Triggered SISG
Finn Hammer, Tue Oct 31 2006, 09:41PM
All,
When I saw the SISG coil, I saw an IGBT stack with a safety gap across it. A basic building block that needed a trigger mechanism to really appeal to me. It should be triggerable even when using DC-Resonant charging, and it should be elegant.
I built a 2 section SISG because it would enable me to find out all the elemental things that I needed to know, and decided to start out with a transformer trigged SCR.
I selected the gate trigger toroid according to the TLAR criterium (That Looks About Right)
and so that the center would hold the insulation from RG214 antenna lead, because the toroids will be parked at elevated potentials.
The trigger wire is passed down the center of the toroids, and current from a 25µF cap is dumped down that wire controlled by an IGBT and a couple of triple5`s.
At the present moment, the circuitry looks like this:
There are 4 250V sidacs in series, and the SCR bridges the bottom 3. Therefore I can trigger from 250Volts and up, and the top SIDAC will turn off the current in the string, in case the SCR should be unable to do so.
Suppose that I connect this gap to a resonant DC supply, which has from 250 to 499 volts on it. I can trigger it. After the ringdown the capacitor will be charged up to 998volts. The gap can hold it off untill I trigger it.
The way I wanted it.
It works great, although I am unable to get the fast charging times that you get, Terry:
Yellow is main cap
cyan is pri. current
magenta is gate
Anyway, the final scope check without breakout:
Before first light @ 988V off a single CD 150nF 942/2000 and about 3 inches
Nothing like a first light, particularly not when it is a real "first"
Cheers, Finn Hammer
Re: Triggered SISG
Terry Fritz, Wed Nov 01 2006, 06:35PM
Hi Finn,
Cool Stuff!
You might only have to trigger just one section. Assume there are three sections at 1000V each for a total normal firing voltage of 3000V. If the voltage is at say 2500V, then just firing one section will fire the other two as well.
It depends on the charging circuit timing, but if the voltage is close to the firing voltage, just triggering one section will bring the whole array into conduction.
Cheers,
Terry
Re:
Triggered SISG
Finn Hammer, Wed Nov 01 2006, 09:36PM
Terry Fritz wrote ...
You might only have to trigger just one section. Assume there are three sections at 1000V each for a total normal firing voltage of 3000V. If the voltage is at say 2500V, then just firing one section will fire the other two as well.
I don`t agree. The DC resonant supply is the well known "Radar Supply" . with a charging choke and a de"Q`ing"diode in series with the transformer and rectifier.
I am sure you know it well
This supply is special because it initially charges the capacitor to the rectified voltage of the transformer(s).
Let`s say ~5000volts.
(3MOT`s in wye: sqrt2*sqrt3*2100V = 5143V
So the very first ringdown happens from ~5000 volts.
However, the 2nd. time, the cap is charged to ~10000volts
To get to the 2nd. charging cycle I have to be able to trigger at the initial voltage ~5000V.But the Sidac`s have to hold of the the ~10000 volts as well because if not, then the breakrate will increase without controll.
For these reasons, I need to be able to trigger the gap at less than the initial voltage (half of the totalstandoff voltage of the gap. and the gap has to be able to stand off the full charging voltage from the second recharge and on.
The gap has to be triggered every time in order to be able to run from a DC resonant supply.
However, it would be neat to be able to start the coil up on a variac, and this would require that the top sidac is a low voltage type, the lowest there is would do fine.
I am using a sidac at the top, because it stops to conduct when the voltage across it is removed, unlike the SCR, which only stops conducting when the current through it drops to zero. I am not sure it would ever do that if not for the top sidac.
Terry Fritz wrote ...
It depends on the charging circuit timing, but if the voltage is close to the firing voltage, just triggering one section will bring the whole array into conduction.
Cheers,
Terry
I hope that the only charging circuit timing I will need to worry about is the size of the charging choke: it should be small enough to allow the primary capacitor to fully charge in the time between ringdowns at the highest breakrate, yet big enough to allow the IGBT`s to turn off safely at the level the current has risen to, during the ringdown, where the supply is shorted.
