Quasar60 DRSSTC build using universal driver/interupter

[RoadReaper]

Hello all,  I have spent the last four years working on a pair of DRSSTCs in my spare time, I don't get much spare time. The plans I purchased from Amazing1.com for there Quasar 60 coil seemed to have a lot of discrepancies when it came to the driver/controller, bill of materials- schematic- picture of finished boards didn't match up.  I am not an electrical engineer so I am probably a little over my head.  I got little help from the company that sold me the plans so I looked for a universal controller. I purchased one i found online (Tesla coil DRSSTC SKP jump pulse driver new leopard arc extinguishing controller) just to see if I could get this up and going for now till I get the the one in the plans figured out. plugged every thing in yesterday and nothing happened.  No smoke came out, but noting to indicate anything was working but the capacitor that the power supply charges on the full bridge did charge.   Am looking to see if anyone has any experience with this coil build or the universal controller and could tell me if the two are just incompatible or where I should start trouble shooting.  I have built spark gap coils but this is my first DRSSTC.  .  I have my build on a face book page but am unclear if I am allowed to post links to other sites here.  Any help in the right direction to of where I need to start testing would be greatly appreciated.

Thanks
Chris

[Mads Barnkob]

Hi Chris

I do not have any personal experience with the amazing1 plans, as I find the open designs from Steve Ward and the further developments of these drivers are top quality.

Supply us with as much information as possible, links to external sites are fine, as long as its related to your post.

No smoke/explosions is a good start, let us work on getting you some sparks, the coil in itself does look good!

[dj.cosmo.esq]

Hi Chris,

I am no electrical engineer myself and I've built and bought my driver's. I have 3 of the drivers that you have so I'll give you the good and bad I've found. If you are using the original toslink cable that came with the package get a new one. I found that I couldn't get the board to trigger because the cable had horrible refractive properties. My 1st driver would not oscillate! So I built a Kickstarter circuit to make it work. Going against  good business practice I bought another one. Everything worked even the feedback and like 5-10v I heard the hummmm! So my first board was a dud but because it kinda worked I bought a another. I would have never known if I never got the second! I've run the driver on mini bricks to cm800s. Right now I'm running a full bridge of cm400s and I'm freaking loving it! Your setup looks great and ready for action. I'm completely up to date on how it's supposed to run so if you have any questions ask away. I've put many many hours of debugging and use on that board.

Courtney

[Teravolt]

where did you get the skip jump driver, is it cinese. can you show us a picture. dose information ulimited offer there own pcb for the plans that you recieved for the driver? there are a lot if circuits and projects hear.

[RoadReaper]

Hello all,  Sorry about taking so ling to reply, my busy time of year at work. I am replying to all responses at once sins this is the first time I have used a forum and not familiar with highlighting and responding to each response.

 I have other fiber optic cable that I bought off of DigiKey, [attachment-1] but it has no connector to fit the universal jump driver and controller,  Where did you find one to fit it?  I got this driver and interrupter off of ebay,  they seem to be all over the internet and lots of people seem to make them work but I have been unable to find any sort of instructions for it.  $30 usd from China,  get what you pay for.  Here is the link to my Road Reaper Performance page on facebook where you can see my build, https://www.facebook.com/Road-Reaper-Performance-320646418046560/ . here is a pic of the full bridge with tank circuit schematic [attachment  the original controller schematic and picture of their finished board [attachment-3] , [attachment-4] , and bill of materials.  [ attachment-5]  [attachment-6]  [attachment-7] ,  they said they had a printed board for an easier build,  dated 2013 but as of 2019 they still did not had one available to purchase.  I was running in to trouble starting with resister 5, stopped matching the BOM and finished board.  other parts were not matching up,  or completely missing form the BOM.  When I messaged them they just responded "looking in to it,  go by the schematic as far as they could tell." Went with this build because it was plans for a coil that a company sells,  should be accurate right?  Guess they really would rather me pay $5500 for one they built.  One change I did make is to the gate drivers input and output when going to the universal controller is their board had both outputs on one jack and return sharing a common pin on another jack. for the universal board I put the out on the center pin on one jack and the return on the outer pin of the jack for each gate driver.  May have been a mistake but made more sense to me,  but with no instructions for the universal driver I was just kinda flying by the seat of my pants.   I would not be opposed to purchasing a cable that fits the universal controller and purchasing a couple more of the controllers to get a good board,  they are cheep enough.  would like to get both coils up and going to display at our local Maker Space relaunch.  We are rebuilding after a fire took out part of our building a few weeks ago.  I would really like to get the original boards figured out as I have already purchased all of the components for them.  I have also purchased some unpopulated boards from Loneoceans Laboratories,  they were based on Steve Wards design,  I though I would populate them with the components that I had already purchased for the original boards. At about $9 a board I thought a good deal but they are way more compact than I expected and my components will not work on them. [attachment-8],  One of Loneoceans Laboratories Microwave oven transformer rotary spark gap coils that made 9' arcs is kinda what made we want to go beyond my simple "Six Pack Tesla Coil".

