Author Topic: First DRSSTC, No Sparks on Breakout  (Read 2051 times)

Offline acmq

  • High Voltage Enthusiast
  • *
  • Posts: 15
  • Karma: +2/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #20 on: August 16, 2020, 11:59:02 PM »
Some things that I notice:  8)
The top load is quite large. This reduces the output voltage.
The beats in the waveforms indicate maybe too high coupling. I would expect beats lasting for about 10-20 cycles.
I never used feedback in my drsstcs, and they work in the same way.  :-X

Online davekni

  • High Voltage Senior
  • *****
  • Posts: 689
  • Karma: +34/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #21 on: August 17, 2020, 12:39:39 AM »
Some things that I notice:  8)
The top load is quite large. This reduces the output voltage.
The beats in the waveforms indicate maybe too high coupling. I would expect beats lasting for about 10-20 cycles.
I never used feedback in my drsstcs, and they work in the same way.  :-X

The top-load might be a bit large for the available power, but I'd first try pushing the power up to its limit.  Larger top-loads generally help performance, unless voltage just can't get high enough to initiate a good arc.  The 52k secondary impedance is great.  A smaller top-load will increase frequency and impedance, the latter being why it might be better for low-power drive.

I think the rapid beat is due to primary-secondary frequency difference, not due to excessive coupling.  If excessive coupling, the amplitude modulation would be deeper.  That's the reason to tune primary somewhat closer to secondary frequency initially until sparks get large enough where more detuning becomes necessary.

Are your DRSSTCs really completely open-loop (driven from an independent frequency source)?  That makes it very tricky to keep zero-current bridge switching.  If you are using FETs instead of IGBTs, or are running lower frequency, perhaps you are getting away without needing zero-current switching (ZCS).
David Knierim

Offline acmq

  • High Voltage Enthusiast
  • *
  • Posts: 15
  • Karma: +2/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #22 on: August 17, 2020, 01:25:24 AM »

I think the rapid beat is due to primary-secondary frequency difference, not due to excessive coupling.  If excessive coupling, the amplitude modulation would be deeper.  That's the reason to tune primary somewhat closer to secondary frequency initially until sparks get large enough where more detuning becomes necessary.

Are your DRSSTCs really completely open-loop (driven from an independent frequency source)?  That makes it very tricky to keep zero-current bridge switching.  If you are using FETs instead of IGBTs, or are running lower frequency, perhaps you are getting away without needing zero-current switching (ZCS).

Beats are easily associated with coupling. Mistuning reduces beating, not increases it.
My drsstcs are totally open-loop. SGTCs are too and work well. Long bursts may cause some problem, but the bursts don't need to be long in the way I design the systems (I have abandoned experiments with drsstcs years ago, but didn't see anything new, except the interesting QCW coils, since then). See http://www.coe.ufrj.br/~acmq/tesla/drsstcexcitation.pdf
Another issue: The IGBTS used are 60A devices. With 1 uF primary capacitor and 170 kHz the primary current is maybe much higher. What are the currents observed?

Offline Yak

  • High Voltage Enthusiast
  • *
  • Posts: 11
  • Karma: +0/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #23 on: August 17, 2020, 12:21:12 PM »
I have separated the case from the primary coil by 12cm. This is as much as I could do in such a short time as I had a mount I had built previously, and put the primary on top of that. I did a high voltage test, at full power 240VAC, I ran the coil at around 100BPS with an on-time of around 220us. At these settings, one of the IGBTs shorted across all terminals. I did not manage to get a scope shot of the primary current. The arcs were still quite small, shooting out the breakout point at around 6cm long.

I replaced the shorted IGBT and did another test, this was at 200VAC, 120BPS with an on-time of around 110us, and no IGBT broke this time. I had a look at the primary current and it rang up, then remained at around 200A, before ringing down after the burst ended. The current was pretty consistent, it just went up, then the amplitude was constant 2V (200A) until the burst ended. The arcs were still quite small. This test lasted around 15s but the IGBTs got very hot. I have them mounted on the same heatsink with insulating pads and thermal paste. I think the insulating pads are preventing alot of the heat from transferring quickly to the large heatsink ive used. I might need to use seperate heatsinks so that I do not need to use insulating pads.

 Anyways, even with the high primary current of around 200A, the arcs were still quite small. I likely did not tune it properly. How will I know if it is tuned properly? will the primary current waveform look different when it is tuned?

I will do another test soon and get some primary + secondary current readings so that I can then calculate the voltage at the topload.

I do not have an isolation transformer, so I cannot measure the circuits while the variac is powering the bridge, as the scope ground is earth.

