Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
GeordieBoy, Sun Jun 14 2009, 12:01AM

Hi Finn,

Looks good from the pictures. How much propagation delay do you think the predicter is compensating for in the whole feedback loop? In other words, how much "timing advance" is the predicting part of the circuit contributing?

It is always good news if the IGBT voltage waveforms look clean when running under power. The absense of overshoot and ringing must mean that the switching transitions are occuring somewhere near the current zero-crossings. Clean voltage waveforms also indicate that there is no significant shoot-through.

It will be interesting to see how it performs in continuous runs, where the spark channels are given more time to grow and wander around!

BTW. What is the DC bus supply voltage to your "intermediate H-bridge" ? The idea here is to use a higher voltage than the IGBT gates require, and use a "step-down" GDT to boost the drive current at the IGBT gates. This technique is highly effective for driving multiple MOS devices in parallel where the total input capacitance is large. It also minimises the detrimental effect of any remaining leakage inductance when a GDT. This is because the leakage inductance from the driver side (primary) is reduced by the square of the turns-ratio when it actually appears at the IGBTs on the secondary windings!

Nice work, and will be interested to see how it develops...

-Richie,

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Sun Jun 14 2009, 03:01AM

We did not measure it, but since it used all of the available capacity there was, I think about 2µS!.
This we did measure, and even at low voltage, where there usually is a lot of spiking on the voltage waveforms, there was nil.
-That is: The first transition spikes a little bit, but the others are clean.
Having to leave the machine without running it flat out was very hard, but time did not allow more experiments. I'm sure Daniel will do this tomorrow.
This intermediate driver does not take advantage of any of these techniques. it is just a booster at 30 or so voltage.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
hvguy, Sun Jun 14 2009, 07:44AM

Congratulations on testing Finn! Your project is showing very promising results; I am going to have to reconsider using this approach in future systems. Can't wait to see a full power video!

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Steve Conner, Sun Jun 14 2009, 11:51AM

Looks great Finn! =) I'd also love to see footage of a continuous run at lower break rate and higher bang energy. Get those peak junction temperatures to the redline

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Sun Jun 14 2009, 08:59PM

All,

Thanks for your kind words.
I was away on family get together, but Daniel recorded this video. 380V or so in, 1000A max primary current, 80BPS and a bit more when it hits the lighting armature, for a full 6 feet discharge.



Everything is running cool, so looking forward to winding current transformers with 0.002V/A as the ones we are using now are hitting the input voltage limit very soon.
And voltage doubler, of course.

Hope this is something near what you were looking for, although no junctions got near the red line.

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
hvguy, Mon Jun 15 2009, 02:14AM

Yep, looks like a DRSSTC

You'll have no problem breaking the 3m mark with a doubler... Will you be using a different pri/sec or will you be setting this one on fire?

The other thread covered just about everything but it would be nice to see a complete schematic set; have one yet, Finn?

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Mon Jun 15 2009, 06:49AM

Aron,

Cannot say about first q, it is Daniels call, really.

I`ve posted sch and brd files down in projects.

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Robert2, Mon Jun 15 2009, 07:19AM

You can give pcb or page with this drsstc ?

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Mon Jun 15 2009, 08:41AM

Robert,

Files are down in projects, but please understand that this design primarily addresses problems encountered with large coils, which has long delays, due to powerfull, but slow switching devices.
Be also aware that this project has not yet been described to the point where it is ascessible to the first time builder.


Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dalus, Mon Jun 15 2009, 10:08PM

Starting to look good Nice to see you're able to run with the electronics in the open without shielding.

I've tried a similar gate driver on my coil some time ago without your predictor schematic. My gate driver was just a halfbridge of irfp450's running on 350Vdc with a step down gdt on each IGBT. This did not change the currents or the performance of the coil when compared to high side drive. So it's quite a nice solution when you need to drive multiple brick IGBT's. (btw I tested it on 2 CM300 igbt's) And I plan to use it on a dual H-bridge with 4 cm-300 modules.