Well I hope this post makes sense to you, and all of the pictures come out in the right place.  Like I said this is the first time I have used a forum.  If I am doing it wrong please let me know.

[RoadReaper]

https://www.facebook.com/320646418046560/videos/1457541547967367/
 
A link to a video of my latest attempt to fire off my coil.  Still no function.  I have boosted the power to the driver from 12 VAC to 24VDC,  Installed a better optical cable but still nothing.  I have an ossilscope that I could connect to the driver outputs to see what is going on but would need to know what I am looking for.  I set my over current protection to one end of the scale to the other and in the middle but did not change anyting.

[davekni]

If I understand correctly from pictures, you have a UD2.7 ECB that isn't stuffed and a breadboard of a driver that is built and being used.  Correct?

Great to hear that you have a scope.  That's the critical tool, and should eventually lead to finding and fixing issues.

First thing is to power the driver only w/o H-Bridge power.  Set the controller for short (~50us) enable pulses at some low rep rate (1-10Hz).  Scope the output of the fiber receiver on the driver with one probe and trigger on that channel.  Scope a gate-drive signal with the other probe.  There should be an initial step on gate-drive at the lead edge of each enable pulse.

If no trigger (no signal at output of fiber receiver), then the issue is with that receiver or optical cable or controller.  If there's a signal there but not at the gate, then trace back through the driver circuit to see where the signal ends.  Feel free to post scope images, being clear where in your circuit probes are connected.

[RoadReaper]

Hello all,  I am not using the bread board driver, I am using a universal driver that seems to be based off of Steve Wards design,  The bread board one is only partly completed,  was having problems with the plans not matching the schematics so I went with a inexpensive universal just for lack of funds at the moment and to try and push this project forward.  Although I think I stumbled across something while on Loneoceans labs site looking over some schematics and diagnostic information they have for there UD2.7C board https://www.loneoceans.com/labs/ud27/#schematic
The two biggest things I noticed is that his test for GDT out seems to be four outputs(correct me if I am wrong), I ask this because I noticed with his scope test he said to connect signal generator set to 100K Hz  to the feedback in on the board and test between all four pins on the GDT to ground on the board.  On the original controller for this project the out put for the drivers were in pairs one pair to the center pin of a jack the other to the outer pin of same jack,  and the returns all four to one pin of a second jack.
sounded a little strange to me.  seemed like it should be one pair out on one pin and the returns to the pair to the other pin of the same jack using two jacks for the GDT.  So I rewired it this way for the universal driver.  Thinking that might be my first problem.  If anyone can help me with this problem would be great.  Second on the UD2.7C I looks like he has his feedback and OCP current transformer with the the lead to the primary coil going through both of them.  The controller in my original plans has the OCP current sensing transformer utilizing the a lead to the primary coil,  The feedback transformer using the ground from the secondary coil.  Here is a link to a short video I posted on my facebook page of the diagnostics and scope readings,  If someone could take a look at it and tell me if think my thoughts of where I think my problems are sound correct.  My scope is old and not working well but I think it gave me a good enough  reading to point me in the right direction. https://www.facebook.com/320646418046560/videos/668932313711068/
 Yes,  I know I need to get a scope from this century.
I want to thank everyone for the help I have been getting on this form.
Chris.

[davekni]

Yes, UD2.7 is designed to feed 2 gate transformers.  I think that's for large IGBT bricks where driving all four gates from a single output would be too much load on the one output.  I think many UD2.7 users don't use or include parts for the second output.

A few still pictures would be more useful for studying in detail.  I can't watch shaky video for long.