« Last Edit: August 17, 2020, 12:23:09 PM by Yak »

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1569
  • Karma: +28/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: First DRSSTC, No Sparks on Breakout
« Reply #24 on: August 17, 2020, 02:54:40 PM »
I have separated the case from the primary coil by 12cm. This is as much as I could do in such a short time as I had a mount I had built previously, and put the primary on top of that. I did a high voltage test, at full power 240VAC, I ran the coil at around 100BPS with an on-time of around 220us. At these settings, one of the IGBTs shorted across all terminals. I did not manage to get a scope shot of the primary current. The arcs were still quite small, shooting out the breakout point at around 6cm long.

I replaced the shorted IGBT and did another test, this was at 200VAC, 120BPS with an on-time of around 110us, and no IGBT broke this time. I had a look at the primary current and it rang up, then remained at around 200A, before ringing down after the burst ended. The current was pretty consistent, it just went up, then the amplitude was constant 2V (200A) until the burst ended. The arcs were still quite small. This test lasted around 15s but the IGBTs got very hot. I have them mounted on the same heatsink with insulating pads and thermal paste. I think the insulating pads are preventing alot of the heat from transferring quickly to the large heatsink ive used. I might need to use seperate heatsinks so that I do not need to use insulating pads.

 Anyways, even with the high primary current of around 200A, the arcs were still quite small. I likely did not tune it properly. How will I know if it is tuned properly? will the primary current waveform look different when it is tuned?

I will do another test soon and get some primary + secondary current readings so that I can then calculate the voltage at the topload.

I do not have an isolation transformer, so I cannot measure the circuits while the variac is powering the bridge, as the scope ground is earth.

It sounds like its not tuned properly, if you can get 200A and little to none sparks. If the energy is not transferred from primary LC to secondary LC circuit, it will ring up quickly. To tune it, calculate tuning point with JAVATC and detune the primary for 5-10% lower frequency.

What is your DC bus voltage at? My small DRSSTC is limiting itself to 280Apeak due to sagging DC bus voltage after a certain burst length.

Only the badly-out-of-tune should explain the small sparks and heating IGBTs.

Save up money for a differential probe, it will make your life with inverters so much easier :)

full power 240VAC, I ran the coil at around 100BPS with an on-time of around 220us

I guess you simply exceeded the thermal capabilities of these TO-247 IGBTs, do not test at full power before you have proper growing sparks from 30-100VAC input.
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics
https://www.youtube.com/KaizerPowerElectronicsDk60/join - Please consider supporting the forum, websites and youtube channel!

Offline AstRii

  • High Voltage Experimenter
  • **
  • Posts: 95
  • Karma: +2/-0
  • Czech Technical University in Prague
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #25 on: August 17, 2020, 04:08:43 PM »
You're very courageous when trying full mains voltage while the coil is not working correctly, but that often leads to failure.
I would say that your coil is simply out of tune, don't worry to detune the primary coil even to much lower frequencies than your secondary, i would try it all the way from 170kHz to 100kHz :)

Offline Uspring

  • High Voltage Technician
  • ***
  • Posts: 197
  • Karma: +16/-0
  • Physicist
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #26 on: August 17, 2020, 05:14:42 PM »
You should lengthen the breakout rod somewhat to at least 3 times its present dimension. That will reduce the breakout voltage and simplify tuning somewhat.

If you want to tune the coil at low power, you can tune for a max secondary current to primary current ratio. That ratio should be independent of input power. The max ratio is not exactly the best for high power operation with long arcs, since these arcs will detune the secondary tank by adding capacitance to it. At that point it is better to decrease primary resonance frequency a bit. But first start looking for the max ratio.

Online davekni

  • High Voltage Senior
  • *****
  • Posts: 689
  • Karma: +34/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #27 on: August 17, 2020, 08:38:35 PM »
Yes, differential probes are wonderful, but also not cheap.  Much can be done to get useful scope traces even without one.  To start, the current transformers are isolating.  A plot of primary and secondary current would be helpful at full power.  Second, since line frequency is so much below 160kHz, a simple C/R filter can pick out the signal of interest without grounding the scope probe.  For example, add a 100k resistor from scope probe tip to scope probe ground, and a 470pF capacitor from probe tip to half-bridge output.  The signal won't be as clean as with a differential probe, but is plenty usable to see how close you are to ZCS and if there are any unintended switching events.  This works even better if you have some small "Y" caps from line (or better VBus) to safety-ground (as many commercial supplies have for EMI reduction).  Then ground your scope probe to that safety-ground connection.

12cm is probably enough.

