Author Topic: Current Sense transformer trouble (solved)  (Read 1397 times)

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 465
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Current Sense transformer trouble
« Reply #20 on: June 15, 2020, 10:22:24 PM »
There's something quite odd about your feedback CT signal.  It starts out in-phase with the measurement signal, then shifts to almost 90 degrees lagging, then turns into some higher-frequency ringing at the end while the measurement CT is showing a continued normal sine wave.  Almost looks like some insulation breakdown in the feedback CT.  I'd suggest chasing that issue before others, as it might be the cause of bad gate signals.

Yes, switching at high-current points in the primary waveform after the end of the enable pulse could be the cause of fried IGBTs.

Does the H-Bridge have some path to ground?  Is your feedback CT on one of the H-Bridge outputs, not at the MMC-to-primary coil connection?  The MMC-primary point is relatively high voltage, so could cause arcing issues to the CT.
David Knierim

Offline TMaxElectronics

  • High Voltage Experimenter
  • **
  • Posts: 73
  • Karma: +2/-0
    • View Profile
    • My random (and very empty) electronics blog
Re: Current Sense transformer trouble
« Reply #21 on: June 16, 2020, 02:44:03 AM »
I noticed the weirdness in the CT waveform too, but think it is more likely to be a measurement error, because the gate drive waveform doesn't really change its phase. I don't really know what that could be from though.

The transformer is on the bridge output going to the primary.

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 465
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Current Sense transformer trouble
« Reply #22 on: June 16, 2020, 02:54:38 AM »
Possible to be measurement issues, often ground-loops through the scope.  If you have any clamp-on ferrites, loop the scope probe cable a few turns through the ferrite and clamp it closed.  That breaks the ground loops for AC signals.  I'd suggest doing that or scoping the feedback CT signal by itself to verify.  I'm having a hard time picturing what measurement interference would cause the signal in your plots.
David Knierim

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 325
  • Karma: +11/-0
    • View Profile
Re: Current Sense transformer trouble
« Reply #23 on: June 16, 2020, 11:44:18 AM »
After the interrupter/IGBTs turns off it is expected that the voltage across the bridge will continue to flip as the energy stored in the tank circuit returns to the bus capacitors via the IGBT anti-parallel diodes. This should be opposite to the voltage you see when the interrupter is on and IGBTs are being driven in sync with the current. See attached example:


This is only relevant to the "some weird additional switching" comment

Offline TMaxElectronics

  • High Voltage Experimenter
  • **
  • Posts: 73
  • Karma: +2/-0
    • View Profile
    • My random (and very empty) electronics blog
Re: Current Sense transformer trouble
« Reply #24 on: June 16, 2020, 01:52:09 PM »
That weird CT waveform was just from bad probing (i was using the ground lead to connect to the CT output. Now I soldered some wires onto it and made a nice and tight connection to the probe and it looks fine (except for those HUGE spikes? Or are they maybe caused by the arc?):


Quote
After the interrupter/IGBTs turns off it is expected that the voltage across the bridge will continue to flip as the energy stored in the tank circuit returns to the bus capacitors via the IGBT anti-parallel diodes.
The additional switching was not measured on the bridge outputs, but was the signal coming from the controller, so i think that's an actual problem rather than an expected phenomenon.

EDIT:
I measured the output from the interrupter and that seems to be the cause (ch2 is the interrupter signal, active low):

It does coincide with the additional switch pulse, so it doesn't seem to be an artifact of my probing (https://i.imgur.com/jMdwSdA.jpg). I measured the supply voltage at the receiver and there was quite some noise (7Vpp if my scope can be trusted), so that might be the cause... that i don't know how to fix. I have copper shielding around the entire controller now, and all of the control cables are shielded (the fan power is not, but the noise is there even if it is disconnected), and there are plenty of decoupling caps around the circuit, so where would such noise even come from?
One solution I thought of, would be to low pass filter the interrupter signal, but that seems like a bad workaround instead of a solution to me. Somehow noise must still make its way inside and i have no idea how (the OCD is going nuts once an arc strikes too)
« Last Edit: June 16, 2020, 11:22:14 PM by TmaxElectronics »

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 465
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Current Sense transformer trouble
« Reply #25 on: June 17, 2020, 07:10:52 PM »
Are you getting arcs from the top of your secondary (top load) that reach some other object?  An arc that hits something else (hopefully a grounded something) makes more noise (spikes) than just the air arcs.  Or are you perhaps getting arcs somewhere else, say lower in secondary to primary, or somewhere within the primary coil or MMC or wiring.  Primary arcs are much higher current and more problematic.