Your 465 scope should be usable, although single-shot capture of digital scopes does have some advantages.  There's clearly some issue with gate-drive output showing up as the bouncy lower-level of the square-wave.  (Since you have a signal generator, that's a great way for initial debug.  Stay with that.)  What enable-pulse were you using during that 100kHz test?  Was your scope probe ground connected to the driver board ground?  Definitely do not connect ground to one of the gate-drive outputs.  Test initially without the GDT connected.

I'm guessing your driver has just a single channel, with two complimentary outputs.  Not clear on the connector issues, but find those two signals and scope them with two probes to see that they are indeed complimentary.  Then wire all four GDT primary windings in parallel, four leads to one output and the other four leads to the other output.  You'll need to twist the GDT leads from GDT to driver and from GDT to each IGBT to minimize leakage inductance.

[RoadReaper]

Ok,  I have some new information and questions.  I think I had the GDT  miss wired, I had all sides of them connected to output none to the ground of the driver board.  I separated them and have two wires from each gate transformer primary going to an out put of each gdt out and the other side going to the ground of the board.  I do not have the ground of the board connected to the ground of the 110 input to the 24vdc transformer.  I am getting a signal at the gdt output of the board on all four pins, I have a signal generator connected to the feed back with 100khz square wave, and a peek voltage meter connected to gdt out and getting 22-23v.  I will upload pictures to my face book link since I have been having trouble getting more than on pic to attach here. Still I have no sign of getting sparks or any sound.  At this point I am still questioning my connections of the gate transformers and or the build of my bridge circuit.  I have step by step build of the bridge circuit on my facebook page under the singing Tesla Coil build( links to the videos are under some of my older post and you should be able to find my build from there,  As soon as I can get facebook to cooperate I will add direct links). and I will post a pic of the schematic that was used under my normal post.  Facebook is not cooperating with my computer(just tells me that I have too many redirects,  switching to old face book solved this for a bit but that expired on the 1st and now can't log in on my computer) and it wont let me add pics to photo albums from my phone, just new post to my normal feed.  The pics of the scope readings are on a cheep scope i ordered off of amazon,  since my 465 has been giving me an attitude the past several days. (not sure if it worth getting some  place to refurbish it or would be cheaper to go pc based scope, suggestions?),  As all ways I really appreciate all of the advice I have been getting on this site.  Thanks to everyone.

[davekni]

Sounds like your GDT was wired correctly initially.  All four primary windings should be in parallel, one side to each of the two GDT driver outputs, none to ground.  The GDT secondary windings need to be phase correctly, so that diagonally opposite IGBTs are on at the same time.

Concerning scopes, I have a Hantek DSO4102C.  It works fairly well.  For high-electric-field conditions such as with my Marx generator, I use it inside a 5-sided foil-lined box.  Otherwise it's processor gets confused.  Haven't checked if my DRSSTC makes enough field to be a problem, as I've left the scope in its foil box.  (My Marx generator makes much more high-frequency field than my DRSSTC does.)  What I find now is DSO4102S.  Looks the same.  About $300 USD:
https://www.ebay.com/itm/2-in1-2CH-70M-200MHz-Digital-Oscilloscope-1GSa-25M-Arbitrary-Waveform-Generator/273433967019?hash=item3fa9ef39ab:g:kMEAAOSwjkxbh77M

My nice scope is a TDS754C.  One of the most accurate 8-bit digitizers around IMHO - for things like intermodulation distortion, effective bits, etc.  I think this is true, but then my brother designed the digitizer chip, along with later ones in high-end Tektronix scopes.  TDS754 and related scopes are heavy compared to today's options, and lack modern connectivity.  Also have a THS720 prototype courtesy of my brother.  Great for its two isolated input channels and portability and battery power, but limited screen size and user-interface relegate it to only rare use.

Yes, I had trouble with more than one image initially too.  Read through the introductory "before you post" message on this forum.  You need to click a note at the bottom about "add more attachments" or something close to that to get a second attachment in the menu.  And again for each additional attachment.

[RoadReaper]

I think I see at least part of my problem,  am at work right now and can't post a picture of it at the moment.  If I am understanding the gate driver out puts correctly I have two outputs each with two terminals that are linked to provide more connection points.  I originally had the two primary leads to one GDT on one pin of the driver output and the other end of the primary leads to the same GDT on the second pin of the same driver output.   And I am seeing that I should have had the other end of the primary leads returning to a pin on the second driver output.  Is this correct?