Were your IGBTs hot, but not the heatsink?  If just IGBTs, then, yes, the insulators are an issue.  Still, at your low 0.22% duty cycle, I wonder if there's other issues causing excess IGBT power dissipation.  That's the value of full-power scoping.  If switching is getting too far from zero-current or behaving erratically, or if the IGBTs are coming out of saturation, power goes way up.  15Vge may not be enough to push these parts to 200 or 250A.  Their peak current rating is 180A.  To get more usually requires pushing Vge higher.  (I run 18Vge, but most people here run 20-24Vge.)  Another possible full-power issue is switching transients causing Vce to spike above 600V.  The IGBT may handle some pulses into avalanche-breakdown, but that dissipates much heat, and sometimes causes immediate failure.  At 220VAC (310 VBus), I'm guessing Vce spikes aren't your particular issue.
David Knierim

Offline Yak

  • High Voltage Enthusiast
  • *
  • Posts: 11
  • Karma: +0/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #28 on: August 18, 2020, 08:08:33 AM »
Thank you all for the useful advice. I will start saving up soon for a differential probe, they are quite expensive but also seem very useful. I could use one for many different projects I work on in the future.

First I thought id show a picture of the setup, the bottom of the primary is 12cm above the top of the ATX housing.



You should lengthen the breakout rod somewhat to at least 3 times its present dimension. That will reduce the breakout voltage and simplify tuning somewhat.

If you want to tune the coil at low power, you can tune for a max secondary current to primary current ratio. That ratio should be independent of input power. The max ratio is not exactly the best for high power operation with long arcs, since these arcs will detune the secondary tank by adding capacitance to it. At that point it is better to decrease primary resonance frequency a bit. But first start looking for the max ratio.


I have also done this, The breakout point now extends 13cm from the top of the torroid.



I ran JavaTC with my coil parameters and used this to tune the coil. According to JavaTC, at exactly 6 turns of my primary coil, I will get a LC resonant frequency of 169.33kHz. With this in mind, I put the primary tap at 6 turns and did a low voltage test (30V dc). The interrupter settings I used was around 160BPS, 170uS on-time. At 30V input, I measured a peak of 44A primary current, and 0.46A secondary current. I could hear and see the arcs in a dark room. The arcs were very small, just some corona around 1mm long.

I then did another test at 55VAC, which is around 77V on the bus. I captured this waveform and have attached it below. Interrupter settings are the same at 160BPS and 170uS on-time. I noticed that the arc grew in size to a few cm long. The waveforms for this test were basically identical to the 30V test, just with the amplitudes much greater. In this test, I measured a primary peak current of 100A and a secondary current of 1.13A.

Yellow is Primary Current, Purple is secondary current. Keep in mind the CT ratios are different. Every volt is 1A secondary current, while every 10mV is 1A primary current. The shot has 500mV per division.



Were your IGBTs hot, but not the heatsink?  If just IGBTs, then, yes, the insulators are an issue.  Still, at your low 0.22% duty cycle, I wonder if there's other issues causing excess IGBT power dissipation.  That's the value of full-power scoping.  If switching is getting too far from zero-current or behaving erratically, or if the IGBTs are coming out of saturation, power goes way up.  15Vge may not be enough to push these parts to 200 or 250A.  Their peak current rating is 180A.  To get more usually requires pushing Vge higher.  (I run 18Vge, but most people here run 20-24Vge.)  Another possible full-power issue is switching transients causing Vce to spike above 600V.  The IGBT may handle some pulses into avalanche-breakdown, but that dissipates much heat, and sometimes causes immediate failure.  At 220VAC (310 VBus), I'm guessing Vce spikes aren't your particular issue.

The IGBTs were hot, and the heatsink was warming up. I know that there is heat transfer between the IGBTs and heatsink as the heatsink slowly warms up. In addition, it didnt take that long for the IGBTs to cool down. However, I do not think heat is transferring fast enough to the heatsink. I may need to consider a different setup to remove the heat as fast as possible.

I have another GDT that I wound that has 10 turns of network cable. The ratio I have made this is a 1:2:2. Meaning that I could increase the Gate-Emitter voltage while driving the IGBT. As my driver circuit is driven with 15V, using this GDT will drive the IGBTs as 30V. I am not sure if doing this will destroy the G-E junction as it is rated as 20V maximum in the datasheet. Alternatively, I could change the driver voltage from 15 to 12V and then use this GDT to drive the IGBTs as 24V instead of the 15V I am currently driving them with. What are your thoughts on this? In addition, wouldn't driving the IGBT above 20V damage the gate as it is stated that 20V is the maximum it can take in the datasheet.

Here is a photo of the GDT I wound:



Finally, if any of you would like to have a look at the JavaTC calculations of my coil, I have attached a txt file that can be used to load my coil into JavaTC.
* SavedCoil (4).txt
« Last Edit: August 18, 2020, 08:24:58 AM by Yak »

Offline johnf

  • High Voltage Technician
  • ***
  • Posts: 147
  • Karma: +5/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #29 on: August 18, 2020, 10:39:01 AM »
get rid of that foil you have put on the piece of wood.
doing what you have done has halved the spacing ie you are only 50-75mm away from a shorted turn this will de-tune your primary and compromise coupling to your secondary

still good progress

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1569
  • Karma: +28/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: First DRSSTC, No Sparks on Breakout
« Reply #30 on: August 18, 2020, 01:14:00 PM »
In addition, wouldn't driving the IGBT above 20V damage the gate as it is stated that 20V is the maximum it can take in the datasheet.