On your first scope image there's a spike in CT signal at the very start of the burst, before any arc should be possible.  That is quite puzzling.  If it can be repeated, it would be worth chasing that down using short bursts to avoid enough energy to damage anything.  Finding the source of that initial smaller spike may lead to understanding the larger ones that are confusing your interrupter. Although, it may be the other way around.  It may be the interrupter malfunction that causes IGBT switching at high current which in turn causes the spikes.  (Perhaps that's something to chase first.  Zoom far into one of the spikes - perhaps trigger on the spike - and see if the gate-drive edge is before or after the start of the spike.)

I'm suspecting there's something more than normal secondary arcs causing these spikes.  If so, that needs fixing.  For normal unavoidable noise (such as secondary arcs), the primary tools for handling such is shielding and common-mode chokes (ferrite beads around cables).
David Knierim

Offline TMaxElectronics

  • High Voltage Experimenter
  • **
  • Posts: 73
  • Karma: +2/-0
    • View Profile
    • My random (and very empty) electronics blog
Re: Current Sense transformer trouble
« Reply #26 on: June 18, 2020, 12:01:38 AM »
I had metal rod next to the toroid, at around 30cm distance as the arc target. I also didn't spot any arcs in the MMCs, so don't think that's an issue either.
I have since fixed those two issues, by adding low pass filtering to both the OCD comparator and the Interrupter. The OCD seems to have been confused by some fairly high frequency noise, not attenuated by the input RC network, now that the filter is in place the transfer function from CT to OCD input is nearly constant from 75-100 KHz and drops off after ~200 KHz, that seems to have done the trick. No more additional pulses and no more false triggering.

HOWEVER when i was setting up to verify the phase shift, one of the gate drivers died again. But i think i figured out why they kept dying (or at least one reason):
the logic inputs are designed for low voltage control (e.g. 5V) but i was giving them 15V, as i switched the driver ICs but not the isolator shortly before making my first version. The reason why they didn't just die instantly would be the 15V absolute maximum rating of these inputs, once high currents came into play though, the noise would have killed the poor things.
So I added some 100 ohm resistors to the board with the gate drivers, making around a 1/3 divider from the output of the isolator, and i also added some capacitors (200pF) in parallel, to quench some of the HF noise coming in. I also increased phase lead a little, so now the transistor switch on occurs at -100ns.
Testing them with no prim. voltage and at 35% duty cycle at 150KHz for 45min didn't cause damage and the temperature of them stayed at <20°C over ambient.

Butit seems the IGBTs still hate me... I just did a test run at 120V, <150us on time (don't know exactly what as it was postscaled by the midi volume), the dutycycle limited to 5% and current limit at 300A (though it pretty much never hit that), and even though it worked for probably 5-10mins and got some really nice 40cm arcs, one leg of the drivers committed suicide again.
Before that test I verified the timings and checked for extra switch pulses and found nothing wrong. I haven't yet done an autopsy (will do that tomorrow) so it might theoretically be a fault in the controller (i had a RESISTOR, that was under basically no load, fail earlier :o ), but i doubt it. I did feel the heatsink temp right before it blew up and it was cold to the touch, so even though shock heating would not transfer the heat all that fast, I don't think it was thermal stress. I didn't add any MOVs or TVSs across the IGBTs yet, might that be worth pursuing?
The coil was pulling ~400W-450W when it blew up, and i still had the rod in place so no arcing to the primary was possible.
So was the additional phase shift maybe too much, and the diodes died? the transistors went bang, but are still shorted so maybe that points to them.
But hey, at least my little Midi USB interrupter works  ;D

I also think i didn't post my dimensions yet:

Prim. turns: tuned to around 8-9
Prim tank cap: 5 parallel x 3 series 150nF Cornell Dubilier film caps (940C16P15K-F)
Prim. dimensions: 30cm dia, 20cm high (10 turns), 33cm off the ground
Secondary turns: ~2100 (didn't count while winding it so this is resistively measured but fits to the resonance frequency calculated by java tc)
Secondary dimensions: 16cm dia, 90cm high, 37cm off the ground.
Toroid (not the final one): 15cm minor dia, 60cm major dia, center height ~140cm (right on top of the secondary), made from aluminium air ducting.
    This can be improved, the air duct is edgy as hell, and I think it is a little small (i'd like around 85KHz Fres). Once my driver&controller works i'll probably make one from multiple pipes)
Fres according to javaTc: 94.5Khz measured 94.3KHz (i adjusted the secondary turns until these matched)
Coupling K: 0.153

