OLTC project / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

OLTC project

first previous 12 13 14 15 next

Move Thread

LAN_403

Finn Hammer

Sat Jan 06 2007, 12:38PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Jrz126 wrote ...

What frequency should I try to measure the L at?

My primary cap is 12.92uF, and it'll be running off ~600V. Maybe alittle less, my diodes are only rated for 1200V each.

The charging inductor size has to be big enough to keep the current rise during ringdown within a reasonable level. The de-Qing diode is a limiting factor here. And it should be small enough to allow the capacitor charging current to drop to zero, before the next Bang.
If it is too big for that, the voltage rise will increase beyond the 2x bus voltage that we ordinarily assume.
Anything over a 100mH choke with 600V buss and 12.92uF cap, the voltage of the cap will increase to over 1500 volts after the first 15 bangs.
To get a predictable voltage across the cap, the inductor should be maximum 25mH for maximum 500BPS.

Another factor that determines the final voltage across the cap, is the ringdown time, or quench time.

If the coil is set to quench at less than 100ÂµS, then the final voltage stays pretty much at 2x the bus voltage, but longer quench times start to have an impact on the final voltage.

If we still assume your 12.92uF cap, 600 Vbuss, 25mH choke and 500BPS, then the final voltage on the cap, and the peak charging current, varies like this with quench time:

100ÂµS.....1199V.....13.7A

200ÂµS.....1220V.... .14.2A

300ÂµS.....1257V.....14.9A

400ÂµS.....130 6V.....16.1A

500ÂµS.....1336V.....17.5A

So you see that things start to get funky, and again, with Tesla cols, there is that little extra "twist" to already established theory, which makes it fun to work with.

What this all boils down to, is, that you really owe it to your self to build a crowbar circuit, so that you don`t send all your igbt`s to the overvoltage failure bin.
It is your guess how I have come to this conclusion

.

To measure your choke, apply 50(60) hz across it with a variac, measure voltage across it, and monitor current trough it.

Then u = wLi

I hope this helps.

Cheers, Finn Hammer

Jrz126

Mon Jan 08 2007, 11:28PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

Thanks for the help Finn. I'm definitly gonna use a crowbar ckt. Which wont be hard at all since I'm using Steve Conner's OLTC2 circuit board.

I did some testing over the weekend and I was able to make some decent progress. Mostly because Steve Conner was in the chatroom and I was able to get some realtime advice.

The problem I am running into is that I'm not able to see the first notch in the primary.
http://www.personal.psu.edu/jrz126/images/oltc/no_notch_1.JPG
The top trace is just a probe hanging 5' away. the bottom trace is across the primary coil. The bottom probe is set to 10X.
Any ideas why I'm not seeing the first notch?
I tried decreasing the coupling by moving the secondary up about an inch. but I didnt see any improvent.
I also removed and added a bunch of caps, and I could see some increase in output as I remove the caps.

Heres a few pics:
The 'powersupply' that I am using:
http://www.personal.psu.edu/jrz126/images/oltc/powersupply.JPG
I'm driving the transformer with my Variac so that my function generator and Oscope dont ground out the supply.

The modified MOT:
http://www.personal.psu.edu/jrz126/images/oltc/modified_mot.JPG

The Cap bank and rogowski coil:
http://www.personal.psu.edu/jrz126/images/oltc/cap_bank_rogowski.JPG
It's not completely finished. Hopefully the stray inductance isnt killing the performance?

Finally, A video of the small sparks I'm getting now:
http://www.personal.psu.edu/jrz126/images/oltc/TC_and_pendulum%20059.mpg

Steve Conner

Tue Jan 09 2007, 10:21AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Hmm, from the look of the scope traces, your coupling is too tight and you're out of tune. The IGBT and capbank assembly will have considerable stray inductance due to the way you laid it out. The loop area between caps and that big thick return wire is pretty huge, and that is probably what makes the tuning be out. You could try moving the return wire round the other side of the cap bank.

Also, you ought to check that the zero current turn-off thingie is working well before you crank the voltage up. I can see quite a spike when it turns off. It should be obvious when it starts working: the Rogowski coil LED will change state, and the gate drive pulse will suddenly get shorter, and will "snap" to the nearest negative half-cycle. As you adjust the hold-on time pot, it will jump along in increments of one cycle. It should never turn off partway through a cycle.

Of course, if you put the Rogowski coil on backwards, it turns off at exactly the WRONG time for maximum spikage

So better check that. I always just tried it both ways round while using a second Rogowski coil to look at the primary current and check that I got it right. The turnoff should always happen during a negative half-cycle, that is to say, one with the opposite polarity to the first half-cycle you see.

The capacitor C28 in the Rogowski coil circuit might need made bigger, since your coil is lower frequency than anything I ever tried, and that cap has to integrate the Rogowski coil signal (they naturally measure di/dt)

Glad to see someone is using my driver board btw!

It was designed for a commercial installation, so it's a bit more complex than it needs to be for a hobbyist project. Good luck with it!

