instead of the hot stick side?

 

[Ian]

The inductor (reactor or "L") is only effective at smoothing current, not voltage (need caps for voltage stabilization), and since the only significant current in a DC system is between the + supply and the load (arc in this case), it has to be somewhere between those two points in the + lead to do any good. If it were on the ground lead, all an inductor would do is add additional load to the circuit and rob current energy from the arc.

 

[fiver]

I thought this was an AC unit with the diodes removed to take the DC out of the equation.
if it's still running as a DC system I would throw a half trashed battery in there as a capacitor to soak up and barely return the power to the system.

 

[Ian]

I bypassed the alternator's rectifier bridge and brought the three phase outside the case, but hooked the legs straight to an external 200 amp, 1600 volt three phase rectifier to get DC. I don't think there's any way to connect three-phase power legs to make single phase due to the angles being 120 degrees off, they'd be swapping the same current between themselves and making heat but no output. A three phase motor direct-driving a single phase ac generator is one way to convert, but not efficient. The black box with the three red wires hooked to it in the last pic of the op shows the extenal rectifier bridge.

 

[Ian]

The half-trashed battery ain't a bad idea and would work if there were three in parallel and they were kept from getting too hot, otherwise just one becomes a 12.6V voltage regulator and that's not enough voltage to sustain a welding arc. I think Freebullet has a battery-only welder setup that I assume uses two 12V batts (most do only two, its for trail repair type stuff with 3/32" rods and very short duty cycles).

 

[freebullet]

This is a spool gun, can be run on almost anything for power supply.

http://readywelder.com/products-page/welders/10000adp/

We originally got it to weld aluminum. Our miller 250x needs a special spool gun & this made more sense. It can run 12,24,36,ect. The drawback on batteries is you can't adjust the voltage unless you do it by adding or subtracting batteries. It works well, the wire is live when hooked up, that takes a moment to get used to, but similar to a stick.

I've run it off batteries & the 250x, it is a really handy tool. The gun shouldn't be used with less than 30 or 35 wire, unless you like frustration exercise. Have used flux core steel, aluminum w/gas flow, & now have some flux core stainless wire to try. Real handy.