Author Topic: Strange gdt waveforms  (Read 537 times)

Offline Marty

  • High Voltage Enthusiast
  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
Strange gdt waveforms
« on: February 24, 2019, 04:08:35 PM »
Hi everyone,

My GDT is behaving very strangely, maybe I'm doing something wrong or maybe i have bad core material.
The core material is 3E25, AI=6900nH, dimensions are: 57.8 x 40.5 x 17.9 mm, cross section Ae=152.4 mm2, 14 turns of CAT5 wire)
When I plug my signal generator into GDT, this is the output on the secondary side (i think its not bad, maybe i should add some turns):


When I connect GDT into my driver (UD2.7C THT, which is working correctly), it behaves strangely. Gdt output with no load looks like this:


With 4,7 Ohms resistor across secondary, it looks like this:


What could be wrong ? Or should I try to connect it to my CM300s ?
I would be thankful for every information or help, thank you :)

Offline Mads Barnkob

  • Administrator
  • High Voltage Senior
  • *****
  • Posts: 924
  • Karma: +11/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Strange gdt waveforms
« Reply #1 on: February 24, 2019, 07:35:58 PM »
Funky waveforms you got there, it is not looking that bad except for the imposed roller coaster effect, but you properly already know about Richies GDT guide? http://www.richieburnett.co.uk/temp/gdt/gdt2.html

About the roller coaster effect, have you checked the power supply rails for the GDT driving voltage? It might be sagging due to lack of near-by filtering caps of the GDT drivers/MOSFETs?

Get those CM300 gates on it, that gives you a much better picture of how the gate waveform really is, with the resistor on the secondary you are missing the capacitive part of driving the gate.
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

Offline Max

  • High Voltage Enthusiast
  • *
  • Posts: 21
  • Karma: +2/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #2 on: February 25, 2019, 01:11:19 PM »
To me it looks like if you're driving your GDT with a signal ranging from 0V to 24V, instead of one from -24V to 24V. At first the GDT is able to maintain that offset (beginning of the waveform), but very soon it saturates and your signal starts walking around. Eventually the DC offset on the output would disappear. The more load you put on it, the faster this happens, as you can see on your measurement with a resistor.
This also explains the amplitude of 12V instead of 24V.

Hook up your two oscilloscope channels to the two inputs of the GDT. My guess would be that one of them is constantly 0V or constantly 24V. Anyway, I'm pretty sure this is not a problem of the GDT itself, rather how it's connected (one input wired to GND?) or which signal it's receiving.


Kind regards,
Max
« Last Edit: February 25, 2019, 01:13:05 PM by Max »

Offline Teravolt

  • High Voltage Experimenter
  • **
  • Posts: 69
  • Karma: +1/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #3 on: February 25, 2019, 03:04:12 PM »
I'm with Max put 2 chanels one on the input of the transformer and one on the out and compare it and poast it. use the o-scope to see if your 24v is sagging and make shure all your ground refrances are common.

Offline Marty

  • High Voltage Enthusiast
  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #4 on: February 26, 2019, 10:04:08 PM »
Thank you for your responses.
Yes, you were right. The driver output is only +24 and the other side is 0.
Its connected correctly, no wire to GND.
I checked all 4 mosfets and they are ok, so I think UCC27423 could be the problem. I will buy new one and post my results.

Offline T3sl4co1l

  • High Voltage Enthusiast
  • *
  • Posts: 40
  • Karma: +4/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #5 on: February 27, 2019, 02:55:22 AM »
That looks like precisely what should happen, for a certain circuit.  The problem is the choice of circuit and/or component values. :)

Let's see... 14 turns is 1.35mH, the LF resonance is about.... 2.5kHz?  There's some saturation in there shortening it.  Looks like you have about 3uF coupling capacitance?

