Author Topic: Strange oscillations  (Read 637 times)

Offline AstRii

  • High Voltage Enthusiast
  • *
  • Posts: 22
  • 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: 22
  • 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: 116
  • 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: 157
  • 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: 291
  • 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: 116
  • 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

High Voltage Forum

Re: Strange oscillations
« Reply #25 on: January 13, 2020, 02:36:04 PM »

 


* Recent Topics and Posts

post Re: GM tube hv transformer
[Detection, Counting and Measurements]
shrad
Today at 12:32:39 PM
post Re: WTS:Power diodes "hockey Puck" MDF 652 02
[Sell / Buy / Trade]
rikkitikkitavi
January 27, 2020, 09:01:23 PM
post Re: WTS:Power diodes "hockey Puck" MDF 652 02
[Sell / Buy / Trade]
Da_Stier
January 27, 2020, 05:56:32 PM
post Re: Mazilli ZVS Driver Modification Problems
[Transformer (Ferrite Core)]
T3sl4co1l
January 27, 2020, 12:54:42 AM
post WTS:Power diodes "hockey Puck" MDF 652 02
[Sell / Buy / Trade]
rikkitikkitavi
January 26, 2020, 10:13:38 PM
post Re: Mazilli ZVS Driver Modification Problems
[Transformer (Ferrite Core)]
davekni
January 26, 2020, 04:55:49 AM
post Re: GM tube hv transformer
[Detection, Counting and Measurements]
futurist
January 25, 2020, 11:01:24 AM
post Re: CT performance investigations (Pearson and DIY)
[Dual Resonant Solid State Tesla coils (DRSSTC)]
T3sl4co1l
January 25, 2020, 06:44:21 AM
post Re: Mazilli ZVS Driver Modification Problems
[Transformer (Ferrite Core)]
T3sl4co1l
January 25, 2020, 06:37:47 AM
post Re: Mazilli ZVS Driver Modification Problems
[Transformer (Ferrite Core)]
Andrew321
January 25, 2020, 02:22:09 AM
post Re: GM tube hv transformer
[Detection, Counting and Measurements]
haversin
January 25, 2020, 01:15:21 AM
post Re: GM tube hv transformer
[Detection, Counting and Measurements]
klugesmith
January 24, 2020, 10:53:42 PM
post Re: WTB [EU]: Cellular Sector Antenna
[Sell / Buy / Trade]
Da_Stier
January 24, 2020, 10:52:48 PM
post Re: GM tube hv transformer
[Detection, Counting and Measurements]
haversin
January 24, 2020, 10:29:14 PM
post Re: Problems with my first DRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
bozidar
January 24, 2020, 08:16:12 PM
post GM tube hv transformer
[Detection, Counting and Measurements]
AlexanderHun
January 24, 2020, 06:24:25 PM
post Re: DIY SMAW welder (stick welding)
[Electronic Circuits]
rikkitikkitavi
January 24, 2020, 05:17:01 PM
post Re: My 833A VTTC
[Vacuum Tube Tesla Coils (VTTC)]
MRMILSTAR
January 24, 2020, 04:37:25 PM
post Re: simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
justin66
January 24, 2020, 04:15:41 PM
post Re: My 833A VTTC
[Vacuum Tube Tesla Coils (VTTC)]
Mads Barnkob
January 24, 2020, 01:15:46 PM
post Re: simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
January 24, 2020, 01:10:38 PM
post Re: simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Hydron
January 24, 2020, 12:39:01 PM
post Re: simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
justin66
January 24, 2020, 11:43:08 AM
post Re: My 833A VTTC
[Vacuum Tube Tesla Coils (VTTC)]
kilovolt
January 24, 2020, 11:10:55 AM
post Re: DIY SMAW welder (stick welding)
[Electronic Circuits]
kamelryttarn
January 24, 2020, 10:11:51 AM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
kilovolt
January 24, 2020, 09:03:52 AM
post Re: CT performance investigations (Pearson and DIY)
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Hydron
January 23, 2020, 11:50:22 PM
post Re: simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
January 23, 2020, 11:42:56 PM
post Re: WTB [EU]: Cellular Sector Antenna
[Sell / Buy / Trade]
Mads Barnkob
January 23, 2020, 11:35:07 PM
post Re: simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
justin66
January 23, 2020, 11:15:12 PM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
MRMILSTAR
January 23, 2020, 10:42:22 PM
post Re: CT performance investigations (Pearson and DIY)
[Dual Resonant Solid State Tesla coils (DRSSTC)]
T3sl4co1l
January 23, 2020, 10:03:17 PM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
kilovolt
January 23, 2020, 09:55:05 PM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
MRMILSTAR
January 23, 2020, 08:39:10 PM
post Re: Universal bus precharge & fan speed controller
[Dual Resonant Solid State Tesla coils (DRSSTC)]
acobaugh
January 23, 2020, 08:37:43 PM
post Re: simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
acobaugh
January 23, 2020, 08:16:02 PM
post simple confusion about UD2.7c LEDs
[Dual Resonant Solid State Tesla coils (DRSSTC)]
justin66
January 23, 2020, 07:48:39 PM
post Re: Bipolar high voltage multiplier project - more than 70cm hot arc discharge
[Voltage Multipliers]
kilovolt
January 23, 2020, 05:07:46 PM
post Re: My 240kv voltage multiplier
[Voltage Multipliers]
kilovolt
January 23, 2020, 05:05:40 PM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
kilovolt
January 23, 2020, 05:01:27 PM
post WTB [EU]: Cellular Sector Antenna
[Sell / Buy / Trade]
Da_Stier
January 23, 2020, 04:34:34 PM
post Re: My X-Ray Machine
[X-ray]
Agent Smith
January 23, 2020, 03:47:06 PM
post Re: DIY SMAW welder (stick welding)
[Electronic Circuits]
rikkitikkitavi
January 23, 2020, 02:24:07 PM
post Re: DIY SMAW welder (stick welding)
[Electronic Circuits]
kamelryttarn
January 23, 2020, 01:29:01 PM
post Re: DIY SMAW welder (stick welding)
[Electronic Circuits]
rikkitikkitavi
January 23, 2020, 12:45:40 PM
post Re: DIY SMAW welder (stick welding)
[Electronic Circuits]
kamelryttarn
January 23, 2020, 11:38:18 AM
post Re: DIY SMAW welder (stick welding)
[Electronic Circuits]
rikkitikkitavi
January 23, 2020, 10:43:36 AM
post DIY SMAW welder (stick welding)
[Electronic Circuits]
kamelryttarn
January 23, 2020, 08:32:56 AM
post Re: Doorknob vs Ceramic Disk
[Voltage Multipliers]
davekni
January 23, 2020, 06:06:20 AM
post Re: Mazilli ZVS Driver Modification Problems
[Transformer (Ferrite Core)]
davekni
January 23, 2020, 05:57:18 AM

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