Hello,
Im new to CNC Zone. First of all I would like to thank all of you guys for keeping this thread alive so long:-). I have recently aquired via the evilbay some massive torroids. They were from a big TIG welder and are 3.4KVa each.
I have done a lot of research since then and hope I can help.
First lets hit on this HF start thing and then work back. If you look at:
You will see that he uses a distributed gap (gaps in series). This allows a lower dissipation in each gap. This is good because it means that each gap will heat up less. Also it allows fine tuning of the total gap. He has labeled his as 8 thou (mills) gap, I guess this is total, so 2 thou each. More information on this set up is easy to find on Tesla coil sites. I really do wonder if one spark plug will last long in this sort of service. I am going to use some 4mm OD tungsten. I got it from the usual place, 1 quid for a 100mm length, I bought two lengths, I will cut it up with a dinky diamond disk. I will mount this in some half inch square brass bar I have using little 3mm grub screws to allow adjustment. Tungsten can be brazed for those with the inclination, this would mean having access to a surface grinder with a green wheel to set them all level, too much messing about for me. Now we know the sort of heatsinking and precision needed for reliable operation over the long term. This is imperitive, I had whack from an old school tig set up that knocked me out. The fact my mate thought it would be funny to zap the back of my neck with it was not the real reason(OK yeh it was, dead leg issued). I was not too pleased as you can imagine. The gap had eroded, ablated, whatever. It was giving out 5-10Kv at one or two pulses a second with a pretty high current. OUCH. Stick the filler to the job, try and use the torch hand to get it off and wollop. Yeh not good.
As for getting a TIG welder from your old stick gear. I think its cracking idea. I have not seen much recent use for full on sine wave AC. So a rectifier will be needed. The really old 50-60Hz ac set ups dont cut the mustard. You will go through electrodes like they grow on trees. So rectifier it is then. But which? There are some choices. Normal bridge type rectifier. Pros- easy, relatively cheap. Cons- no current control. This should not be a problem if you already have current control on you stickamajig. SCR controlled rectification. Well this is the one really for a DC set up. Fully controllable, some use an auxilliary supply to give really low arc currents. This is how they weld razor blades, ally foil etc. (they dont but it is possible, LASER is the way for this sort of thing).
The one trouble with any sort of semiconductor output stage is adding the HF start. The HV will puncture the diode/IGBT/MOSFET/SCR junctions easily. The thing is its HF, normally >500KHz or so. This makes it reasonably easy to bypass to ground (a relative term, the earth clamp) some film caps will do the trick. Most also use tranzorb type avalanche diodes. If anyone knows more about ratings type numbers etc I would be very pleased to hear about them? Also another good reason to ensure you have the best spark gap possible, as any wear over time will lower the frequency and destroy your expensive semis.
Anyway, this is the stuff I have worked out so far. Here are some Patents that make interesting reading if you can get over the usual " cover it all and sue" style of most US patents.
US6388232 - interesting
US5229930
EP 1 782 907 A2 - this a Euro one.
This thread is pretty old, but I felt I had to thank the contributors for setting me straight. I have just built an HF arc starter using spark plugs. It had a nice spark at the gap, but no arc at the output. Then I saw the warning about about resistor spark plugs and DUH!!!! the light bulb came on! I had used resistor plugs! I switched them to gaps using coat hanger wire to try it, and it worked great! Thanks to this forum! Now I'll tweak it a bit to optimize performance. The basic circuit I used is Chris McKinnon's. I can post pictures or schematic if anyone's interested.... Thanks again...
Thanks for that, I was thinking of bumping this topic anyway. Please post your schematic from Chris McKinnon.
Any updates from anyone? Punisher, et al?
Anyway, regarding the spark plugs - I noticed before that my "non-resistor" plugs were still reading as 5k ohms, so after reading your post, I wonder if this is why I couldn't get it working.
There are 2 arrangements I've seen people use for this project, and they are similar to the diagrams here about Tesla coils. Notice the spark gaps in parallel with the HV output is preferred to protect the transformer.
Anyway, when I was working on this (nearly a year ago), when I connected it up that way (diagram 1), I got no sparking on the gaps (though I did get maybe 0.25mm arc from the torch to the earth clamp). I believe it's because I have a 20kHz switchmode NST and the path of least resistance was through the capacitors/coil, not the spark plugs (4 at 0.5mm and 5k ohms of resistance each), so it was essentially shorting the gaps. I'll try as you say, with coathanger wire, and see if it works.
My other theory is that a higher frequency supply (20kHz vs 50/60Hz), would need lower capacitance for the same reactance, so instead of maybe 1500-2500pF, something like 5pF should be used to give the same reactance. There's also the coupling transformer though, which increases reactance as frequency increases, so it's possible that the number of turns would need changing too.
Has anyone had any luck with using switchmode supplies at these higher frequencies? Or any comments from someone with a really good understanding of AC theory?
I've already bought all the stuff to make this with the dimmer/cap/ignition coil, but it would be nice to get it working with the NST.
I've attached the schematic and text description of the McKinnon Arc Starter. The spark gaps can be in series and don't have to be on opposite sides of the coupling coil. Same for the 5 ohm resistors. When I get a chance I'll try to make a photo of my setup.
After I saw the note about resistor plugs I went to the garage and dug out my spare lawnmower plugs - Champion CJ6, definitely NOT resistor, but I decided to measure them anyway. One was zero ohms to the electrode, the other was open! And these are brand new, never used plugs. I tried using just the good one, got a nice spark at the plug, but no arc. This did not surprise me since I had read somewhere (on a Tesla site I think) that you need several gaps in series to facilitate quenching the spark as the input voltage drops. Then I switched to coat hanger wire, which worked fine and I can replace with Tungsten TIG electrodes if the wire burns away too quickly.