Cheers, Finn Hammer
Re:
Triggered SISG
Steve Conner, Wed Nov 01 2006, 11:01PM
Wow, neat work Finn!
Re:
Triggered SISG
Terry Fritz, Thu Nov 02 2006, 03:40AM
Hi Finn,
You and I are thinking of completly different charging systems. It's all cool
Charge on!!
Cheers,
Terry
Re:
Triggered SISG
Daniel Uhrenholt, Thu Nov 02 2006, 03:13PM
Yearh, i cant wait to see this thing in action!
Re:
Triggered SISG
Finn Hammer, Fri Nov 03 2006, 10:47PM
Terry Fritz wrote ...
Charge on!!
Charging the triggeed SISG
Tonight I got the fiber optic trigger transfer link adapted to the SISG`s SCR gate trigger driver, and also, added a bare secondary from a MOT as a charging choke, so that I could try the resonant charging system at various breakrates, 60 to 1200 BPS.
This first picture shows the voltage across the primary capacitor (yellow) during the first 3 ringdowns of a 1000BPS burst.
(Magenta is the gate voltage on the IGBT that fires the current pulse down the trigger transformers. Soft recovery to avoid spikes!
Cursor 1 refers to the initial voltage on the primary cap, which is 352V (200x differential probe)
After the ringdown, the voltage jumps up to 816 volts, as pointed out by cursor2. This is more than double, so I suppose that is "good".
Next picture shows one ringdown,
revealing that the quench happens after 400µs. This could probably be shortened down a bit and I just now realize that a look at the charging current here is needed. I`l take that tomorrow. It will probably bring clues to the high recharge level on the primary cap.
This trigger stuff removes the super repeatability of the original SISG coil, but boy is it cool to have a coil that changes breakrate at the twist of a knob.
This coil is such great fun, It feels just like getting back to playing with a.....a _real_ Tesla Coil again.
Cheers, Finn Hammer
Re:
Triggered SISG
Terry Fritz, Sat Nov 04 2006, 06:30AM
Hi fine Finn ram,
revealing that the quench happens after 400µs. This could probably be shortened down a bit
It looks like the SISG is stying in conduction way too long (cool scope BTW
)... Once the primary energy is gone - plus a little, you might as well turn it off so it can start recharging again. Of course, there are all kinds of timing issues now with this triggered stuff
This trigger stuff removes the super repeatability of the original SISG coil, but boy is it cool to have a coil that changes breakrate at the twist of a knob.
The SISG is sort of "chaotic" anyway... But now the firing voltage is varying too... COOL!!!
I just figure the coil "knows what it is doing"
The sparks seem to love it!!!
For the PIRANHA drive circuit, I am posting a big "WARNING" tonight...
"""
Now that others are interested in the PIRANHA SISG design, be VERY aware that the PIRANHA primary circuitry is VERY DANGEROUS!!! There is NO ballast or current limiting for the MOT which pretty much directly connects and drives everything on the primary!!! The 165nF caps hold 4 joules at 8000V and can recharge at 500BPS!!! It is an easy and simple design, but it packs a LOT of "serious killing power"!!! They are "very different" than any other coils...
""""
Cheers,
Terry
Re:
Triggered SISG
Finn Hammer, Sat Nov 04 2006, 09:53AM
Just changed the Gate drain resistor to 900 ohm, for a nice 184µS quench, which takes place while the current out of the charging reactor has risen to 2 amps.
Well, this is low power testing with just one SISG section and low voltages.
I guess it`s time to wrap the stuff up on a soldered board and add some power.
Cheers, Finn Hammer
Re: Triggered SISG
Finn Hammer, Thu Nov 09 2006, 09:57PM
I am building a board with 10 SISG sections, all triggered.
I forgot to buy 33 Volt TVZ`s to protect the gates of the IGBT`s.
