Author Topic: UD 2.7 second start pulse  (Read 962 times)

Offline Phoenix

  • High Voltage Experimenter
  • **
  • Posts: 75
  • Karma: +2/-0
    • View Profile
UD 2.7 second start pulse
« on: July 24, 2020, 11:45:24 AM »
Hello

I am currently building a QCWDRSSTC. Here are some images:



The QCW is currently working fine, I just noticed something: Without any feedback signal, there seems to be a second start pulse at the end of the ontime.



This pulse is not there in normal operation, it is only there when there is no feedback or when the bus voltage coming from the buck converter drops to under 10V while the UD 2.7 is still getting an interrupter pulse. Normally this does not happen because i adjusted the reference ramp such that there are always 30V on the bus.

A possible reason for this second impulse might be that the n channel fets on the UD 2.7 output stage are not on between the pulses which would mean that the GDT is floating, but this is not the case. The output stage of the 2.7 is performing as expected, i checked everything with the oscilloscope and there appears to be no problem. The n channel fets are always on between the pulses and pull the gdt low. I also checked the coupling capacitors on the output stage but there is nothing wrong either.

Does anyone have an idea what else might be causing this?

Thanks for reading

Greetings
Phoenix

Edit: I just figured out how to remove the second pulse: I bridged the DC blocking cap that is in series with the 15 ohm output dampening resistor. Now the n channel fets can pull down the gdt directly over the 15 ohm resistor. Loneoceans only added this cap in order to prevent the 15 ohm resistor from burning in case the UD 2.7 gets permanently turned on. It is not needed if you make sure to never turn on the UD 2.7 permanently. The cap is also not there on the original 1.3b driver.


I know the starting pulse is ugly but that does not matter because the coil will be running from feedback way before the ugly parts appear. If i make the 15 ohm resistor smaller the pulse also becomes better because it cant ring as much if it is coupled tighter.


« Last Edit: July 24, 2020, 02:50:44 PM by Phoenix »

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 470
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: UD 2.7 second start pulse
« Reply #1 on: July 24, 2020, 09:58:37 PM »
Impressive coil!  Would love to see schematics and operating scope traces of bridge current and voltage etc.

Concerning UD2.7 behavior without feedback:  The gate drive is not designed to handle DC.   Your 700us enable pulse is effectively DC relative to time constants of the circuit.  The initial ~20us pulse you see is likely due to GDT core saturation.  It starts with the lead enable pulse edge and ends when the core saturates.  By 200us or so into the enable pulse, the PFETs of the gate-drive have lost voltage due to being AC-coupled from the UCC27423 (0.1uF into 1k ohms, ~2 time constants).  I'm guessing that the half-cycle of resonance during that 200us is GDT inductance with the 2.2uF coupling capacitors.  Before shorting out the cap in series with 15-ohms, those caps are left charged after the initial 200us when the PFETs have lost gate drive.  At the end of the enable pulse, both NFETs are on, discharging the 2.2+1uF caps into the GDT, creating the second pulse.

Your second scope trace doesn't go out to 700us.  Was that case a shorter enable pulse?  Have you tested with <100us enable pulses?  That may still show a second pulse at the end of enable, as the PFET won't have turned off yet, so the 15-ohm resistor won't have been able to discharge the 2.2uF coupling cap.
David Knierim

Offline Phoenix

  • High Voltage Experimenter
  • **
  • Posts: 75
  • Karma: +2/-0
    • View Profile
Re: UD 2.7 second start pulse
« Reply #2 on: July 24, 2020, 11:23:46 PM »
Thank you for your help :)

Your explanation with the charged coupling caps getting discharged into the GDT causing the second pulse makes a lot of sense. The second scope trace was a very long enable pulse (20ms). I just tested it with a very short enable pulse of 100µs and you are right, there is a second pulse. That is luckily not a problem because the QCW will be running with enable times in the region of 2-20ms (I had to change a value in the ramp code first to get down to a 100µs pulse  ;D).

Here is some more information on the QCW itself. The buck converter is based on loneoceans buck converter using an ultrafast opamp configured as a schmidt trigger in order to utilize bang bang control. Here is the basic schematic: https://www.loneoceans.com/labs/qcw/buckdriver1.png

I changed a few things on the schematic. I replaced the rather weak opto gatedriver with a proper 30A IXD630 gatedriver and an optocoupler. I also changed the supply for the gatedriver. Instead of utilizing a bipolar supply which can cause problems I simply placed 5V onto the emitter of the buck IGBT. When the gatedriver is off, there will be -5V referenced to the gate and when it is on there will be 19V (24-5V) on the gate. I also made the hysteresis span of the schmidt trigger wider to make the buck run more stable (too little hysteresis causes interference problems). In order to compensate for the lower switching frequency caused by the bigger hysteresis i decreased the value of R11.  I also designed my own PCB for the buck controller. 
Here you can see all the PCBs:


Here is the bus voltage on the bridge where you can see the ramp:


The buck currently uses an Infineon FF400 Igbt module and the fullbridge uses Ixys IXXN110N65C4H1 minibricks.
« Last Edit: July 25, 2020, 12:53:00 AM by Phoenix »

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 470
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: UD 2.7 second start pulse
« Reply #3 on: July 25, 2020, 04:57:54 AM »
Is there a 10x factor on the buck converter output scope trace?  The scope trace looks like a ramp to 45V, while 450V seems more likely.