Your Tesla site says that the HiV transformer needs to be able to survive high frequency voltages. The McKinnon circuit incorporates a low pass filter to protect the transformer - nice!
My understanding of how the spark gap generates High Freq is that the HiV transformer charges up a capacitor which is then shorted through a tuned circuit, formed by the capacitor and the primary of the coupling transformer,
by the spark gap firing. This happens 120 times a second for a 60Hz transformer. The ringing frequency is determined by the reactances of the capacitor and the inductance of the primary winding. If you use a 20Khz source, then the repetition rate is merely higher. The frequency is still determined by the tuned circuit.
My next step is to set the coat hanger wire gaps to 8 thou each and vary the number of turns on the coupling transformer. It's now 14:14. I think I'll try 8 0r 10:14 to try to raise the output a bit. I put some ferrite rods into the winding core to improve the coupling, but I'm nervous about using too much to avoid coupling any 60Hz, dangerous, voltages....
Thanks for posting that. Hmm, okay, so if the resonance is determined by the caps and the coil, then the frequency of the HV supply doesn't matter?
As it's been nearly a year, and I'm trying to relearn all the stuff I researched so I can get this finished. The other thing I didn't understand - if the suppression network (say, 250V 0.47uF cap and resistor in parallel) is the only path for the HV/HF current after the coupling transformer, how come the capacitor doesn't have to be 15kV? I know it's fine in practice - I think mine is about 500V, but why doesn't it need to be an HV cap?
Yes, that is my understanding. The high frequency ringing voltage is determined by the tuned circuit, not the excitation frequency (EF). Of course if the EF were close to the ringing frequency then that's another matter.
The high frequency high voltage is developed across the secondary winding of the coupling transformer. One side goes to the torch, the other to the supply, from the buzz box, which is relatively low voltage above the "Work" wire (about 70 volts open circuit). The suppression cap is to keep any high frequency high voltage away from the rectifiers in the buzz box (if any). If no rectifiers, it just protects the transformer. That's why the "Work" wire has to be brought to the HFAS, so that the high freq high voltage is suppressed right in the HF box, and can't get to the buzz box. At high frequency, that suppression cap is almost a short, so very little voltage develops across it. I don't know if this helps explain things, but that's how I see it.....
I just checked my spark gaps and they're around 10 tho. I rewound the coupling transformer to 9:14 ratio. With no ferrites, the arc is around 1/8 inch, with one rod, it grows to 3/16 " and with two ferrites, it blasts along at 1/4". Now I want to rig up a probe to my oscilloscope so that I can determine the frequency to reassure myself that it's high enough not to fry me. With all the arcing going on, the 'scope is almost useless. I need to add suppression filters to the HFAS input power lead and the 'scope supply to see if I can get a stable image.....
Thanks for trying to help, but the translation of that forum to English is still pretty much impossible to understand in any sort of useable detail. Can anyone make sense of it and explain it?
I too could not make much of Google's translation of the Polish forum. Pity!
I finally got around to taking some pix of my breadboarded arc starter. I've attached three. They show it from various angles, one with a nice arc. The coat hanger spark gap is clearly visible. Right now I only have two ferrite rods in the coupling coil. I'm waiting to get some more from an eBay seller. The transformer is 5000 v @ 30 mA. Its type is listed as "Fly Grid". I guess it was made to zap flies. The 500 pF doorknob across the HV terminals is part of the low pass filter keeping the RF out of the transformer. The resistor part of the filter is seen vertical. I only used three 591 pF doorknob caps for the resonant circuit. With four I thought the frequency would be too low. it arcs fine. Now I'll take it apart and build it into a proper metal shielded box. I'll use a remote foot switch to power a relay to turn the arc on, so that I can control it easily and not run it more than necessary. A noise suppression filter on the line supply is just barely visible in one of the pics. Otherwise the circuit is as described by McKinnon, which I had posted earlier.
The white doorknobs I actually got from eBay for a few dollars each about a year ago. I see that now they are going for about $35 each!!!! I looked around the electronics surplus stores in Toronto a year ago and they were all out of them - it seems that Tesla fever had struck and they had all been gobbled up. It seems that that might still be the case, hence the high price in light of short supply. The little brown doorknob came out of some high voltage research equipment that had been disposed of in the building where I worked.
I can make a capacitor bank for about $5.00 to $10.00 with surplus capacitors, it would have bees simpler to just buy one that would do the job.
E-Bay had some doorknob capacitor, a little while ago, that where reasonably priced but the shipping rate was indecent.
I noticed, in Post #88, that you mentioned that you measured the resistance of a brand new spark plug and it read "open", some spark plug have a "spark gap" inside them, it could be what you have, if it's the case, it might still work.
Hi Orville: Yes, I knew that some plugs had an internal gap, but here I had two IDENTICAL, brand new Champion CJ6 plugs, presumably bought at the same time, yet one had zero ohms to the electrode, the other open. Maybe the one with zero ohms shorted out an internal gap. At any rate, I'll try them both in my lawn tractor and it'll be interesting to see what difference there is.
I'm still waiting for my eBay ferrites to finish putting the arc starter into its final box. More pictures when I'm done....
Reply with Quote
Home Made AC TIG weld in Poland
"5,0"
Posting Permissions
