[NewCandle] Schlicher 'drive'

[Kyle R. Mcallister]

Keith Nagel wrote:
> Hi Kyle + John,
> 
> While FETs are great, they get pretty pricey when you are
> trying to control large currents. The situation is somewhat
> improved with IGBTs, but still you're talking about a
> bit of cash ( but a good investment all the same ). SCR's
> are the cheapest thing for the job, and for relatively
> slow risetimes like what we have been discussing more than
> adequate for the task.

Alright, then I will concentrate on using SCRs as you suggest. 
Suggestions for power ratings with a decent margin of safety? Peak 
current, etc..? What voltages are you suggesting to use? Obviously this 
will depend on the resistance of the antenna itself, but can you give me 
a ballpark idea?

> Which brings us to your problem, getting them to turn off.
> The trick here is to design the circuit around the SCR.
> There are two issues here. First, your gate drive pulse
> must be >10 times shorter than your power pulse duration.
> This ensures that the gate is off when the time comes
> to extinguish the SCR. Second, your power pulse voltage must
> be near zero, or better, reversed. To achieve this, we
> design the capacitor and the load resistor so that we
> have either critical damping or underdamping.

Noted. Some is a bit Greek to me, as what I've learned over the years is 
a bit patchy. Some things I am very familiar with, others, not so much. 
I understand the need to get the voltage to zero or below.

> So let's put some numbers to this. You tell me how long the
> pulse needs to be, and I'll tell you what value cap you'll
> need and how much resistance in the total circuit will
> critically dampen it. 

Unfortunately, I don't know how long the pulse needs to be. John didn't 
discuss it very carefully...he indicated pulse rep rates of maybe 1 to 
50 cycles per second. He spoke of short duty cycles. 10 percent duty 
cycles, so what, 100 milliseconds? It may be much smaller than that. The 
Schlicher patent and paper aren't entirely clear on this. The patent 
suggests 1 to 100 cps. The pulse rise time is said to need to be shorter 
than the "magnetic diffusing time" of the flat conductors. Not entirely 
clear on this. Also in the patent, 10 volts is suggested (John said 2 or 
  3 volts is fine), with peak currents of 1 to 10kA. Assuming 3 volts at 
1000A, and a duty cycle of 10%, that's 300W continuous.

> That resistance value will only be
> a guide, because at these currents the distributed resistance
> of the antenna and the rest of the output circuit will be
> substantial. I use a special low resistance meter to measure
> such circuits, not sure how you're set for such things but
> I'm guessing not? An ordinary multimeter _might_ be used
> but most of the units I have seen only go to .1 ohm and
> that's not very useful. Perhaps you can hit up the local
> uni's ( U of R. or my old haunt RIT might be able to help )
> or prevail upon some other local source for the meter.

I might be able to rig up something to measure milliohms. I remember a 
project in Radio Electronics for this. Might be worth heading up to the 
library and digging through the archives.

> As concerns the capacitor, and oil filled cap would
> be best suited for the application. Often these can be found in surplus or
> at ham fests ( big one up in your area every summer, keep
> an eye out for that ). Polycarbonate caps could also be used,
> it makes sense to buy a lot of small units ( say .2 microfarad
> @ 1KV ) and gang them up to get larger values of capacity
> or voltage. 

I've got oil filled caps as well. A lot of motor-run (NOT motor start!) 
caps laying around. 10uF to 30uF apiece, ranging from 175VAC to 660VAC 
in rating. A few 1uF caps scavenged from microwave ovens as well. Yes, I 
confess...I am an electronics dumpster-diver. And lately, the gang of 
'hangers-on' that come over on Friday nights has been gathering things 
for me to use. So I get plenty of junk for no cost at all.

> You mentioned electrolytics; we _might_ be able
> to use these, but again, you risk destroying that cap if
> our inductance is too high and damping resistance is too
> low. But given the cost, they're the cheapest solution and
> I'm guessing you'll want to go with that. Best to spend
> what money you have available on the switching circuit.
> Try ebay for the SCRs, often those hockey puck kinds are
> available at reasonable prices. For this application,
> a _small_ capacity electrolytic will be preferrable to
> a large one. The ones you mention are maybe too big for
> what we are considering, but I'll know when you spec the
> pulse width.

Alright. Again, not sure of the pulse width we'll need. I'll go find the 
AIAA paper and see if it gives any specs. I don't think it did. But 
knowing 1 to 100 cps rep rate...maybe we can ball park this, and make a 
circuit that can be interchanged with various cap sizes?

> I'll post a bit later with a pic of what a typical SCR switch
> looks like, along with a circuit diagram. It may take some time
> as I have neither pic nor diagram at hand, but all my stuff is
> unpacked now so at least the kit is at hand.
> 

Sounds good, appreciate the help with this project.

--Kyle