Also i got the biggest arcs with the primary tuned to ~84KHz, is the deviation from the measured secondary resonant frequency normal or did i mis-tune?
Also also should i maybe continue my other thread, as this issue doesn't really have anything to do with the CT anymore...
« Last Edit: June 18, 2020, 12:32:45 AM by TmaxElectronics »

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 465
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Current Sense transformer trouble
« Reply #27 on: June 18, 2020, 04:38:11 AM »
84kHz isn't surprising.  Arcs add capacitance to the secondary, lowering its resonant frequency.  My DRSSTC has a 93kHz unloaded secondary resonance, with primary at 80kHz.  (I've been experimenting with 90kHz down to 80kHz, and will shortly be going lower to get longer arcs.)

It's always difficult to figure out why electronics fail.  My first guess is that there was still occasional erroneous gate signals switching at high current points.  Second guess would be poor thermal contact to the heat sink as you discussed.  Smooth flat uncontaminated (by dust or particles) surfaces with good thermal compound and appropriate mounting force are key.  TVS diodes can help, but with your nice low-inductance VBus interconnect, they shouldn't be necessary.  (Or, did you say somewhere that the VBus film capacitors didn't all fit?)  I doubt that the diodes are an issue, other than a secondary issue if gate signal timing is erroneous.  100ns should be fine for turn-on phase lead.

For monitoring thermals, my favorite tool is a small thermal camera.  Cost won't fit with everyone's budget.  Mine was $300.  It is wonderful for viewing an entire circuit to find what's hottest, Then checking the temperature of that part.  Use one at work regularly too.  With a close-up lens, I can see which bank of an FPGA got fried by the location of the hot-spot on top of the package.  (I also remember when the cheapest thermal cameras got down to $60,000 and we purchased one at work.)

To further minimize noise, I'd suggest more common-mode chokes.  Ferrite E-cores work fine - the ones designed for switching power supplies.  Make cables longer, then wrap a few turns around the core, and close the path with an I-core or another E-core.  Or loop the cable a few turns through a large ferrite bead or toroid.  Wrap the entire cable including any grounds or shields.
David Knierim

Offline TMaxElectronics

  • High Voltage Experimenter
  • **
  • Posts: 73
  • Karma: +2/-0
    • View Profile
    • My random (and very empty) electronics blog
Re: Current Sense transformer trouble
« Reply #28 on: June 18, 2020, 12:03:06 PM »
Quote
It's always difficult to figure out why electronics fail.
It sure is... Especially if each failure costs quite a bit of money ::)

I guess I'll have to set up my scope with probes on every important signal and in normal trigger mode (maybe even record the screen) and wait for another death. So i can see if any issues occur at higher power levels.
The thermal contact was probably fine though, the surface was flat, all of the burrs where removed and I had plenty of thermal grease on it.

Also I just thought about how weird it is that in the last two failures, only two of the four transistors died. If it was an issue with the switching signal both should die i think, as the signal or them is the same just at 180°. So maybe it is the gate driver dying again and not discharging one gate before turning on another? Or I'm just reading too much into this, and the other ones were only close to death :P
EDIT: So i just took the bridge apart and the Gate drivers are fine, but the gate resistor of the transistor that blew up is open. So now I'm wondering if that might be the cause, or if it died when the transistor blew. It is only a 1/4W resistor but has a 2A pulse rating, so i think it died with the IGBT.

Quote
For monitoring thermals, my favorite tool is a small thermal camera.
I'd love to have a thermal camera, but that just does not fit into my budget ;)

Oh and i continued the other thread, the issues fit there a lot better I think.
Thanks for the help btw davekni :D
« Last Edit: June 18, 2020, 07:15:41 PM by TmaxElectronics »

High Voltage Forum

Re: Current Sense transformer trouble
« Reply #28 on: June 18, 2020, 12:03:06 PM »

 