The problem is three-ish things:
1. You're asking for more flux than the transformer can handle, causing it to saturate.  This causes the sharp decline around div 4 (from the left side of the screen).  Fix is to use shorter bursts, or more turns.  Also an option to use lower voltages, and some kind of receiver on the high side (which can be just another gate driver -- TTL levels are a far sight easier to transform than full 15V+ gate signals are!).
2. The LF cutoff must be well below the burst duration.  You don't want the square wave to sag much during a burst, so you need the cutoff that many times slower.  250us burst implies a time constant of maybe 1.25ms or more, which means more like 10x the coupling capacitance.
3. The LF cutoff should also be well damped.  I'm guessing the ~3uF is a single ceramic capacitor and that's it?  (Or any parallel equivalent, or film or aluminum polymer, but probably not tantalum or electrolytic.)  The cap should have ESR equivalent to sqrt(L/C), for magnetizing inductance L (~1.35mH) and coupling capacitance C (~30uF), or 6.7 ohms.  This does drop signal which is unfortunate, but signal loss can also be minimized by damping the capacitor with a bigger, lossier capacitor.  In that case, probably 22uF would be fine, and you'd put double or more in parallel with it, with the same ESR.  Probably an electrolytic of 47-100uF 16V would be a damn close fit, in terms of typical ESR.  (But if you want to be sure, use a lower ESR cap and add the remaining resistance yourself.  Electrolytics are notoriously weak on their ESR specs, after all.)

Mind that I've done these calculations for 1.35mH, but if you need more turns for the flux, the inductance goes up accordingly.  Which means C doesn't need to be quite as large, and the damping capacitance and resistance also change accordingly.

Also note that, if you're using this for direct transformer coupled gate drive, you probably want a pretty low impedance.  Added winding length hinders that (the winding length is directly proportional to its equivalent leakage inductance).  This gives all the more reason to use multiple pairs (of primary and secondary) in parallel.  Particularly good is the "star quad" configuration (four wires twisted together, giving a square cross section, connecting opposite corners of the square in parallel for each winding).  If still lower impedances are needed (i.e., < 25 ohms or so), more pairs/quads will need to be connected in parallel.

By the way, transformer impedance has everything to do with the wire cross section -- the shape, not the size.  You can save space by using, say, wirewrap wire, or even enameled wire if it'll have enough insulation value for your inverter.

Tim
« Last Edit: February 27, 2019, 03:00:10 AM by T3sl4co1l »

Offline Max

  • High Voltage Enthusiast
  • *
  • Posts: 21
  • Karma: +2/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #6 on: February 27, 2019, 09:56:45 PM »
@T3sl4co1l:
Never saw a GDT in combination with direct gate driving. A transformer is a pretty complicated way to transmit logic signals isolated. Usually you use an optocoupler which also removes the need for a symetric signal.
Anyway, GDTs normally are a cheap and simple way to drive gates. No need for a powerful driving circuit and isolated power supply (which supports the high floating voltages) on each gate.
Also most people have no problem to transform +/-20V signals.

Considering the sagging of the signal, I'm with Mads recommendation to follow Richie Burnett's excellent GDT Waveform guide

@Marty:
Before buying "randomly" new components, continue measuring with your oscilloscope to find the real culprit. What if no logic signal reaches that UCC input?
Simply follow the signal path until you find the component where the signal goes in, but doesn't come out ;)


Kind regards,
Max

Offline T3sl4co1l

  • High Voltage Enthusiast
  • *
  • Posts: 40
  • Karma: +4/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #7 on: March 01, 2019, 07:43:55 AM »
Ahh, so hey, this is the circuit..? http://loneoceans.com/labs/ud27/UD27Cschematic.png

Hah, not bad at all, and there you have it, just a component value change will do.  C21 and C30 are 2.2uF, in parallel with an R+C of less, not more!  No wonder it's bouncing.  Increase C24 and C31 to 4.7-10uF, and probably decrease R19 and R24 accordingly.  Still won't fix the saturation, for which more turns are needed.