What frequency is your QCW running?  How much does frequency change during the ramp?  How does primary coil/MMC current change during the ramp?  Do you measure secondary current?  What's the MMC capacitance, primary and secondary coil inductances and coupling factor?  (I've been getting more intrigued by QCW, so am interested in examples.  Especially looking at designs that are farther from conventional to learn where boundaries are.  May eventually explore some boundaries myself, especially behavior at lower frequencies and longer ramps.  Wondering if double-ramp-time might compensate for half-frequency, for example.)
David Knierim

Offline Phoenix

  • High Voltage Experimenter
  • **
  • Posts: 75
  • Karma: +2/-0
    • View Profile
Re: UD 2.7 second start pulse
« Reply #4 on: July 25, 2020, 09:38:23 AM »
The scope shot of the ramp was during low voltage testing, normally the ramp goes to about 310V. My buck capacitors are 350V rated and I have a total capacitance of 30000uF.

The system is running at ~370kHz. The frequency changes quite a lot because of sparkloading, I have seen frequency drops down to a little below 300kHz. The primary current aproximately starts at half of the maximum primary current and the peak current is where the ramp has the maximum voltage. No, i have never measured the secondary current. My MMC capacitance is 16.5nF currently. The coupling factor is 0.4k. The inductance of the primary is ~11uH. My secondary has 600 turns of 0.4mm with dimensions of 125×250mm.

Yes i have also thought about the relation between ramp duration and system frequency. With a longer ramp and lower frequency you can still get sword sparks, altough I don't think that this relation is linear where you can say that double the ramp time and half the frequency will still give sword sparks. I think you have to increase the ramp duration way more than just doubling in order to get sword sparks at half the frequency.

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 470
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: UD 2.7 second start pulse
« Reply #5 on: July 25, 2020, 06:33:00 PM »
Thank you for the parameters!  I'd guessed a lower frequency just looking at your images.  It's nice to see that fast IGBTs can work well at 370kHz under ZCS conditions.

Using UD2.7, I presume oscillation starts at the frequency with the strongest current signal.  Do you have primary tuned slightly higher frequency than secondary to make it start at the upper resonant frequency?  I read that QCW coils run at the upper pole, but don't know if that's universal.
David Knierim

Offline Phoenix

  • High Voltage Experimenter
  • **
  • Posts: 75
  • Karma: +2/-0
    • View Profile
Re: UD 2.7 second start pulse
« Reply #6 on: July 27, 2020, 03:04:56 PM »
Yes exactly, my primary is tuned a bit below the upper pole frequency but above the resonant frequency of the secondary. For a typical QCW this is probably the best choice. I was pretty lucky with my calculations, I got about 2m sparks with my initial tap point.

There is a trick how to get longer sparks. The primary gets tuned to the lower pole but at the beginning of the pulse you feed in the upper pole frequency into the feedback of the UD with a startup oscillator. This leads the UD to switch with the upper pole frequency while the primary is tuned to the lower pole. This makes it possible to "track" the optimal frequency during the whole discharge event (the secondary resonant frequency drops during the discharge event because of spark loading). I haven't tried this trick yet but i heard that it boosts the performance quite a bit.

Offline davekni

  • High Voltage Engineer
  • ****
  • Posts: 470
  • Karma: +26/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: UD 2.7 second start pulse
« Reply #7 on: July 27, 2020, 06:19:44 PM »
Great, that makes sense.  I'd read about that "trick" of forcing upper-pole operation with lower tuning.  I was unsure if that was a necessary feature of QCW coils or just a performance booster.  You clearly show good performance without forced startup.  If you ever decide to add forced-startup (with lower primary tuning), I'd love to see how much improvement is obtained.
David Knierim

High Voltage Forum

Re: UD 2.7 second start pulse
« Reply #7 on: July 27, 2020, 06:19:44 PM »

 