See http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/ for references.

Quote
Maximum Gate-to-Emitter Voltage (VGE)

The gate voltage is limited by the thickness and characteristics of the gate oxide layer. Though the gate dielectric rupture is typically around 80 volts, the user is normally limited to 20 or 30V to limit current under fault conditions and to ensure long term reliability.
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics
https://www.youtube.com/KaizerPowerElectronicsDk60/join - Please consider supporting the forum, websites and youtube channel!

Online davekni

  • High Voltage Senior
  • *****
  • Posts: 689
  • Karma: +34/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #31 on: August 19, 2020, 05:05:03 AM »
The aluminum ground perimeter on the wood piece is fine IF there is a break somewhere around the perimeter.  A couple mm is plenty, as voltage shouldn't be over ~500V for one turn.  (I'm presuming the foil has some connection to ground to protect electronics below from a streamer-strike.  Otherwise I don't know what the foil's purpose would be.)

Especially for low-power testing, I'd raise your primary frequency closer (but still below) secondary frequency.  Hopefully you will see more of the beat-frequency as energy transfers to the secondary and back, and get higher peak secondary current for a given bus voltage.  As streamers get longer, secondary resonant frequency drops.  That's the reason to have primary lower.  But closer is better until streamers get longer.

Even though the IGBTs may survive +-30V, I'd be concerned about the resulting 4x load on your UCC27423 driver chip.  I gather you are running the UCC27423 at its maximum 15V.  How large bypass capacitors are on the 15V supply?  UCC27423 will draw high currents during enable pulses, even higher as GDT transforms-up the voltage, by square of turns ratio.  Even though it's more tricky to construct, I'd suggest 1.5:1 for your GDT (+-22.5Vge).  I'd use two sets of twisted wires, 5 wires per set, to wind the GDT.  Of each twisted set, two wires are in series for the primary and three in series for the secondary to an IGBT gate.  Both primaries wire to the UCC27423.
David Knierim

Offline Uspring

  • High Voltage Technician
  • ***
  • Posts: 197
  • Karma: +16/-0
  • Physicist
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #32 on: August 19, 2020, 01:48:19 PM »
JavaTC provides impressively accurate figures for the tank quantities, but varying external conditions, like distance to the next wall, metal objects nearby that it does not know about, can lead to errors, so it might be worthwhile to try out different taps on your primary coil.

A look at your JavaTC file indicates a coupling of 0.134. A larger coupling is probably better, since it increases energy transfer from the primary to the secondary tank and also makes the coil less sensitive to detuning. I'd try lowering the secondary a bit with respect to the primary. Too much coupling can lead to racing sparks or flashovers, but presently your arcs don't seem big enough for that.

Your primary current seems to flatten out too soon. In principle, the primary current should increase as long as you supply energy to it. The flattening indicates, that you are loosing energy somewhere. Since you don't have big arcs, the energy is presumably lost in and/or

(i) too thin leads to your primary coil
(ii) excessive C-E voltage drop
(iii) bus capacitor ESR

I think (i) is the most likely candidate, but all possibilities can contribute and (ii) has particularly explosive consequences.


Offline Yak

  • High Voltage Enthusiast
  • *
  • Posts: 11
  • Karma: +0/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #33 on: August 19, 2020, 03:35:25 PM »
Thank you all for the helpful advice.

The aluminum ground perimeter on the wood piece is fine IF there is a break somewhere around the perimeter.  A couple mm is plenty, as voltage shouldn't be over ~500V for one turn.  (I'm presuming the foil has some connection to ground to protect electronics below from a streamer-strike.  Otherwise I don't know what the foil's purpose would be.)

Especially for low-power testing, I'd raise your primary frequency closer (but still below) secondary frequency.  Hopefully you will see more of the beat-frequency as energy transfers to the secondary and back, and get higher peak secondary current for a given bus voltage.  As streamers get longer, secondary resonant frequency drops.  That's the reason to have primary lower.  But closer is better until streamers get longer.

What exactly do you mean by beat-frequency waveforms? I have tried tuning the coil from 6.0-6.5 turns as from experimentation this is the best, and also according to JavaTC. From multiple points between 6.0 and 6.5, and the waveforms look pretty similar. It usually takes the length of the on-time 170uS to ring up to peak primary current, then it rings down.