Can I use BZV85 zeners back to back instead. They are 1.3W types, and allow a non repititive peak reverse curent of 1.2A @100µS
I hope they will do, so that I can test the board during this week-end
Cheers, Finn Hammer
Re: Triggered SISG
Terry Fritz, Thu Nov 09 2006, 11:12PM
Hi Finn,
It will "probably" be ok... TVSs are super fast and powerful, but the gate to source voltage really should not over shoot anyway. The 24V one in the SIDAC string commonly goes to about 30V with a significant current hit. but the Vgs TVS should just be "asleep" normally. The 100nF cap and 50 ohm resistor too it really control the voltage there.
Cheers,
Terry
Re:
Triggered SISG
Finn Hammer, Sat Nov 11 2006, 08:46PM
Terry Fritz wrote ...
Hi Finn,
It will "probably" be ok...
Cheers,
Terry
I was lucky to find some 1500W TVSès, so I have been building intensively today.
So far all 10 sections work well, when tested individually, but of course, the final proof lies in applying power.
I`l probably do that tomorrow.
I found a center point in the 250V SIDAC`s, which gave access to a 135V SIDAC, so this gap can be made to trigger from 1350V to 10000V
I give a couple of pics. to show today`s progress.
Here is the Fiber optic trigger transmitter, the receiver and driver board, and the SISG itself.
And here is the 10kV SISG itself in all it`s gory detail.
Big files, but what.....
Cheers, Finn Hammer
Re:
Triggered SISG
Terry Fritz, Sun Nov 12 2006, 01:07AM
Hi Finn,
So far all 10 sections work well, when tested individually, but of course, the final proof lies in applying power.
I`l probably do that tomorrow.
Beware that it will go from dead silent to a big arcing in 1uS!!! It still scares the crap out of me the way SISGs "suddenly" start up at full power!!!! You WILL understand soon
I found a center point in the 250V SIDAC`s, which gave access to a 135V SIDAC
Yeah, they are center tapped. See how I have brass shorting jumpers on the SIDACs:
They do seem to like being "evenly" distributed.
Very cool!!! No worries about not heat sinking the IGBTs yet. they run pretty cool even at 250 BPS!!
Cheers,
Terry
Re:
Triggered SISG
Steve Conner, Sun Nov 12 2006, 01:26AM
wrote ...
It still scares the crap out of me the way SISGs "suddenly" start up at full power!!!!
Finn just said that his new design could be triggered over a 1:10 range of DC bus voltages. So it can be started at low power and ramped up.
Re:
Triggered SISG
Terry Fritz, Sun Nov 12 2006, 02:37AM
Yes, of course!!!
In my case, the PIRANHA just sort of grew "UP"!! I never "reduced" the firing power
Even though the coil's "form" has changed very much, I still just fire it off "first time" at 4800V!! It always does just the same thing. But after major form changes, it is always "interesting"
But it has never done anything "bad".
Unlike a conventional coil on a variac, SISG coils are either "full off" or "full ON"!! Silent, or firing like crazy!! Once you get used to it, it is fine, but it is very different that we are used too... We'll see if Finn notices this crazyness in his work
) It is sort of hard to change the firing tap voltage, so why bother
I always warn "spectators" and myself before that it is going to go "SUDDENLY FULL ON!!", so prepare one's spine
The last spectator was Bob Golka (yes THAT "Bob Golka"
) He did not flinch
Cheers,
Terry
Re:
Triggered SISG
Finn Hammer, Sun Nov 12 2006, 05:55PM
Today i finally had all 10 sections engaged at the same time. It functioned wonderfully!
The SISG gap, with these IGBT`s seem to favour operating conditions below 400Apeak, so to test the coil, I used 2 MOT`s in series, into a 50nF cap. At 150kHz, this gave me 400 Amps Pri Peak.
Notice that I am running the gap without any cooling finn`s, so I have to be real carefull to kep run times short not to overheat them.
I have a video here that shows the Triggered SISG gap in operation- Windows Media player will not play it, so right click and save!
What it shows is the coil, John Freau 4x13"toroid. The voltage applied to the primary is 4300 volts.
There is no breakout untill I press the trigger button, at which time it starts at 67BPS. First bang is then 0,46J, the following are 1,85J due to the charging choke.
I then crank the coil up to 1270BPS and back.
