ZVS Driver Induction Heater As A Bolt Buster & Power Supply Question

[fierygt]

I have a few questions, which the first is about taking one of the ebay 1000w ZVS drivers, and converting it into a bolt buster for rusty bolts on cars.  I assume it's possible, but I am wondering if anybody might know how much power that sort of usage might draw.  Because my second question, is I was wondering if it was possible to take an 18-20v lithium ion tool battery, and use that as the power supply in that scenario.  Would there be any issues?  I understand it would drain the battery pretty quick, but the commercial bolt busters usually heat a rusty bolt up within a minute or so.  Would it provide enough power for the issue of start up on these zvs drivers to get them resonating?  From what I have read, that is the important step to keep them from blowing components.  I don't need it to run for a long time, just enough to heat up rusty bolts once in a while.  Also, would reshaping the coil or making a coil to a typical bolt size cause problems, assuming the rest of the project was feasible.  I only learned about these a couple months ago, so I am trying to gather as much knowledge as I can before doing damage.

One of the zvs drivers I am looking at has specs of Voltage: DC12V-30V, Max. Current: 20A, Max. Power: 1000w  which I can provide a link if necessary, but figured anybody here seeing this would probably already know what I am talking about.  Thanks for any help.

[T3sl4co1l]

Yes.

Obviously you can only run at, whatever 20A max, for as long as the battery can provide it at current state of charge.  So, a 5Ah pack could sustain that for 5/20 = 1/4 hr or 15min.  Enough for a few bolts, it would seem.

Power: as much as you can.  For a given supply voltage, power goes up as number of turns in the work coil is reduced (and preferably, the capacitance is increased proportionally* to keep the operating frequency within range).  You should probably have a means of controlling supply voltage (buck switcher), so it's not just full throttle as soon as you plug it in.

*Proportional to change in inductance.  Which actually goes as turns squared, generally speaking, so you tend to need a huge amount more capacitance to do like one or two turns, versus three or four.  At some point this is impractical, and you'll need a transformer to impedance-match the load, or a boost supply to get more input voltage.

But also not too much power, so as to honor the current rating.  This can be improved by rebuilding the circuit with bigger transistors, heavier traces, more capacitors... Or same idea, get the bigger model (there's a 2 or 3kW version innit?).

Water cooling will be needed for the coil, even at fairly low power (100s W), and is a good idea for the power supply, too.

Tim

[Doctorandus_P]

One of the the main failure modes of these cheap ZVS circuits is a too low input voltage, which results in insufficient gate drive for the MOSfets and that releases the magic smoke from the MOSfets.

Overall it's just the minimum circuit that sort of works. Adding some decent gate drivers and some under voltage protection is a nice upgrade for these.

[fierygt]

Thanks, too low of input voltage was what I was also wondering about.  Would an 18 volt lithium ion battery still be too low, even if it's rated between 12-30v?

[T3sl4co1l]

That should be fine.

Tim

[fierygt]

Ok, so I finally got the zvs driver here and tried twice to run it off an 18v 24Wh battery and twice, the terminal block that the coil screws onto, melted the solder and fell off within less than 2 seconds!  I soldered it on once, and tried again, but I don't want to keep doing the same thing over and over and damage something unless it came damaged. So I am stuck and can't figure out what I am doing wrong to cause that. 

[strawberry]

reactive power in LC tank cause loses from ESR
20..40A here and not optimal frequency or coil design could cause
cable terminals might be made out of iron(high electrical and magnetic loses)
I ran water cooled coil 90kHz 50V 2.5kW and ~100W loss in coil and circulating current ~100A

[fierygt]

I understand water cooling, if it's being ran.  But...again, I could only power on this thing for two seconds, before the terminals fell off the pcb from melting solder.  I tried once more with my bench top power supply, and ran it at 24 volts, and it looks like it pulled about 6 amps or so before the terminals again fell off the pcb.  Each time, I resoldered them back to the board.  I haven't even had a chance to put anything in the work coil to even test it, before it seems to fail.

[DavidAlfa]

Absolutely impossible that those little pins can stand the huge LC tank currents.
While the input might be only a few amps, the current in the coil can reach hundreds.
You need a much beefier connector, or directly solder the coil to the pads.
Would be better to reinforce the traces going to the capacitors with thick copper wires.

[BrokenYugo]

Looks like they cheapened the design since the last time I looked at these, I recall the coil mounts being made of brass standoffs screwed to the board.

[strawberry]

here is my mini version with M6 coil terminals and ATX 12V connector

[fierygt]

Ah...that looks to be the issue.  I soldered a coil directly to the board and it seems to be working like it should!  The mosfets and caps are definitely getting warmer than I'd like, but I can work all that out through testing and cooling.  The good news is, it seems to be doing ok with a ryobi 18v 24Wh battery, at least in the short term.  I'm looking around for some screw terminals that will be less dense enough to be able to solder to the board, and also hold up without popping off like the factory one kept doing.  Any ideas?  I'll solder directly to the heavy duty trace.

[fierygt]

I took davidalfa's advice, and went right to the traces underneath the board.  Replaced the little terminal screws for a couple of pretty beefy copper lugs soldered directly to the trace.