Even though the IGBTs may survive +-30V, I'd be concerned about the resulting 4x load on your UCC27423 driver chip.  I gather you are running the UCC27423 at its maximum 15V.  How large bypass capacitors are on the 15V supply?  UCC27423 will draw high currents during enable pulses, even higher as GDT transforms-up the voltage, by square of turns ratio.  Even though it's more tricky to construct, I'd suggest 1.5:1 for your GDT (+-22.5Vge).  I'd use two sets of twisted wires, 5 wires per set, to wind the GDT.  Of each twisted set, two wires are in series for the primary and three in series for the secondary to an IGBT gate.  Both primaries wire to the UCC27423.

The aluminium was originally meant to be a ground, it was there on the mount previously for a SSTC I had made. I got rid of the aluminium foil as the electronics are housed in a grounded case. Later, I will add a strike rail above the primary coil so that any streamers do not hit the primary and destroy the IGBTs. I have been trying to tune my tesla coil but I am pretty sure all the stuff on my desk, metal shelf nearby, as well as the wall being nearby is throwing off my tuning. I have, however figured out that the best tuning point is between 6.0 and 6.5 turns. These points create the largest sparks at a lower voltage.

I apologise, but I also forgot to mention that I am not using a UCC27423 gate driving chip as used in the original driver 1.3b. This was because I could not find any locally or online for cheap. Instead, I had some UCC27322 chips which are also low-side drivers, but with a single output. These chips are 9A peak, just over double the 4A of the UCC27423. I am using two of the UCC27322 chips in a push-pull configuration for the same effect as the single UCC27423 chip used in the original driver. I have 47uF electrolytic on each chip and some ceramic capacitors. I made a new GDT and this GDT has a ratio of around 1:1.5. The GDT drives the IGBT gate at 22V peak. I have attached a waveform below.



The wave is looking less square than it did before. However, it still switched the IGBT and I could see primary current flowing on the scope.

JavaTC provides impressively accurate figures for the tank quantities, but varying external conditions, like distance to the next wall, metal objects nearby that it does not know about, can lead to errors, so it might be worthwhile to try out different taps on your primary coil.

A look at your JavaTC file indicates a coupling of 0.134. A larger coupling is probably better, since it increases energy transfer from the primary to the secondary tank and also makes the coil less sensitive to detuning. I'd try lowering the secondary a bit with respect to the primary. Too much coupling can lead to racing sparks or flashovers, but presently your arcs don't seem big enough for that.

Your primary current seems to flatten out too soon. In principle, the primary current should increase as long as you supply energy to it. The flattening indicates, that you are loosing energy somewhere. Since you don't have big arcs, the energy is presumably lost in and/or

(i) too thin leads to your primary coil
(ii) excessive C-E voltage drop
(iii) bus capacitor ESR

I think (i) is the most likely candidate, but all possibilities can contribute and (ii) has particularly explosive consequences.



I have lowered the secondary coil+topload assembly 2cm. This increases coupling to 0.172.

I was also thinking that the wires I have used from the bridge to primary circuit are too small. The thickest wire I had was AWG16, which is quite thin but it is all I had. I thought that it wouldn't be as important as the thickness of the primary coil as the length is much shorter compared to the length of the primary coil. Therefore, I used copper pipe for the primary coil. Ill see if I can buy some thicker wire to use.

I increased the voltage with the variac up to 100VAC and saw that the sparks grew larger. The sparks were likely around 7-8cm long. The on-time was around 170uS, at 100BPS. I did not capture primary current + secondary current waveforms, however I did see them on the scope. It looked similar to the low voltage test, just the primary current flattens out at a peak before the end of the burst. I will post some waveforms soon.





Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1569
  • Karma: +28/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: First DRSSTC, No Sparks on Breakout
« Reply #34 on: August 19, 2020, 07:50:44 PM »
I apologise, but I also forgot to mention that I am not using a UCC27423 gate driving chip as used in the original driver 1.3b. This was because I could not find any locally or online for cheap. Instead, I had some UCC27322 chips which are also low-side drivers, but with a single output. These chips are 9A peak, just over double the 4A of the UCC27423. I am using two of the UCC27322 chips in a push-pull configuration for the same effect as the single UCC27423 chip used in the original driver. I have 47uF electrolytic on each chip and some ceramic capacitors. I made a new GDT and this GDT has a ratio of around 1:1.5. The GDT drives the IGBT gate at 22V peak. I have attached a waveform below.



The wave is looking less square than it did before. However, it still switched the IGBT and I could see primary current flowing on the scope.

This is why you have heating IGBTs!

You are running them in the linear region almost 100% of the time, just when it reaches maximum, its time to turn off.

You need to lower your gate resistance or improve your driving current capabilities

Gate drive waveform troubleshooting: http://www.richieburnett.co.uk/temp/gdt/gdt2.html

http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics
https://www.youtube.com/KaizerPowerElectronicsDk60/join - Please consider supporting the forum, websites and youtube channel!