[davekni]

I think you had it correct initially.  Ignore the second GDT output pair.  Parallel all four primary windings and connect the two ends to the two terminals of one output.  The second GDT output pair isn't needed unless driving large IGBT bricks.

[RoadReaper]

Here is the image from my scope, I am not sure if probing the same signal on two terminals or if it is two separate signals.  I don't have much practice using scopes, still learning.  This is the board I am using, the two terminals on the lower right with the yellow circles around them are the two I get the signals from.  The other terminals on the same connectors don't have a signal.

[davekni]

It's normal for only one side of GDT drive to have a signal when the driver's enable input is pulsing (from the interrupter), but CT feedback is inactive.  (If you have your signal generator running into the CT feedback input, then there is some issue with that CT path through the driver.)  Is the scope trace of a single output?  I recommend using just one of the two output connectors, with the one connector's two pins wiring to the two sides of the primary windings (all four in parallel).  (BTW, looks like your 10x scope probe needs to be adjusted.  Actual trace likely has flat tops.  Use a fine screw driver to tweak the compensation of the probe to get flat signals.  This will matter more later when checking signals that might have such a shape or the opposite, such as gate-drive after being loaded with GDT.)  I'm also presuming that the gate output traces are actually at 10V/div using a 10x probe, but the scope is still set for a 1x probe, so is displaying 1V/div.

860Hz is a high interrupter frequency for initial testing.  I'd suggest dialing your interrupter down to 1Hz, or at least to 10Hz, to make it less likely for problems to cause fried IGBTs.  I'd also reduce interrupter pulse width to 100us initially for the same reason.  Short infrequent enable pulses are much safer for initial testing.

[Netzpfuscher]

Ignore the second GDT output pair.  Parallel all four primary windings and connect the two ends to the two terminals of one output.  The second GDT output pair isn't needed unless driving large IGBT bricks.

This is a dangerous advice. With a CPLD pulse skipping driver it is likely that both GDTs are needed. I think in a overcurrent event only one GDT gets inhibited like on the UD3.


From the UD3:

The UD3 implements freewheeling via a dual GDT drive of the H-bridge.  When the current level exceeds the
threshold, one of the GDTs will be inhibited for for the next drive cycle.  Inhibiting 1 GDT, while driving the other,
puts the H-bridge into a mode where energy will be trapped in the primary circuit (rather than adding energy in, or
rectifying it back to the DC bus capacitor).  The gate-driver logic alternates which GDT port will be inhibited during
current limit so that the switch/diode losses can be spread evenly between half-bridges.  Consequently, half-
bridge operation, or operation from a single GDT requires special firmware (which does not yet exist) to work
properly, and is not ideal.

[RoadReaper]

Beginning to think my best bet is to dump my $30 driver/interrupter and order the parts from Mouser to populate the UD2.7 boards I have.

[davekni]

Netzpfuscher:  Good point about pulse-skipping.  If that mode is enabled, then the driver may require separate GDTs for each half-bridge.  I think the scoped gate-drive outputs are still consistent with what a UD2 driver would generate with enable pulses and no feedback.

The existing driver may work, and pulse-skipping can be useful.  However, a UD2.7 is easier to debug, as the schematic is known.

[RoadReaper]

I ordered parts off of Mouser to populate my UD2.7 board,  Most should be here Friday,  a couple were on back order.  Would like still to try to make use of the cheep boards I have purchased but want to move forward.  My concern is surface mount soldering,  have had little practice with it and these boards have a lot of surface mount components.

[Hydron]

Get a decent pair of tweezers, some flux and some good leaded solder.

Tin one pad first (a corner for a SO package, whichever is easiest ergonomically for 0805 etc), use tweezers and iron to melt the solder on the pad while placing and aligning the part, then finish other pads at your leisure (you may need to revisit the first pad again to touch it up later). Adding flux before putting the part down will help a lot, especially with clearing bridges etc and with smaller parts, but you can probably get away without it for the UD2.7 if you have good flux-core solder (still get some if you can though!).

Can take a little time to get the hang of it, especially if you get solder bridges etc (less likely with SO sized packages), but once you're up to speed I find it significantly faster than through-hole work.