Starting power is 124W@67BPS, peak power is 2350W@1230BPS
Distance to the ground point is 30 inches, so I am not into spark length record breaking territory yet. But I guess that "Taming the SISG" is a feat in itself, so that can wait.
Scantesla! Is it able to figure out the best cap size @150kHz for maximum power and 400Primary amps max?
The help file says that the primary current limit is only for DRSSTC`s.
Cheers, Finn hammer
Re:
Triggered SISG
Terry Fritz, Sun Nov 12 2006, 09:09PM
Cool!!
You coil has a neat sound to it!! I think yours is the first SISG in Europe! - And the first triggered one!!!!
The help file says that the primary current limit is only for DRSSTC`s.
I "think" it will still work fine. Those functions are always turned on.
Scantesla! Is it able to figure out the best cap size @150kHz for maximum power and 400Primary amps max?
It should. You could also just graph peak current as a function cap size (mode 2). If you have any problems, send me the input file and I'll fix it.
Cheers,
Terry
Re:
Triggered SISG
Finn Hammer, Mon Nov 13 2006, 07:21AM
Terry Fritz wrote ...
You coil has a neat sound to it!!
Perhaps that is because the charging waveform still has a destinct 50Hz overlay on it. I
have not yet rigged up the 3 phase supply, because I have to go to the garage to do that, and it is cold there. (56 lat) So everything else better work well by then.
Anyway, someone asked for a schematic, here it is:
Pls notice, that when the SCR fires, the gap still is a SISG gap, only with a lower breaktown voltage. In my case, the section breakdown voltage at is 135Volts in the triggered state, and 1080Volts in the untriggered state.
Re: Triggered SISG
Steve Conner, Mon Nov 13 2006, 01:24PM
This is great! I reckon our "toys" have more or less caught up with the commercial state of the art in pulsed power now.
Re:
Triggered SISG
Finn Hammer, Mon Nov 13 2006, 09:04PM
Conserve Ten wrote ...
I reckon our "toys" have more or less caught up with the commercial state of the art in pulsed power now.
Thanks!
In that case I will have to change the source of current into the pulse line trough the gate transformers.
At present it is a humble 22µF cap @16volts dumping current in 10µS, producing a far from ideal (despite sufficient) waveform.
Now I will upgrade to a nice Rayleigh Line pulse forming network, where the inductance of the pulse line is also the output inductance.
Nice!
Cheers, Finn Hammer
Re:
Triggered SISG
Finn Hammer, Thu Nov 16 2006, 09:20PM
All,
I`ve just started a webpage with some information about the triggered SISG.
Hope it helps.
Cheers, Finn Hammer
Re:
Triggered SISG
Daniel Uhrenholt, Fri Nov 17 2006, 06:32PM
Hi Finn.
Your SISG coil is looking great!
But why is there stuff about your RAT coil on the new webpage
Some day next year I`ll drop by your house and see it in action
Cheers, Daniel
Re:
Triggered SISG
Finn Hammer, Sun Nov 26 2006, 09:42PM
All,
In another thread, Aaron said:
J. Aaron Holmes wrote ...
Heheh....well, that means the title of "First To Blow Up An SISG" will probably be mine
Regards,
Aaron, N7OE
And in yet another thread, Terry noted:
Terry Fritz wrote ...
I note Finn has not blown any IGBTs either unlike the DRSSTC systems tend to until they are all figured out
Cheers,
Terry
But of course, I took the price:
In an attempt to get beyond 44 inches, I did something silly:
I added smoothing caps, 330µF/4500V and then I hooked up a charging inductor without soldering the leads, or crimping them. Just twisted the wires together.
Well, the 2 wires of the inductor touched each other, and thus offered a free and unrestricted path for the 3341 joule residing in the smoothing caps.
That cracked the IGBT`s open. All 10 of them
I am not able to tell if the rest of the circuitry survived, but time will tell.
Cheers, Finn Hammer
Re: Triggered SISG
ragnar, Mon Nov 27 2006, 01:39AM
**weeps with you for IGBTs**
..a moment's silence..
did you get any more spark/setup photos prior to explodage?