Online davekni

  • High Voltage Senior
  • *****
  • Posts: 689
  • Karma: +34/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #35 on: August 19, 2020, 11:25:21 PM »
Yes, as Mads says, gate waveform is definitely too rounded now.  As you reduce resistance to fix the rounding, watch for overshoot.

Reading between the lines when you say "I made a new GDT and this GDT has a ratio of around 1:1.5.", it sounds like primary and secondary windings were added separately, not as twisted sets.  That makes GDT leakage inductance higher, which exacerbates overshoot once gate resistors are reduced.  It's better to wind GDTs as twisted pairs.  1:1.5 would require 12 twisted pairs to implement properly, so a bit awkward.  That's why I suggested twisting sets of 5 wires.  Not as good as pairs, but much better than individual primary and secondary windings.

For primary lead wire, a few 16AWG wires in parallel will work great, even slightly better than equivalent copper area in a single conductor.  Just a bit of home-made litz wire.  (I picked up a large spool of surplus 16AWG wire years ago.  Paralleling that wire is cheaper than buying new wire of the "correct" gauge.)

Do you have a good clamp for the primary coil movable tap?  Most options are fine, just not cheap plated-steel alligator clips.
David Knierim

Offline Yak

  • High Voltage Enthusiast
  • *
  • Posts: 11
  • Karma: +0/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #36 on: August 20, 2020, 02:01:59 PM »
Hello, thank you for the advice. I have adjusted the gate resistance in order to get a nice looking waveform. Currently I have 5ohms of resistance on each IGBT gate. Here is the waveform switching the transistor, while the transistor is switching 30V through the primary LC. The first image is of the high side IGBT, and the second is of the low side IGBT. Both waveforms are very similar, as expected. However, it looks like the low side IGBT has less rounding on the waveform. I was unable to get a perfect square wave, nor did I get any overshooting.





Yes, as Mads says, gate waveform is definitely too rounded now.  As you reduce resistance to fix the rounding, watch for overshoot.

Reading between the lines when you say "I made a new GDT and this GDT has a ratio of around 1:1.5.", it sounds like primary and secondary windings were added separately, not as twisted sets.  That makes GDT leakage inductance higher, which exacerbates overshoot once gate resistors are reduced.  It's better to wind GDTs as twisted pairs.  1:1.5 would require 12 twisted pairs to implement properly, so a bit awkward.  That's why I suggested twisting sets of 5 wires.  Not as good as pairs, but much better than individual primary and secondary windings.

For primary lead wire, a few 16AWG wires in parallel will work great, even slightly better than equivalent copper area in a single conductor.  Just a bit of home-made litz wire.  (I picked up a large spool of surplus 16AWG wire years ago.  Paralleling that wire is cheaper than buying new wire of the "correct" gauge.)

Do you have a good clamp for the primary coil movable tap?  Most options are fine, just not cheap plated-steel alligator clips.

You would be correct in the way I wound the GDT. I did not have a choice as that large torroid I had sent in a previous post, I was unable to get any decent waveforms, or any transistor switching from it. I therefore had to use one of the smaller green torroids I used for all my CT's. I wound 20 turns of two twisted wires, and then wound 12 turns of two twisted wires for the primary coil. I just used wire from some network cable, as it was already in twisted pairs. With the method you described, I was unable to fit all the wires into the torroid and it would have been quite tedious, even though that method would likely yield better results.

I have not yet made the thicker primary leads yet, I will do that soon as I will have more time tomorrow and over the weekend. The current primary tap I am using is a fuse clip which I can just press onto the copper pipe and it fits nicely. I have attached a photo. Currently I am only using one of them, however once I make the thicker wire with multiple AWG16 wires in parallel, I will use two.







Online davekni

  • High Voltage Senior
  • *****
  • Posts: 689
  • Karma: +34/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #37 on: August 21, 2020, 04:20:08 AM »
Gate drive waveforms look much better!  One slight reason for the lower gate waveform appearing more square is that it's at a lower frequency, ~166kHz.  The upper trace is at ~190kHz.  The same edge rounding appears more square when combined with longer flat sections.

There's probably no need to further optimize.  If you wanted more square waveforms, the winding I'd suggested should be possible in the same core size.  Your design has a total of 64 individual wires (32 pairs) passing through the core.  If you made the two 5-wire twisted bundles and wound 6 turns each, that's a total of 60 individual wires passing through the core.  The primary windings have the same 12 turns each as in your version, two 6-turn wires in series.  The secondary windings would have 18 turns each (three 6-turn wires in series), so slightly lower gate voltage.  That would reduce leakage inductance, allowing you to further drop series resistance.  Again, looks good enough, so I'm not suggesting bothering.  (Also requires making the series connections on the 5-wire bundles close to the core to avoid extra inductance.)