Anyway keep up the good work, I'm still considering to try it on my coil to see how it affect performance. Though that will be in the future.

Still very curious to see what you guys and your coil can do

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Tue Jun 16 2009, 11:44PM

Today we (Daniel and I) worked till the blod was squirting from our nails, to finish a new primary coil, a bigger topload, and a voltage doubler too.
Goal: To make as much spark as possible.
We did not have time to reconfigure the current transformers, so we were limited to 1400Apri. which corresponds to 7V on pin 3 of IC8, the AD790 down in the current controll deck. With 1:100 current transformers, that is.....
It's as easy as that: just dial the right voltage, and we have predictable over current protection. No experimenting, here.
We also did not have time to make snubbers, but then did not seem to need them: We were running 800V into the bridge, switching 0 current, but peaking 1400A and making 2.5 meters of spark. That's 8feet4 to you yanks.
It was a bit windy, and we had only 10A 240V so mainly bursts.
The primary capacitor was 0.75uF and Fres 32kHz.
The "predikter" was out in the open. Not even a ground connection has been routed out to it, it is just floating out there.
I think Daniel is making a small video from this evening right now.

Enjoy


Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Coronafix, Tue Jun 16 2009, 11:54PM

Well, well. Bloody nice work both of you!!

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dalus, Wed Jun 17 2009, 09:00AM

Nice results guys, glad that your r&d led to something useful to the whole community. Think I'll redesign my controller to incorporate your predikter after all. I'll keep you updated on how it works out for me. Will take a long time to complete though since I don't have money for a new MMC atm though I'm working on that problem literally.

Anyway still looking forward to see how it preforms on a 3~ 32A line

btw Finn you where the one that could buy those brass bolts? If so would like to buy some from you for my coil.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Wed Jun 17 2009, 06:45PM


Dalus,

Thanks for youir kind comments.

About brass bolts. I am not against going a bit out of my way to help a fellow coiler in a tight spot, getting parts. Being situated in the old "east" may be such a tight spot.
But in the Netherlands? I would assume that the local Bauhaus has them on a rack. Have you checked?

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Linas, Wed Jun 17 2009, 08:37PM

Have you oscillation start in your circuit ? because i have some problems with my driver startup, need some frequency to start oscillate, and just then feedback catch the frequency of primary.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Wed Jun 17 2009, 08:55PM


No. The bridge starts to conduct when the interuptor is *on*, then switches at the zero crossings untill the interuptor is off, and then stops at the next following zero..

Sounds like you need to reverse the current transformer.

You are running standard Ward controller?

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
GeordieBoy, Wed Jun 17 2009, 09:31PM

Finn, that video is awesome! Particularly the longer bursts at the end where you vary the break rate. It's good that you have found the phase-lead network to improve operation. The sparks look really fat and bright too. What was the burst length set to during your video?

-Richie,

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dalus, Wed Jun 17 2009, 09:42PM

No haven't checked them yet. I'll sure do. I've searched quite a while myself but couldn't find a company that offers them. Thanks for the tip

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Thu Jun 18 2009, 05:08AM


Thanks!

Yes we were quite exited, also since my own spark length record was set 10 years ago, 2.4 meters on the prototype of the old Museum coil.

We were at times using as high as 300µS with good results, which surprised me, since 200µS usually seems to be the prime duration.

Daniels motor is now going to pass 3 months testing, as it is on display (in a proper faraday cage, for some children fare at the university). Great chance for some long time tests, although at reduced power.


Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dr. Drone, Fri Jun 19 2009, 04:39AM

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
teravolt, Fri Jun 19 2009, 07:15PM

Hi Finn exlant work it is enspiring, I have orcad lite and I am not able to open the schematic portion. was it created with a more current version of orcad. thanks

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Fri Jun 19 2009, 08:01PM

Nope, It is eagle.
Didn´t I write that clearly right next to the download link?????