* Recent Topics and Posts

post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
AstRii
Today at 01:04:06 AM
post Re: Moving to a new lab!
[Laboratories, Equipment and Tools]
profdc9
Today at 12:55:12 AM
post Re: Air Gap: why and how much?
[Transformer (Ferrite Core)]
profdc9
August 13, 2020, 10:59:27 PM
post Re: How about some Tesla Coil memes
[General Chat]
TMaxElectronics
August 13, 2020, 09:55:27 PM
post Re: Air Gap: why and how much?
[Transformer (Ferrite Core)]
johnf
August 13, 2020, 09:45:58 PM
post Re: Air Gap: why and how much?
[Transformer (Ferrite Core)]
klugesmith
August 13, 2020, 08:26:56 PM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
August 13, 2020, 07:02:24 PM
post Air Gap: why and how much?
[Transformer (Ferrite Core)]
HighVoltageRulezz
August 13, 2020, 06:23:24 PM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
HighVoltageRulezz
August 13, 2020, 06:15:16 PM
post Re: How about some Tesla Coil memes
[General Chat]
AstRii
August 13, 2020, 05:26:47 PM
post Re: First DRSSTC, No Sparks on Breakout
[Dual Resonant Solid State Tesla coils (DRSSTC)]
TMaxElectronics
August 13, 2020, 05:25:43 PM
post Re: How about some Tesla Coil memes
[General Chat]
TMaxElectronics
August 13, 2020, 05:23:40 PM
post Re: First DRSSTC, No Sparks on Breakout
[Dual Resonant Solid State Tesla coils (DRSSTC)]
AstRii
August 13, 2020, 05:23:37 PM
post Re: Bipolar high voltage multiplier project - more than 70cm hot arc discharge
[Voltage Multipliers]
kilovolt
August 13, 2020, 01:44:24 PM
post How about some Tesla Coil memes
[General Chat]
AstRii
August 13, 2020, 01:09:16 PM
post First DRSSTC, No Sparks on Breakout
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Yak
August 13, 2020, 12:41:06 PM
post Re: X-ray image intensifier
[X-ray]
AlexanderHun
August 13, 2020, 11:01:46 AM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
AstRii
August 13, 2020, 02:09:13 AM
post Re: Charging a LiPo Battery with an LED driver
[Electronic Circuits]
davekni
August 12, 2020, 11:08:35 PM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
davekni
August 12, 2020, 10:49:52 PM
post Re: Moving to a new lab!
[Laboratories, Equipment and Tools]
Mads Barnkob
August 12, 2020, 10:30:54 PM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
Teravolt
August 12, 2020, 09:00:50 PM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
SteveN87
August 12, 2020, 07:16:43 PM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
Teravolt
August 12, 2020, 02:54:33 PM
post Re: Charging a LiPo Battery with an LED driver
[Electronic Circuits]
TMaxElectronics
August 12, 2020, 12:42:50 PM
post Re: Charging a LiPo Battery with an LED driver
[Electronic Circuits]
Twospoons
August 12, 2020, 12:23:58 PM
post Re: Single sided use of Xray XFMR
[Transformer (Ferrite Core)]
HighVoltageRulezz
August 12, 2020, 11:15:07 AM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
davekni
August 12, 2020, 04:41:24 AM
post Re: Charging a LiPo Battery with an LED driver
[Electronic Circuits]
davekni
August 12, 2020, 04:06:09 AM
post Re: Charging a LiPo Battery with an LED driver
[Electronic Circuits]
TMaxElectronics
August 12, 2020, 01:58:47 AM
post Re: Charging a LiPo Battery with an LED driver
[Electronic Circuits]
Twospoons
August 12, 2020, 12:56:02 AM
post Charging a LiPo Battery with an LED driver
[Electronic Circuits]
TMaxElectronics
August 11, 2020, 11:24:18 PM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
Teravolt
August 11, 2020, 06:34:05 PM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Lightning On Demand
August 11, 2020, 04:19:01 AM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Lightning On Demand
August 11, 2020, 04:17:27 AM
post Re: Fun with wireless power
[Solid State Tesla Coils (SSTC)]
Lightning On Demand
August 11, 2020, 04:10:34 AM
post Re: Single sided use of Xray XFMR
[Transformer (Ferrite Core)]
klugesmith
August 11, 2020, 03:33:02 AM
post Single sided use of Xray XFMR
[Transformer (Ferrite Core)]
HighVoltageRulezz
August 10, 2020, 12:23:05 PM
post Re: XRay Tank- any ideas about what is inside
[Transformer (Ferrite Core)]
HighVoltageRulezz
August 10, 2020, 12:18:04 PM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
johnf
August 10, 2020, 10:24:54 AM
post Re: making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
davekni
August 10, 2020, 06:20:04 AM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General Chat]
RoadReaper
August 10, 2020, 05:40:19 AM
post Re: Easy to build Polyphonic MIDI Interrupter
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Max
August 10, 2020, 02:07:26 AM
post making HV bobbins for ferrite transformers
[Transformer (Ferrite Core)]
Teravolt
August 10, 2020, 01:27:22 AM
post A Tesla coil and an antenna
[General Chat]
TheFreq
August 09, 2020, 09:02:32 PM
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

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