@T3sl4co1l:
Never saw a GDT in combination with direct gate driving. A transformer is a pretty complicated way to transmit logic signals isolated. Usually you use an optocoupler which also removes the need for a symetric signal.

Not in SSTCs, no.  Drivers are often seen in combination with GDTs in commercial products.  Usually just to boost the falling edge a bit, or in industrial drives where the cost of a full signal+power isolator is justified by the high drive requirements of the (usually) IGBTs used.

In a recent project, I did a bit of both, for a couple reasons:
1. Transformer coupling is easier to control, in terms of isolation, total bandwidth and propagation delay (all aspects can be designed in, no testing needed).  Pure digital isolators (internally, capacitor or transformer coupled) have fast enough propagation delay, but I don't have any on hand.
2. I set a high switching frequency, for which direct drivers are almost obligatory.  (In fact I used MOSFET drivers followed by complementary emitter followers, because the drivers I had on hand were rather weak.)
3. I tried with bootstrap power at first, but this fell over fairly quickly (UF4007 really isn't up to that kind of service), so I plugged in a DC-DC module I made years ago.  Handy!

For a lot of SSTC applications, a bootstrap driver like IR2186S would do very nicely indeed.  Propagation is a bit sluggish, but more than adequate for low-100s kHz application.  Quite a bit cheaper than a GDT, and quite a bit simpler than an optoisolator circuit!

Tim

Offline Marty

  • High Voltage Enthusiast
  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #8 on: March 01, 2019, 04:48:39 PM »
@T3sl4co1l: No, its not that circuit. Its UD2.7C THT version (its written above second picture).
Here : * ud27-schematic.pdf

Next week I will post my results, thank you.

Offline T3sl4co1l

  • High Voltage Enthusiast
  • *
  • Posts: 40
  • Karma: +4/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #9 on: March 02, 2019, 05:03:49 PM »
Thanks. Same output circuit, so the same analysis applies. :)

Tim

Offline Marty

  • High Voltage Enthusiast
  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #10 on: March 09, 2019, 04:24:45 PM »
It is working !
Problem was with LT1016 ultra fast compator. It has output Q and Q' (with negation). Q works normally, but Q' not. There was square wave, but only 1.92V. This goes into 74HC08 (AND gate), so I think it was not enough to be logical 1. And thats why UCC27423 wasn't switching two mosfets.  So I put 7404 invertor on Q and replace that Q'. And it works ! Update: On 74HC14 (hex schmidt invertor) there is one invertor unconnected, so I don't need another IO after all.


This is the driver output.

Today I'm going to test my IGBTs with some bus voltage and load to see real gate waveforms. Thanks for your advice,

Offline Marty

  • High Voltage Enthusiast
  • *
  • Posts: 7
  • Karma: +0/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #11 on: March 18, 2019, 05:32:43 AM »
Sorry for double post but...

Its f***ing working !!!! my first DRSSTC.. after hundreds of hours and a lot of efort :D



1.8m long spark, OCD set to only 500A, 230VAC, CM300s are cold, but the coil need still tune a bit.

Thank you guys
« Last Edit: March 18, 2019, 02:31:31 PM by Marty »

Offline FilipŠebík

  • High Voltage Enthusiast
  • *
  • Posts: 28
  • Karma: +0/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #12 on: March 18, 2019, 12:22:58 PM »
Sorry for double post but...

Its f***ing working !!!! my first DRSSTC.. after hundreds of hours and a lot of efort :D



1.8m long spark, OCD set to only 500A, 230VAC, CM300s are cold, but the coil need still tune a bit.

Thank you guys
Glad to see it working and good job, but I would say to change the underscored word that you used. The word begins with F and I don't think that you can use it here :D so change it
« Last Edit: March 19, 2019, 12:11:30 PM by Mads Barnkob »

Offline Hydron

  • Administrator
  • High Voltage Engineer
  • *****
  • Posts: 216
  • Karma: +4/-0
    • View Profile
Re: Strange gdt waveforms
« Reply #13 on: March 18, 2019, 02:27:35 PM »
I personally don't have an issue with it, especially in this context - I certainly understand the emotion of getting a long and tricky project going!. If anyone feels differently though then please feel free to get in touch with Mads about it.