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Bennem, Sat Jun 20 2009, 05:28AM

Excellent work Finn!!
and thanks for posting and documenting all the stages of your project!
this is truely a great thread with some much needed additional DRSSTC info
into why and how delays can cause 'spike' problems in high powered coils.
Your 'Predikter' circuit seems to over come this.
once again.....well done Finn and Daniel too!

Mel

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Sat Jun 20 2009, 08:57AM

Thanks, Mel.

I learned this approach from Terry Fritz. He said it was the easiest way to keep records of the progress of a project.
Other good reasons are the valuable, and at times indispensible, input from other coilers, as well as being part of a big rush towards the holy grail of solid state tesla coiling: zero current switching.

We will meet in Cambridge this year, I think?

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Bennem, Sat Jun 20 2009, 02:18PM


Hopefully i'll be there again this year.
i still have some great footage of your Midi playing DRSSTC from Cambridge,
it took weeks to get 'popcorn' out of my head!......lol

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Sun Jun 21 2009, 08:29PM

All,

I took a stab at writing a "paper".



Enjoy

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Linas, Sun Jun 21 2009, 08:36PM

link is broken...

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dalus, Sun Jun 21 2009, 09:46PM

Nicely documented. Great to learn from and maybe even find a way to improve on. Think it will be sure some variation of your driver will end up driving my quad CM-300 coil. So yes you'll have a coil to compete with in the EU

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Terry Fritz, Thu Jun 25 2009, 03:31AM

A very important solution to an age old problem!!!!!!!

Woke me up it did!!!

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Daniel Uhrenholt, Fri Jul 03 2009, 10:50AM

Hi All,

It's been a while since my last post, I have been stressed at work, as I now design mechanical parts for robots and satellites at Aalborg University.

But I am now in to coiling again:-)

The DRSSTC I made is at a children's fare at the University now, and it has passed some EMI tests today, with no comments at all! So I´m really happy about the results, as I used over two months building the darn cage and coil:-)

Here is a video of the coil

Cheers, Daniel Uhrenholt

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Paul Benham, Tue Jul 14 2009, 12:41PM

Hi Finn and Daniel,

I like your predikter circuit. It looks like a very sound design, and you have great results!

You could take the output of your precision rectifier and feed it into an RMS-DC converter with long time constant and then have a measure of the RMS current in your primary which could be interesting.

One thing I was not sure about was running the gate drivers from a signal with different timings for each driver. Do these not drive a gate drive transformer? I would have thought that you would want them both to change at the same time if driving a GDT. Or is the additional delay in the dead time circuit to prevent shoot through or something? I think I must be missing something.

Cheers,

Paul.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Tue Jul 14 2009, 02:05PM


Paul,

Thanks for your kind comments.
About the RMS circuit: Yes, sounds interesting, and would be nice with another meter on the front of the controll box.
The deadtime does not work with the gate transformers, as you noted. Deadtime was a leftover from the revision where optocouplers were used, but they were postponed for later refinement.

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dzejwor, Sun Mar 21 2010, 10:17PM

Hi Finn and Daniel,

I woud like to pass a question about your drive. I recently build my first DRSSTC using clasical Steve's driver. I have a problem with IGBT failures. I think that the reason is ringing and spikes on IGBT. I want to try use predikter to solve this problem. Daniel hanged supply voltage for UCC drivers from 5 to 15V and I want to do the same thing in my predikter but I would like to know from where did you get 15V ? I think it is possible to use 15V from supply for analog section (AD790 and LM7171) but I fear if this does not adversely affect the work of this part of the circuit. Resonant frequency of my secondary is about 60kHz power consumption is about 5kW. I need to make accurate measurement but it is not at this moment the most important.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
mmt, Tue Mar 23 2010, 09:12PM

plz... the link is broken, cant wait to see it

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Tue Mar 23 2010, 10:15PM


I will try to locate .BRD and :SCH files for re-upload.
You know how it is: new PC's, not too well backed up structure, but it is there somewhere, if only I could remember...