* Recent Topics and Posts

post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Steve Ward
August 08, 2020, 10:22:04 PM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
T3sl4co1l
August 08, 2020, 03:36:40 PM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
acmq
August 08, 2020, 02:07:52 AM
post Re: LSI Nitrogen Laser tear-down and Homemade TEA Nitrogen and Dye Lasers
[Light, Lasers and Optics]
LesWright
August 06, 2020, 10:20:12 PM
post Re: LSI Nitrogen Laser tear-down and Homemade TEA Nitrogen and Dye Lasers
[Light, Lasers and Optics]
Weston
August 06, 2020, 08:45:59 PM
post Re: Cathodes in Compact Fluorescent Lamps
[Light, Lasers and Optics]
klugesmith
August 06, 2020, 05:41:04 PM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
Jun
August 06, 2020, 03:53:58 PM
post LSI Nitrogen Laser tear-down and Homemade TEA Nitrogen and Dye Lasers
[Light, Lasers and Optics]
LesWright
August 06, 2020, 03:28:14 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
AstRii
August 06, 2020, 12:59:03 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Mads Barnkob
August 06, 2020, 10:51:43 AM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
UB_Serpent
August 06, 2020, 08:40:57 AM
post Re: compact midi interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
TMaxElectronics
August 06, 2020, 03:22:38 AM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
davekni
August 05, 2020, 11:48:55 PM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
davekni
August 05, 2020, 11:37:13 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
AstRii
August 05, 2020, 08:01:11 PM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Lightning On Demand
August 05, 2020, 06:34:08 PM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
Uspring
August 05, 2020, 12:56:32 PM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Mads Barnkob
August 05, 2020, 12:55:52 PM
post Re: Bipolar high voltage multiplier project - more than 70cm hot arc discharge
[Voltage Multipliers]
Teravolt
August 04, 2020, 09:21:09 PM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
johnf
August 04, 2020, 09:16:55 PM
post Re: Does anyone know how this site works.
[General Chat]
thedoc298
August 04, 2020, 05:53:03 PM
post Re: Does anyone know how this site works.
[General Chat]
thedoc298
August 04, 2020, 05:51:03 PM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Lightning On Demand
August 04, 2020, 05:17:26 PM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Max
August 04, 2020, 02:14:56 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
August 04, 2020, 02:08:07 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
futurist
August 04, 2020, 10:48:35 AM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
UB_Serpent
August 04, 2020, 10:02:39 AM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
Jun
August 04, 2020, 08:47:07 AM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
Jun
August 04, 2020, 08:43:00 AM
post Re: Does anyone know how this site works.
[General Chat]
Mads Barnkob
August 04, 2020, 08:33:47 AM
post Re: Does anyone know how this site works.
[General Chat]
MRMILSTAR
August 04, 2020, 05:09:26 AM
post Does anyone know how this site works.
[General Chat]
thedoc298
August 04, 2020, 01:54:49 AM
post Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Lightning On Demand
August 03, 2020, 11:55:41 PM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
Teravolt
August 03, 2020, 09:17:40 PM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
MRMILSTAR
August 03, 2020, 08:30:38 PM
post Re: Budget USB cameras
[High Speed Filming]
missoum0514
August 03, 2020, 06:56:08 PM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
Jun
August 03, 2020, 05:23:15 PM
post Re: Is my xray transformer a ferrite core and how would I drive it?
[Transformer (Ferrite Core)]
nielsquake
August 03, 2020, 05:02:26 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
August 03, 2020, 04:31:04 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
futurist
August 03, 2020, 10:05:59 AM
post Re: Is my xray transformer a ferrite core and how would I drive it?
[Transformer (Ferrite Core)]
Mads Barnkob
August 03, 2020, 08:21:26 AM
post Re: Is my xray transformer a ferrite core and how would I drive it?
[Transformer (Ferrite Core)]
nielsquake
August 02, 2020, 11:45:38 PM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
davekni
August 02, 2020, 08:10:47 PM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
Uspring
August 02, 2020, 07:27:33 PM
post Re: Corona lockdown thread, tell about your situation
[General Chat]
Uspring
August 02, 2020, 07:14:17 PM
post Re: Is my xray transformer a ferrite core and how would I drive it?
[Transformer (Ferrite Core)]
nielsquake
August 02, 2020, 06:58:34 PM
post Re: current limiting to parallel resonance circuit
[Beginners]
davekni
August 02, 2020, 06:49:59 PM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
Jun
August 02, 2020, 05:10:12 PM
post Re: Dual MOT tesla coil
[Spark Gap Tesla Coils (SGTC)]
MRMILSTAR
August 02, 2020, 04:55:14 PM
post Re: How much is the voltage used in Corona Discharge tubes for ozone Generation?
[Transformer (Ferrite Core)]
UB_Serpent
August 02, 2020, 04:20:13 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