I'd suggest a quick test to see if the fuse-clips stick to a magnet.  If they do, then performance depends on how thick plating is.  If not, then they are likely made of spring-brass, making plating thickness less important.  Or, just see how hot the clips get immediately after a run.  (They will cool off quickly since against copper tubing and soldered to copper wire, so need to check quickly.)
David Knierim

Offline Yak

  • High Voltage Enthusiast
  • *
  • Posts: 11
  • Karma: +0/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #38 on: August 21, 2020, 06:56:20 AM »
I have replaced the primary lead wires with 3x AWG16 wires twisted in parallel. I have also changed my primary tap from one to two fuse clips. The fuse clips are not magnetic. After doing all of this, I ran the coil at 30V DC, to ensure that all was working properly. I ran the coil at 100BPS at 175uS on-time. I was pleasantly surprised with how good the results were. I noticed that primary current was much, much higher than I expected. In addition, the secondary current was the highest I have ever seen it with 30V at the input. I also believe that I found a nice tap point on the primary for close to optimal low-voltage tuning. Here is the waveform attatched.

Yellow is primary current at 1v per division (10mV = 1A), and secondary current is at 500mV per division in purple (1V = 1A). Primary current went up to peaks of 100A. I am not sure if this is too much as I am scared at full power the OCD will be on the entire time. Secondary current peaked at around 1.08A



Here is a photo of the primary tap. I have tapped it at like 6.1 turns.



Finally, I ran the coil at 44VAC from a variac. In this run, I got the biggest sparks I had ever seen out of this coil. I was originally planning to do a test at 100VAC, but I was too scared to do this as I believe the streamers would have been quite long, possibly hitting stuff on my cluttered desk, and causing damage. I will need to do this test in a more open room, or maybe outside. At this voltage, I did not feel the fuse clips get hot, nor did the IGBT's warm up much. Here is a photo of the arcs, it isnt the best photo as the coil was messing around with my phones touchscreen.



I think that if I can find the best tap point on the primary for tuning, this coil would be ready for a full-power test run. However, I still do not have a strike ring above the primary. I will first make one before I do a high power test as I do not want to risk the transistors exploding, or a strike possibly damaging something.

« Last Edit: August 21, 2020, 06:58:09 AM by Yak »

Offline acmq

  • High Voltage Enthusiast
  • *
  • Posts: 15
  • Karma: +2/-0
    • View Profile
Re: First DRSSTC, No Sparks on Breakout
« Reply #39 on: August 21, 2020, 05:16:57 PM »
The waveforms indicate too high coupling. The peaks of the output voltage don't have enough time to rise. With proper adjustments and light load the maximum voltage can be reached at the first beat, coinciding with a current notch.

High Voltage Forum

Re: First DRSSTC, No Sparks on Breakout
« Reply #39 on: August 21, 2020, 05:16:57 PM »

 