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
William Monk, Tue Mar 23 2010, 11:09PM

Excellent work! I would love to re-create this.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dzejwor, Wed Mar 24 2010, 12:38PM

Changing the PCB is not a problem for me. I have the original designs of predikter from "Thumper" thread (btw. tkis is the most amayzing tesla project I've ever seen ). I just know if what I want to do: get 15V for UCC drivers from AD790 supply will not cause any undesirable effects as the transfer of interference by the power rails or other bad things that could cause incorrect operation of sensitive comparators. I'm not so good in designing analog circuits.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Wed Mar 24 2010, 04:21PM


Well that's easy, then.
Feel free to tap into the 15V supply, to feed the UCC drivers.
Remember to add generous decoupling next to UCC's.

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dzejwor, Sat Mar 27 2010, 01:27PM

I changed the design today. UCC now are powered from 15V. There was one new small track on the top layer. I asked for the review and any comments if I did something wrong. I would like to order board for my coil in PCB company - I tried to eth them at home but I have a big problem with drilling small pads. Project is here

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dzejwor, Fri Apr 30 2010, 09:09PM

I have finished predikter pcb now and if I connect gdt into output of driver I have totlay bad signal on it - high frequency osclilations and generaly signal is not a sqoare wave. When I put 15ohm resistor in series with GDT primary everything looking proper. I speaking about small intermediate bridge GDT. I didnt connect IGBT GDT yet. Without intermediate GDT everything is good.

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Finn Hammer, Sat May 01 2010, 03:32PM


I know what you are talking about.
The predikter was designed to drive the bridge via optocoupled gatedrivers, and in that configuration this problem did not arise. When i abandoned the gate driver boards, and went to gate drive transformers, I got the same problem.

The problem arises from the fact, that the enable pin has a propagation delay of 60nS, which allows the gate signal to go high according to the feedback signal, before it is pulled low by the enable pin at the end of each burst. This seems to inject some charge in the transformer primary, which then gets trapped and has no other way but to oscillate.
Putting a resistor in the circuit, as I also did, seems to solve the problem, although I agree with you that it looks ugly, and far from elegant.
I have since forgotten about it, and I intend to resurrect my driver boards anyway, so....
It should be trivial to solve this problem, but some proficiency with logic chips is needed. You could borrow the logic from Steve Ward's current design, it is a better design than the one I used.
We had the same problem on Daniel's exhibition coil, however, it did perform excellently, so we just forgot about it.
I welcome an easy solution from anyone, because I am not so creative today as I was in those magical momths last year.

Cheers, Finn Hammer

Re: Hammertone "Predikter", the DRSSTC driver for the rest of the century.
Dzejwor, Mon May 03 2010, 08:43PM

After several hours of fighting I have something what looks like proper signals but stil have a lot of problems with adjusting that driver. Can you give me some advice how to proper adjust this ? I know how they should look correct waveforms on load But I dont know how to adjust everything and I have a question about a few issues which I do not understand first what is the potentiometer marked on your schematic as R1 and how to proper adjust this. I don't use deadtime so I set the R7 and R8 to minimal resistance - no delays here. R12 I set to proper rectified sine. R13 sets treshold of OCD. Now I ned to set R5 and R6 delays to arhive proper ZCS switching. Another thing which I do not understand the signal at the igbt gate. looks like at this picture below. I dont know whai is it I didnt se nothing like this before the only idea of what I have is that something goes through gate to collector capacity in the IGBTs when the primary current ringdown after the bridge is stopped by the interrupter/OCD. Next question what I have is why you didn't use a diodes with gate resistors. At this moment I have diodes in parallel with the gate resistors. Perhaps I should to try without them.



Edit:
I removed the diodes and it seems to me that the output voltage from the bridge looks better but now I have quite a different signal to the gate and I still have these strange oscilations after each cycle. What is the proper gate waveform without diodes - I need to adjust gate resistors.