As for the coil - looks like it's working quite well if you're getting that sort of performance 8)

Offline Mads Barnkob

  • Administrator
  • High Voltage Senior
  • *****
  • Posts: 924
  • Karma: +11/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Strange gdt waveforms
« Reply #14 on: March 19, 2019, 12:15:22 PM »
Congratulations on making your first long sparks! You are now in the exclusive club of people that have sparked longer than their own body height! (I assume you are around 180cm) :)

Swearing depends on the context and generally I got no problems with it, I just want people to understand it paints a forever online picture of themselves what they leave written :) Google, indexing and archives work faster than we think and some day you are held accountable for it.

Do not worry about double posting as you are not breaking the rules, there is exceptions for the exact reason for your post.

Quote
3. Avoid double-posting, double posting is posting twice in a row in a thread, without allowing any replies from other users in between. Double posting to bump a topic to the top to get it marked as new is not allowed. Every topic deserves a chance. If yours slips into obscurity, its chance came and went. If 8 hours have elapsed, and you have some new text information to add, you may double post. If there is much information, a large update with many lines of text and many attachments, it is allowed to double post to avoid editing long and complex posts with many attachments. This is the only two exceptions, there are no other exceptions! Posting an identical or nearly identical question in two individual threads is also considered double posting.
http://www.kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics

High Voltage Forum

Re: Strange gdt waveforms
« Reply #14 on: March 19, 2019, 12:15:22 PM »

 


* Recent Topics and Posts

post Re: GDT Experiments
[Solid state Tesla coils]
Teravolt
March 20, 2019, 11:15:50 PM
post Re: GDT Experiments
[Solid state Tesla coils]
Thunderstruck
March 20, 2019, 09:07:04 PM
post Re: Keep or throw away ?
[Beginners]
Teravolt
March 20, 2019, 05:18:07 PM
post Re: New small SSTC
[Solid state Tesla coils]
Bambinz
March 20, 2019, 03:52:15 PM
post Re: New small SSTC
[Solid state Tesla coils]
Mads Barnkob
March 20, 2019, 01:23:52 PM
post Re: GDT Experiments
[Solid state Tesla coils]
T3sl4co1l
March 20, 2019, 01:18:29 PM
post Re: GDT Experiments
[Solid state Tesla coils]
Mads Barnkob
March 20, 2019, 01:06:24 PM
post Re: cute-qcw Tesla Coil
[Dual Resonant Solid State Tesla coils]
Mads Barnkob
March 20, 2019, 01:00:46 PM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic circuits]
Mads Barnkob
March 20, 2019, 12:41:22 PM
post Re: Keep or throw away ?
[Beginners]
Thunderstruck
March 19, 2019, 10:02:41 PM
post Re: Keep or throw away ?
[Beginners]
Hydron
March 19, 2019, 08:14:33 PM
post Re: Keep or throw away ?
[Beginners]
Teravolt
March 19, 2019, 05:12:03 PM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General chatting]
Experimentonomen
March 19, 2019, 01:32:32 PM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General chatting]
Mads Barnkob
March 19, 2019, 12:20:34 PM
post Re: Strange gdt waveforms
[Dual Resonant Solid State Tesla coils]
Mads Barnkob
March 19, 2019, 12:15:22 PM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
Mads Barnkob
March 19, 2019, 10:46:50 AM
post Re: Keep or throw away ?
[Beginners]
Thunderstruck
March 19, 2019, 09:11:09 AM
post Re: Keep or throw away ?
[Beginners]
Mads Barnkob
March 19, 2019, 08:24:31 AM
post Re: Keep or throw away ?
[Beginners]
Teravolt
March 19, 2019, 01:51:51 AM
post Re: Keep or throw away ?
[Beginners]
Thunderstruck
March 18, 2019, 09:56:42 PM
post Re: cute-qcw Tesla Coil
[Dual Resonant Solid State Tesla coils]
Weston
March 18, 2019, 07:55:52 PM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic circuits]
petespaco
March 18, 2019, 06:11:07 PM
post Re: Strange gdt waveforms
[Dual Resonant Solid State Tesla coils]
Hydron
March 18, 2019, 02:27:35 PM
post Re: cute-qcw Tesla Coil
[Dual Resonant Solid State Tesla coils]
Uspring
March 18, 2019, 12:52:35 PM
post Re: Strange gdt waveforms
[Dual Resonant Solid State Tesla coils]
FilipŠebík
March 18, 2019, 12:22:58 PM
post New small SSTC
[Solid state Tesla coils]
Bambinz
March 18, 2019, 11:46:18 AM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General chatting]
Bambinz
March 18, 2019, 11:29:02 AM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic circuits]
Mads Barnkob
March 18, 2019, 10:35:04 AM
post Re: Strange gdt waveforms
[Dual Resonant Solid State Tesla coils]
Marty
March 18, 2019, 05:32:43 AM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General chatting]
Experimentonomen
March 17, 2019, 09:13:47 PM
post Re: DRSSTC Driver Logic
[Dual Resonant Solid State Tesla coils]
Netzpfuscher
March 17, 2019, 04:08:33 PM
post Re: WTB: Large Heatsink with a Baseplate at least 10mm thick
[Sell / Buy / Trade]
Phoenix
March 17, 2019, 02:46:44 PM
post Re: cute-qcw Tesla Coil
[Dual Resonant Solid State Tesla coils]
flyglas
March 17, 2019, 02:21:45 PM
post Re: DRSSTC Driver Logic
[Dual Resonant Solid State Tesla coils]
Laci
March 17, 2019, 01:54:47 PM
post Re: WTB: Large Heatsink with a Baseplate at least 10mm thick
[Sell / Buy / Trade]
Mads Barnkob
March 17, 2019, 10:15:16 AM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
FilipŠebík
March 17, 2019, 09:33:50 AM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
Teravolt
March 17, 2019, 05:32:48 AM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
Thunderstruck
March 17, 2019, 02:39:25 AM
post WTB: Large Heatsink with a Baseplate at least 10mm thick
[Sell / Buy / Trade]
Phoenix
March 16, 2019, 10:06:39 PM
post Re: Large Ferrite Core with IGBT-Brick Driver
[Transformer (ferrite core)]
rikkitikkitavi
March 16, 2019, 08:35:51 PM
post Re: cute-qcw Tesla Coil
[Dual Resonant Solid State Tesla coils]
Weston
March 16, 2019, 08:04:52 PM
post Re: cute-qcw Tesla Coil
[Dual Resonant Solid State Tesla coils]
flyglas
March 16, 2019, 06:44:14 PM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
FilipŠebík
March 16, 2019, 09:57:55 AM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
Thunderstruck
March 16, 2019, 09:40:55 AM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
FilipŠebík
March 16, 2019, 08:57:54 AM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
Thunderstruck
March 16, 2019, 08:46:24 AM
post Re: Which varnish to use to coat the Secondary?
[Solid state Tesla coils]
FilipŠebík
March 16, 2019, 08:27:24 AM
post Re: UD 2.7C Build
[Dual Resonant Solid State Tesla coils]
Thunderstruck
March 16, 2019, 07:35:58 AM
post GDT Experiments
[Solid state Tesla coils]
Thunderstruck
March 16, 2019, 06:32:36 AM
post cute-qcw Tesla Coil
[Dual Resonant Solid State Tesla coils]
Weston
March 16, 2019, 06:01:44 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