* Recent Topics and Posts

post Re: Have you ever tripped a breaker during normal operation ?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
profdc9
Today at 06:20:35 PM
post Re: Have you ever tripped a breaker during normal operation ?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
panjaksli
Today at 05:24:23 PM
post Re: Have you ever tripped a breaker during normal operation ?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
buchtawill
Today at 03:25:59 PM
post Re: Have you ever tripped a breaker during normal operation ?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
panjaksli
Today at 03:02:01 PM
post Have you ever tripped a breaker during normal operation ?
[Dual Resonant Solid State Tesla coils (DRSSTC)]
panjaksli
Today at 01:36:04 PM
post Re: mini tesla (hopefully not a fail)
[Beginners]
bogdan
Today at 08:55:06 AM
post Re: HF Litz wire as a primary coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
johnf
Today at 07:43:14 AM
post Re: How to design FPGA based on EDA technology?
[Electronic Circuits]
Genterman
Today at 06:53:52 AM
post Re: DRSSTC trouble
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
Today at 03:29:14 AM
post Re: DRSSTC trouble
[Dual Resonant Solid State Tesla coils (DRSSTC)]
buchtawill
Today at 02:43:33 AM
post Re: SKP DRSSTC = aliexpress driver + medium coil + large bricks ;)
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Maju
December 01, 2020, 10:44:31 PM
post Re: mini tesla (hopefully not a fail)
[Beginners]
davekni
December 01, 2020, 07:44:14 PM
post Re: transistor markings question
[Beginners]
davekni
December 01, 2020, 07:39:48 PM
post Re: HF Litz wire as a primary coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
December 01, 2020, 07:38:53 PM
post Re: How to design FPGA based on EDA technology?
[Electronic Circuits]
petespaco
December 01, 2020, 06:30:33 PM
post Re: Question about arduino polyphonic MIDI interrupter
[Dual Resonant Solid State Tesla coils (DRSSTC)]
bozidar
December 01, 2020, 04:58:29 PM
post Re: Question about arduino polyphonic MIDI interrupter
[Dual Resonant Solid State Tesla coils (DRSSTC)]
TMaxElectronics
December 01, 2020, 04:22:10 PM
post Re: Question about arduino polyphonic MIDI interrupter
[Dual Resonant Solid State Tesla coils (DRSSTC)]
bozidar
December 01, 2020, 03:02:51 PM
post Re: Half-brigde with pot on mosfet gates question
[Solid State Tesla Coils (SSTC)]
TMaxElectronics
December 01, 2020, 12:08:56 PM
post transistor markings question
[Beginners]
NOOBPASTE51
December 01, 2020, 12:02:19 PM
post Re: Question about arduino polyphonic MIDI interrupter
[Dual Resonant Solid State Tesla coils (DRSSTC)]
TMaxElectronics
December 01, 2020, 11:55:01 AM
post Re: HF Litz wire as a primary coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
TMaxElectronics
December 01, 2020, 11:50:19 AM
post Re: Analog HFBR amplifier
[Electronic Circuits]
TMaxElectronics
December 01, 2020, 11:21:56 AM
post Half-brigde with pot on mosfet gates question
[Solid State Tesla Coils (SSTC)]
costas_p
December 01, 2020, 11:17:16 AM
post Re: Question about arduino polyphonic MIDI interrupter
[Dual Resonant Solid State Tesla coils (DRSSTC)]
bozidar
December 01, 2020, 11:11:50 AM
post Re: mini tesla (hopefully not a fail)
[Beginners]
bogdan
December 01, 2020, 09:29:24 AM
post Re: HF Litz wire as a primary coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
johnf
December 01, 2020, 07:03:56 AM
post Re: How to design FPGA based on EDA technology?
[Electronic Circuits]
johnf
December 01, 2020, 06:49:26 AM
post Re: SSTC or DRSSTC with GaN Transistors
[Solid State Tesla Coils (SSTC)]
Steve Ward
December 01, 2020, 04:59:28 AM
post How to design FPGA based on EDA technology?
[Electronic Circuits]
Genterman
December 01, 2020, 04:02:06 AM
post Re: DRSSTC trouble
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
December 01, 2020, 03:46:10 AM
post Re: DRSSTC trouble
[Dual Resonant Solid State Tesla coils (DRSSTC)]
buchtawill
December 01, 2020, 02:00:54 AM
post Re: HF Litz wire as a primary coil
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Steve Ward
December 01, 2020, 01:47:10 AM
post Re: DRSSTC trouble
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
December 01, 2020, 12:02:23 AM
post Re: Site Info desperately needed: http://www.stevehv.4hv.org/
[General Chat]
Mads Barnkob
November 30, 2020, 10:28:56 PM
post Re: I started making a drsstc
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Miki_407
November 30, 2020, 08:58:41 PM
post DRSSTC trouble
[Dual Resonant Solid State Tesla coils (DRSSTC)]
buchtawill
November 30, 2020, 08:53:00 PM
post Re: I started making a drsstc
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
November 30, 2020, 08:49:30 PM
post Re: Simple H-Bridge construction with low parasitic inductances (for SSTC or ...)
[Beginners]
davekni
November 30, 2020, 08:45:46 PM
post Re: mini tesla (hopefully not a fail)
[Beginners]
davekni
November 30, 2020, 06:56:13 PM
post Re: I started making a drsstc
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Miki_407
November 30, 2020, 01:16:04 PM
post Re: I started making a drsstc
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
November 30, 2020, 12:56:05 PM
post Re: mini tesla (hopefully not a fail)
[Beginners]
bogdan
November 30, 2020, 12:03:06 PM
post I started making a drsstc
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Miki_407
November 30, 2020, 11:58:29 AM
post Re: Interesting 2400W power supply circuit
[Electronic Circuits]
Kiki
November 30, 2020, 07:58:52 AM
post Re: QCW with replaceable ferrite-core primary
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
November 30, 2020, 06:37:23 AM
post Re: Analog HFBR amplifier
[Electronic Circuits]
USAMAN
November 30, 2020, 03:07:30 AM
post Re: SKP DRSSTC = aliexpress driver + medium coil + large bricks ;)
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
November 29, 2020, 08:14:37 PM
post Re: mini tesla (hopefully not a fail)
[Beginners]
davekni
November 29, 2020, 08:04:33 PM
post Re: QCW with replaceable ferrite-core primary
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
November 29, 2020, 08:04:01 PM

Sitemap 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 
SimplePortal 2.3.6 © 2008-2014, SimplePortal