Class E/TVS coil upgrade / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Class E/TVS coil upgrade

1 2 next

Move Thread

LAN_403

Mates

Thu Mar 13 2008, 12:01AM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Hi guys,
my old GhettoI (the project discussed here

) got some upgrade. It has a new primary and I'm also trying a new much bigger top load (runs at 150KHz). It still needs some tuning. It sucks almost 2KW from the wall and it can run 2-3minutes with the fan on. I think it is not bad for BUP314 (two in parallel)

. The temp of all the heat sinks is not critical after two min run, but I feel it can be still better. My problem is I have only one channel scope so proper classE tuning is a problem for me. On the other hand, because of the design, I don't need to care about the fall time of the transistor because the caps are discharged long before the transistor is closing, this allows me to keep the driving pulses very short (1us) which makes the classE tuning relatively easy.

I managed some new nice pictures and a movie so enjoy...

Comments and questions welcomed...

Movie link:

Logan Kennedy

Thu Mar 13 2008, 12:54AM

Registered Member #1103 Joined: Mon Nov 05 2007, 06:02PM
Location: Houston
Posts: 80

Cool!

I've been trying to make the move to E class, but I haven't been able to succeed thus far. I'm beginning to wonder if my o-scope isn't calibrated properly.

uzzors2k

Thu Mar 13 2008, 03:06PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Nice pictures as usual!

No offence but for 2kW you aren't getting very large streamers, are you sure the TVS diodes aren't burning most of that power? If so you should consider fixing your circuit or tuning. When tuned for proper class E, 3x Vin or more can be expected which is within the safe range of your two BUP314s. I would remove all the TVS diodes which are unnecessary for class E. As long as they're needed you're not switching in class E, but class TVS.

Mates wrote ...

I don't need to care about the fall time of the transistor because the caps are discharged long before the transistor is closing, this allows me to keep the driving pulses very short (1us) which makes the classE tuning relatively easy.

Which caps, the class E tuning ones? Their voltage is supposed to just touch zero just as your switch turns on, not before and not after. (Ideally) That's all class E tuning is.

Mates

Thu Mar 13 2008, 04:04PM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Uzzors wrote ...

Nice pictures as usual!

No offence but for 2kW you aren't getting very large streamers, are you sure the TVS diodes aren't burning most of that power? If so you should consider fixing your circuit or tuning. When tuned for proper class E, 3x Vin or more can be expected which is within the safe range of your two BUP314s. I would remove all the TVS diodes which are unnecessary for class E. As long as they're needed you're not switching in class E, but class TVS.

Yes you right. Those streamers are not very long (max 35cm) - but, it's a new design so it still needs a lot of tuning, mainly matching the resonant cap with the coil is a tricky part (this time my primary is not naked, so I have to try different capacitances which makes the tuning rather difficult). BTW: In case I run only 600W the sparks are almost the same size, but they are not as hot.

Uzzors wrote ...

Which caps, the class E tuning ones? They're voltage is supposed to just touch zero just as your switch turns on, not before and not after. (Ideally) That's all class E tuning is.

That's my mass in explaining things - sorry. So let's make things hopefuly more clear:

The old scheme is here:

The cap I mean is the 300uF motor foil cap used as a power supply. This cap is discharged into the cap/coil resonant circuit after the transistor gets open. This discharge pulse longs less than 100ns. That's why I do not care about the fall time of my transitor and the driving pulse last only 1us. While the transistor is closed my 300uF power source is being loaded from the wall and the resonant circuit makes an AC wave response at 150 KHz (or 180 in my previous setup). It's quite easy to place the 1us pulses between the AC pulses generated by the coil/cap, because half of the wave longs 13us. But, there two important issues here. In the resonant circuit made of the coil primary and resonant cap (aprox. 100nF 5000V) is induced quite a high voltage which once reach a 900V is clamped by the TVS. So the TVS are absolutely necessary and I will never remove them. But would be nice to keep them cold while the coil is running the question is whether itâ€™s even doable? I hope that once I achieve ideal resonance and coupling there will be voltage induction lower than 900V...

uzzors2k

Thu Mar 13 2008, 07:18PM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

I see. Using a larger resonant capacitor and smaller primary inductance will drop the rise. You should also find some TVS diodes with a clamping voltage closer to 1200V so you don't clip the resonant rise at such a low voltage. Even 1kV would allow for most of the rise if not all of it, although you can probably push the rise under 900V if you try hard enough.

Good luck with the tuning!

Mates

Thu Mar 13 2008, 09:02PM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

I decreased the nr. of primary turns (6), and also played with the resonant cap size. I also decided to go back for my old top load - it looks much better.
So, the performance (spark size and heating) seems much better now...

Mates

Thu Mar 13 2008, 10:47PM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

(Sorry for double post - admin please join with the previous answer...)

I connected my gate to the probe and left the ground open. I can see the driving pulses and also signal from the running coil (powered at 12V AC in this case). It's a try to show the classE performance using only one channel o-scope

Here are the driving pulses with grounded probe connection

Here the gound of the probe is left unconnected. You can see on the picture that the AC sinus generated by the running coil peaks right in the midle between the driving pulses (negative part of the sinus is not visible because of the driving pulses)...

And here is one more picture after little bit of additional tuning

Mates

Sat Mar 15 2008, 12:28AM

Registered Member #1025 Joined: Sun Sept 23 2007, 07:53PM
Location: Czech Rep.
Posts: 566

Iâ€™m little bit surprise that my coil attracts so little of interest

. It has so many advantages to compare to standard half or full bridge designs:
1) Relatively slow (thus cheap) IGBT transistors like BUP314 can be used for freq. up to 250KHz (maybe more)
2) Direct 220V AC powered (no transformer needed)
3) Power regulation (not continual but step like â€“ by adding or removing caps)
4) No antenna or current transformer feedback needed!
5) Stable operation resistant to capacitance environmental changes
6) No dead transistors!
7) Itâ€™s a single switch (does not need special demands on driving circuit) and allowing easy transistor paralleling
8) Even totally uneducated electronic beginner (like me) can manage to built it
9) Spark size can easily compete with the best full bridge designs Iâ€™ve seen on the net

It would be so nice to get at least somebody on my side and develop this design a bit. Iâ€™m ready to help and discuss thingsâ€¦

BTW: After last tuning I achieved cold TVS â€“ finally! Here are the last pictures of sparks and a movie â€“ hopefully a little bit of motivation for followers.

movie:

Cheers Mates

uzzors2k

Sat Mar 15 2008, 09:10AM

Registered Member #95 Joined: Thu Feb 09 2006, 04:57PM
Location: Norway
Posts: 1308

Looks like you've tuned it perfectly! The reason I haven't tried an offline class E coil is that I have piles of IRFP450s which are only 500V devices. If I'm to run anything offline it has to be via a bridge.

vasil

Sat Mar 15 2008, 12:08PM

Registered Member #229 Joined: Tue Feb 21 2006, 07:33PM
Location: Romania
Posts: 506

...maibe because it has a precedent (simple enough to build and good performance...the all known Royer SSTC). Even it has not an IC (as TL494) involved.

PROJECT HERE

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob