Author Topic: Strange oscillations  (Read 1129 times)

Offline AstRii

  • High Voltage Enthusiast
  • *
  • Posts: 26
  • Karma: +0/-0
  • Czech Technical University in Prague
    • View Profile
Re: Strange oscillations
« Reply #20 on: January 12, 2020, 06:33:22 PM »
No i don't mind the question, i'm very grateful for your help. It would be a dream come true for me to see the coil running at full power by the end of the month.
Here is another waveform: low side emitter - high side collector

Offline AstRii

  • High Voltage Enthusiast
  • *
  • Posts: 26
  • Karma: +0/-0
  • Czech Technical University in Prague
    • View Profile
Re: Strange oscillations
« Reply #21 on: January 12, 2020, 06:39:33 PM »
Oh and here are again the outputs of the inverter referenced to Vbus- more closed up.
I wound the GDT by taking wires from cat5, twisted them and wound them together on the GDT to create a primary and 4 secondaries.
« Last Edit: January 12, 2020, 07:07:34 PM by AstRii »

Offline T3sl4co1l

  • High Voltage Technician
  • ***
  • Posts: 118
  • Karma: +9/-0
    • View Profile
Re: Strange oscillations
« Reply #22 on: January 12, 2020, 09:08:17 PM »
Yikes, looks like it's either dropping out or shorting momentarily.

Tim

Offline davekni

  • High Voltage Technician
  • ***
  • Posts: 188
  • Karma: +5/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Strange oscillations
« Reply #23 on: January 13, 2020, 12:34:48 AM »
I'm fairly confident that the artifacts are caused by shoot-through (momentary short-circuit when a high-side IGBT turns on before the corresponding low-side IGBT turns off and visa-versa).  Two options come to mind on how to chase this down.

First thought is to search for the asymmetry.  Especially in some of your earlier scope traces at 30VBus, such as CEwave.jpg, the bridge output (low-side Vce) looks asymmetric, with one output having <50% duty cycle and the other output >50% duty cycle.  The gate drive waveform shows some duty-cycle offset (slightly over 50%), but it's not as much as the bridge outputs.  The shoot-through appears to occur on every half-cycle, but it may all be on one of the half-bridges, or perhaps only on rising edges (toggling between half-bridges, which ever is rising at the moment).

Do you have a signal generator that you could feed into the UD2.7 feedback input instead of the current transformer?  If so, disconnecting the Tesla primary (MMC and/or primary winding) and running with no H-Bridge load may make it easier to debug shoot-through.  If shoot-through is the correct diagnosis, there will still be 1MHz ringing even without load on the H-Bridge.  Once that's shown, you could disconnect the gate-drive from one half-bridge (and short gate-to-emitter) to see if shoot-through occurs on that side.  Repeat for the other side.

The other thought is to ignore the asymmetry and focus in general on reducing shoot-through (speed-up gate falling edges and slow-down rising edges).  That implies reducing the 5.6-ohm pull-down resistor and increasing the 15-ohm pull-up resistor.  However, reducing the 5.6-ohm resistor will increase gate-drive overshoot.  To avoid that, gate-drive transformer (GDT) leakage and wiring inductance needs to be reduced.  One way to do that is to wind the GDT with 8 wires (4 twisted pairs).  Parallel 4 wires, one from each twisted-pair, to use for the primary.  The other wire from each pair is one of the four secondaries.  For each of the secondaries, twist the two ends of the winding from the GDT to where it solders to the power ECB.  For each of the 4 paralleled primary windings, twist the two ends of each winding as a pair back to the UD2.7, then parallel the four windings at that UD2.7 board.  (Take care to label each of the four windings before twisting to make sure they are all parallel and not anti-parallel.  Use an LRC meter to measure inductance before and after paralleling as a double-check.  If the inductance drops significantly after paralleling, one or more of the windings is anti-parallel.)

Reducing the UD2.7 gate-drive supply voltage will allow more room for overshoot, and may help some with shoot-through too.

What core are you using for the GDT?  If you don't have full specifications, roughly what is the cross-sectional area of the core (area within one turn)?  At 220kHz, it may be possible to lower the turn-count, which will lower leakage inductance as well.

Finally, once shoot-through is fixed, the UD2.7 phase-lead inductor needs to be increased.  The gate-drive waveform at the driver is in-phase with current.  However, with delays to the actual IGBT Vge, and further delays within the IGBTs, the actual switching times are significantly after the current zero-crossings.  IGBT switching losses are lowest when switching is just before current zero-crossing rather than after.k

Good luck!
David Knierim

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 293
  • Karma: +9/-0
    • View Profile
Re: Strange oscillations
« Reply #24 on: January 13, 2020, 10:19:03 AM »
Regarding the leakage inductance of the GDT, it's also important not to forget the length of the leads, both to the primary and secondary. At ~100 ohms Z0 they have a fair amount of inductance per unit length, which can easily exceed the leakage inductance of a well designed GDT if they are too long.

In my 300+ kHz coil I've actually used a ribbon cable to connect the gate drive to the half bridges - this way I can parallel multiple pairs to reduce inductance significantly.

Beyond this it's probably best to eat the cost and complexity of doing away with the GDT and having an isolated driver per IGBT - I'd be doing this if I had fewer gates to drive (16!).

Offline T3sl4co1l

  • High Voltage Technician
  • ***
  • Posts: 118
  • Karma: +9/-0
    • View Profile
Re: Strange oscillations
« Reply #25 on: January 13, 2020, 02:36:04 PM »
Useful way to understand GDTs: the lead length doesn't stop at the bulk of the transformer.  It continues for the entire winding length.  Which makes sense right, it's just a ball of wire?

Ideally you'd use something like two star quads in parallel, for Zo around 10 ohms, a much closer match to the gate resistance.  These can be thin wires so you don't have to take up any more space.  And obviously it's repeated for each gate, and all the primaries are wired in parallel.

Or, to put it another way, consider using a bunch of transformers, each using twisted pairs, all wired in parallel.  Well obviously, if they're all the same core and number of turns, we don't need to use separate cores, we can wind all the pairs on the same core, that's easy.  Again, lots of pairs in parallel means the individual pairs can be thin wire (wirewrap wire is good for this, as is multiconductor cable with thin wires: say, parallel port cables, some USB, HDMI, have seen non-ribbon style IDE cables too).

One trick: "star quad" arrangement is about half the impedance of two twisted pairs in parallel.  So it's a huge space savings.  Just take two pairs and twist them around each other, wired so that the cross section alternates PSPS in a circle.

GDTs aren't very good for IGBTs because there's no way to apply negative gate bias during turn-off.  Turn-off and opposite side turn-on are synchronized by force.  Preferable to use gate driver modules, and drive them with complementary logic signals.  Much easier, too.  If you're salvaging IGBTs from equipment, pay attention to what they're connected to; the power and driver boards are likely quite usable with a little tracing.  Otherwise, off the shelf driver modules run $20-100, not a bad deal for a $100+ IGBT that you'd prefer to not use up as a "wear item".  (That is to say, these also offer desat protection and maybe current limiting.  Make use of them!)

Tim

Offline AstRii

  • High Voltage Enthusiast
  • *
  • Posts: 26
  • Karma: +0/-0
  • Czech Technical University in Prague
    • View Profile
Re: Strange oscillations
« Reply #26 on: January 28, 2020, 08:36:51 PM »
I've tried to completely bypass the charing/discharing resistors, since i've found out that my GDT cannot supply much current and the turnon/turnoff takes a lot of time. Now the waveforms are pretty clear and i tried to run the coil with 260V on the bus and everything seems to work very well. Temperature of IGBT heatsink is just few degrees above ambient. I wanted to wait before sharing the results until i push it to the maximum power, but i'm so excited that it works that i just had to play MIDI :D. So here is the result
/>And yes i know, i need larger topload and a better camera than my phone :D

Offline davekni

  • High Voltage Technician
  • ***
  • Posts: 188
  • Karma: +5/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Strange oscillations
« Reply #27 on: January 29, 2020, 05:59:12 AM »
Congratulations on getting it running!

If the only change is removing gate resistors, I suspect a bit more lead compensation in the driver circuit would help.  Best to have the actual H-bridge output switching just slightly before the zero-current points.
David Knierim

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1261
  • Karma: +17/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Strange oscillations
« Reply #28 on: January 30, 2020, 10:01:37 AM »
I've tried to completely bypass the charing/discharing resistors, since i've found out that my GDT cannot supply much current and the turnon/turnoff takes a lot of time. Now the waveforms are pretty clear and i tried to run the coil with 260V on the bus and everything seems to work very well.

Good work, that you came to fly some sparks of a decent length, just be careful not to blow up the bridge in excitement (we all did that) over its working with some _minor_ flaws that turns out significant at higher voltages :)

Did you measure/scope your DC supply rail to the GDT drive circuit? Is the voltage sagging here?

I am not completely sure which charge/discharge resistors it is you have bypassed?
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

Offline AstRii

  • High Voltage Enthusiast
  • *
  • Posts: 26
  • Karma: +0/-0
  • Czech Technical University in Prague
    • View Profile
Re: Strange oscillations
« Reply #29 on: January 30, 2020, 07:16:12 PM »
Mads, i mean the gate resistors, i had 15R gate charge resistor and 5R6 discharge through diode gate resistor. I think(hope) there aren't any more flaws so i can run it at full power but i didn't find the courage to do that yet :D As soon as i scope some waveforms at higher voltages i will post them here :) Now i need to solve how to lower primary currents, i have OCD set to 300A and already at 260V on the bus i get OCD on pretty much every interrupt even with very short on times.

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1261
  • Karma: +17/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Strange oscillations
« Reply #30 on: January 30, 2020, 10:17:33 PM »
Mads, i mean the gate resistors, i had 15R gate charge resistor and 5R6 discharge through diode gate resistor. I think(hope) there aren't any more flaws so i can run it at full power but i didn't find the courage to do that yet :D As soon as i scope some waveforms at higher voltages i will post them here :) Now i need to solve how to lower primary currents, i have OCD set to 300A and already at 260V on the bus i get OCD on pretty much every interrupt even with very short on times.

Running without gate resistors should cause even more trouble from ringing on the gates. Be careful with that.

How well did you tune the coil? If it is out of tune the current in the primary circuit will ring up fast as not all of it is transferred to the secondary.
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

Offline AstRii

  • High Voltage Enthusiast
  • *
  • Posts: 26
  • Karma: +0/-0
  • Czech Technical University in Prague
    • View Profile
Re: Strange oscillations
« Reply #31 on: January 30, 2020, 10:42:52 PM »
It's not tuned well i would say, i only tuned it while running on low bus voltage and checked when the arcs are the longest. Running JavaTC is not an option for me since i don't know exact parameters of my secondary as i didn't make it, i bought it online :D

« Last Edit: January 31, 2020, 09:58:02 AM by Mads Barnkob »

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1261
  • Karma: +17/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Strange oscillations
« Reply #32 on: January 31, 2020, 09:58:09 AM »
It's not tuned well i would say, i only tuned it while running on low bus voltage and checked when the arcs are the longest. Running JavaTC is not an option for me since i don't know exact parameters of my secondary as i didn't make it, i bought it online :D

The seller must have written some parameters, but even if they didn't you only have to measure what you have to get it.

You can measure diameter, winding length and wire diameter with a caliper and calculate amount of turns or other missing number: http://kaizerpowerelectronics.dk/calculators/helical-coil-calculator/

Running a JavaTC simulation for tuning and tune the primary a little lower is the absolute best way to get a good starting point.

Trial and error for largest sparks without a proper starting point could be running at a harmonic frequency and you would get sparks but its horribly out of perfect tune.
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

Offline Teravolt

  • High Voltage Technician
  • ***
  • Posts: 103
  • Karma: +2/-0
    • View Profile
Re: Strange oscillations
« Reply #33 on: January 31, 2020, 06:25:02 PM »
I don't know how well your driver is built but it looks like mabey you have little dead time or none at all. noise at the transitions is symptom of shoot through.

High Voltage Forum

Re: Strange oscillations
« Reply #33 on: January 31, 2020, 06:25:02 PM »

 


* Recent Topics and Posts

post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Uspring
Today at 02:03:12 PM
post Is this flyback OK for my Plasma ball ?
[Transformer (Ferrite Core)]
warmania
Today at 11:24:58 AM
post Re: HP T620 Thin Client - PSU Hack And Windows PC Modification
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
latethannever
Today at 11:00:15 AM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
rikkitikkitavi
Today at 10:43:13 AM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Peregrine
Today at 06:54:14 AM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
klugesmith
Today at 06:51:52 AM
post Re: 230 volts spark gap
[Spark Gap Tesla Coils (SGTC)]
davekni
Today at 03:53:13 AM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
ElectroXa
February 18, 2020, 09:44:25 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
MRMILSTAR
February 18, 2020, 09:34:06 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Peregrine
February 18, 2020, 08:27:31 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Weston
February 18, 2020, 07:55:14 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Peregrine
February 18, 2020, 07:32:04 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
klugesmith
February 18, 2020, 05:58:39 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
MRMILSTAR
February 18, 2020, 05:16:55 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Peregrine
February 18, 2020, 04:19:37 PM
post Re: Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Mads Barnkob
February 18, 2020, 04:12:11 PM
post Re: 230 volts spark gap
[Spark Gap Tesla Coils (SGTC)]
Quentief
February 18, 2020, 01:25:06 PM
post Help!Capacitor that burns out after a while!
[Voltage Multipliers]
Peregrine
February 18, 2020, 10:40:58 AM
post Re: How can I make my driver better
[Solid State Tesla Coils (SSTC)]
Mads Barnkob
February 18, 2020, 09:02:23 AM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
Mads Barnkob
February 18, 2020, 08:58:38 AM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
davekni
February 18, 2020, 05:33:22 AM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
davekni
February 18, 2020, 05:08:28 AM
post How can I make my driver better
[Solid State Tesla Coils (SSTC)]
Irido
February 18, 2020, 04:45:06 AM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
klugesmith
February 18, 2020, 02:44:42 AM
post Re: HP T620 Thin Client - PSU Hack And Windows PC Modification
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
latethannever
February 18, 2020, 01:58:48 AM
post Re: UV-cured cyanoacrylate glues?
[Light, Lasers and Optics]
Twospoons
February 17, 2020, 11:30:03 PM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
Twospoons
February 17, 2020, 11:21:14 PM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
Mads Barnkob
February 17, 2020, 10:45:52 PM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
ElectroXa
February 17, 2020, 08:20:42 PM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
Twospoons
February 17, 2020, 08:02:29 PM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
Mads Barnkob
February 17, 2020, 08:02:15 PM
post Re: DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
haversin
February 17, 2020, 04:26:10 PM
post Re: HP T620 Thin Client - PSU Hack And Windows PC Modification
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Mads Barnkob
February 17, 2020, 01:44:41 PM
post Re: HP T620 Thin Client - PSU Hack And Windows PC Modification
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
latethannever
February 17, 2020, 11:18:33 AM
post DIY 250W LED Flood Light for High Speed Filming
[Light, Lasers and Optics]
Mads Barnkob
February 17, 2020, 10:58:17 AM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
kilovolt
February 17, 2020, 08:55:30 AM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic Circuits]
petespaco
February 17, 2020, 06:00:59 AM
post Re: 230 volts spark gap
[Spark Gap Tesla Coils (SGTC)]
davekni
February 17, 2020, 03:59:59 AM
post Re: 230 volts spark gap
[Spark Gap Tesla Coils (SGTC)]
johnf
February 17, 2020, 03:19:32 AM
post Re: 230 volts spark gap
[Spark Gap Tesla Coils (SGTC)]
Quentief
February 17, 2020, 12:14:07 AM
post Re: 230 volts spark gap
[Spark Gap Tesla Coils (SGTC)]
Quentief
February 17, 2020, 12:10:28 AM
post Re: Saxophone controlled DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
February 16, 2020, 07:39:33 PM
post Re: WTB [EU]: Cellular Sector Antenna
[Sell / Buy / Trade]
Da_Stier
February 15, 2020, 03:25:10 PM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
davekni
February 15, 2020, 05:41:13 AM
post kilovolt: website, youtube
[Hobby Sites]
kilovolt
February 14, 2020, 11:53:39 AM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
kilovolt
February 14, 2020, 08:27:03 AM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
davekni
February 14, 2020, 05:42:19 AM
post Re: Help with Van De Graaff Generator Build
[General Chat]
pianomanjcb
February 14, 2020, 03:29:37 AM
post Re: Help with Van De Graaff Generator Build
[General Chat]
haversin
February 14, 2020, 01:07:47 AM
post Re: Help with Van De Graaff Generator Build
[General Chat]
Mads Barnkob
February 13, 2020, 10:14:27 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