The schematic:

A page describing how I tested my primary circuit:

Some more pics I found while looking for the schematic

Jrz126

Tue Jan 09 2007, 01:29PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

Steve Conner wrote ...

Hmm, from the look of the scope traces, your coupling is too tight and you're out of tune. The IGBT and capbank assembly will have considerable stray inductance due to the way you laid it out. The loop area between caps and that big thick return wire is pretty huge, and that is probably what makes the tuning be out. You could try moving the return wire round the other side of the cap bank.

I was talking to Steve Ward last night on AIM. He also mentioned the tuning. So I went and removed 8 more caps and I started to see the notches. I couldnt fire it up because everyone was sleeping. I'm down to 9.4 uF and still not in tune

And I cant run the wire around the back side because there are 60 more (unused now) caps back there.

I think I might try to go down to a single turn primary, but I'm gonna remove some more caps and see if I can get it to throw some bigger sparks.

The rogowski coil is going to need some debugging. I think I may have swapped a couple resistors. I will also try changing C28.

Well I better get to work, I'm planning on ducking out alittle early though.

Finn Hammer

Tue Jan 09 2007, 07:01PM

Registered Member #205 Joined: Sat Feb 18 2006, 11:59AM
Location: SkÃ¸rping, Denmark
Posts: 741

Your coupling is up towards 0.28 to 0.3. You should raise the secondary up to the point where the bottom turn is level with the top edge of the primary.
Later you can lower it a bit.
The tuning is off, and may well be off by what amounts to 1/3 the primary caps.
This is a suggestion based on simulating and trying to get similar waveforms as yours.

With a 2 turn primary, I`d think you have just enough Q to resonate the circuit up with a function generator across the cap. The IGBT has to be shorted out with a soldered connection, or perhaps just turned on permanently.
However, in a situation like this, solid evidence supplied by a measurement is the way to go.
Measure your primary and secondary resonant frequency ASAP, and you will have your answer.
Building low inductance caps is not at all easy. But an OLTC can produce great sparks even quite far out of tune.

Cheers, Finn Hammer

Jrz126

Tue Jan 09 2007, 10:16PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

I got the rogowski coil circuit to trigger! It was a bad solder joint. Its turning off when the voltage across the primary is near the negative peak. does that mean the current is 0 at that point?

Finn,
Thanks for the info, what program are you using for the simulation?
You were just about dead on with the 1/3rd extra capacitance. I'm down to 7.5 uF and I made that measurement when I was at about 10uF.
I'm seeing the first notch after 2.5 cycles.

Still no improvement in the arc length though

What should I try next?

Steve Conner

Tue Jan 09 2007, 10:32PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

The peaks in primary voltage are the zeros of primary current (V=L*di/dt and all that). A negative peak means that the primary current is passing through zero, changing from positive to negative. IOW, the current just finished flowing in the IGBT and is going to spend a half cycle going through the inverse diode. This is just the right time to turn off.

That assumes that the voltage probe is wired up correctly such that the first quarter cycle of primary voltage was positive.

The next step is to adjust the timing so the IGBT turns off at or immediately after the notch. Then the setup is complete and you can crank the voltage higher. Probably best not to crank it too high until you've installed the crowbar circuit, though. Note it needs a high voltage resistor which isn't installed on the board for safety reasons. A 1M resistor gives a trigger voltage of about 1kV.

Jrz126

Wed Jan 10 2007, 01:11PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

Well I tried going alittle higher on the charging voltage and I was able to gain some arc length (up to ~5" now). I also started causing some interference on the TV too.

Should I try to switch over to a 1-turn primary so I can get my tank cap back up?

EDIT:::

I decided to switch to a 1 turn primary and I was able to get it to notch much better than before.
http://www.personal.psu.edu/jrz126/images/oltc/single_turn_tuning_4.JPG
I am able to get ~1' arcs now with a DC bus voltage of about 180V (~400V after the choke) and 60 bps.

Steve Ward mentioned that many people have blown their garage door openers. Does this happen from a direct strike to the opener itself? or a strike to anything the opener is connected to? What should I do to make sure I dont kill my opener?

Woot Fed-ex just dropped off my fiber optic supplies. Time to get back to work.

Jrz126

Fri Jan 12 2007, 02:28PM

Registered Member #242 Joined: Thu Feb 23 2006, 11:37PM
Location: Erie PA
Posts: 210

Time for some pictures:
Me and the coil (I'm still trying to come up with a clever name for it)

The cap bank and rogowski coil. I'm gonna try to minimize the stray inductance sometime soon

The work area where I did the low voltage testing

My view of the coil. (yes, it is quite scary)

Sparks!

Still no breakout to air, I'm trying to debug the crowbar circuit. It trips every time I draw an arc. I think I just need some sheilding. (scared the crap out of me when it tripped the first time, not used to everything just going silent all of a sudden)
The strike rails are about 90% complete. Just need to add another half turn and connect them to ground.
Once I get those 2 things fixed, I should be able to give it some power. If I'm not too scared