Re:
Triggered SISG
Terry Fritz, Mon Nov 27 2006, 11:11AM
Finn Wrote:
Well, the 2 wires of the inductor touched each other, and thus offered a free and unrestricted path for the 3341 joule residing in the smoothing caps.
I blew a six string too in a ball lightning experiment early on too. It was at a party and I did not think it though.
We dumped about 900J into the SISG string which shorted all the IGBTs but did no other damage.
So they can be blown due to accident or giant energy discharge. At 3341J, Finn holds the record for the most energy
Mine did not blow the IGBTs apart or anything, they we just all shorted.
So SISGs "can" be broken if one really tries hard
So far, directly discharging really big caps into them is the preferred way. I think overheating is the biggest danger in a normal situation.
Cheers,
Terry
Re:
Triggered SISG
Steve Conner, Mon Nov 27 2006, 03:05PM
Don't be sad BP, it's all in the name of science
Re:
Triggered SISG
Daniel Uhrenholt, Mon Nov 27 2006, 03:16PM
Firm He-man(n) wrote ...
I added smoothing caps, 330µF/4500V and then I hooked up a charging inductor without soldering the leads, or crimping them. Just twisted the wires together.
I think i heard that one before
Did this happen while testing it on the Millinium coil?
Re: Triggered SISG
Finn Hammer, Mon Nov 27 2006, 08:13PM
Autopsy report.
Hoping for the best and expecting the worst, I measured the components of the SISG.
The SCR`s are fried, The 24V TVS`s are fried, and the first section of the top SIDAC`s are fried. The other half of the top SIDAC, together with the 3 bottom SIDACs are the parts bypassed by the SCR, so it all makes sense.
Well, that is it then, I am not going to perform a rebuild. The 650Apri(max) restraint is not my cup of tea. I`m going to pursue my goals by stacking larger IGBT`s.
I got my 44" of spark, a lesson learned, and proof that the trigger circuit worked. That makes it a worthwhile experiment.
So let it RIP
Cheers, Finn Hammer
Re:
Triggered SISG
Terry Fritz, Tue Nov 28 2006, 05:20AM
Hi Finn,
Cool!! A LOT more damage than I was able to do with "just" 800J
I rebuilt mine (easy), but lost interest in that design for bigger and better things. So I shipped the old one off to Conner. Knowing that, someday, he will do something far greater
He is one with that "God's eye" perspective over all this...
Two years ago I sent about 10 CM600 bricks to Ward... Starting to payoff now in that other post about 14 foot DRSSTCs
Looking forward to seeing what is next!!!!!!!
I have been all lazy here these days and have not done anything
Cheers,
Terry
Re:
Triggered SISG
Steve Conner, Tue Nov 28 2006, 11:06AM
Well, I was going to try the triggered DC resonant charging thing. But Finn beat me to it and came up with a better triggering arrangement than I'd have thought of. As for what's next, it looks like we now have mini versions of all the toys the particle accelerator guys like Greg Leyh use, so not sure what to do next
I do think the SiSG is the way to go though: it's more powerful than the OLTC and not as temperamental as the DRSSTC.
Re: Triggered SISG
Tom540, Mon Oct 29 2007, 08:47PM
Would it be possible to use scr's instead of IGBT's? Since they are cheaper. just have a chain of them and trigger them simultaneously? Or am I missing something?
Re: Triggered SISG
Ken M., Tue Oct 30 2007, 12:19PM
If ur using DC, the SCR's will trigger then remain on, so long as there is current.
Re: Triggered SISG
Tom540, Tue Oct 30 2007, 02:14PM
Good point. I want to try this but I don't like having to use 10 or so IGBT's .
Re: Triggered SISG
Electroholic, Tue Oct 30 2007, 03:18PM
Or you can use a big IGBT, i think its called BRISG.
Re:
Triggered SISG
Dr. Dark Current, Tue Oct 30 2007, 03:43PM
Tom540 wrote ...
Would it be possible to use scr's instead of IGBT's? Since they are cheaper. just have a chain of them and trigger them simultaneously? Or am I missing something?
No it would not be possible, because the turn-on of SCRs is way too slow. Maybe you could use fast type SCRs but I don't think they are cheap.