Author Topic: Using UCC37322/21 with Steve Ward UD1.3b Driver  (Read 664 times)

Offline WayneSeng

  • High Voltage Enthusiast
  • *
  • Posts: 6
  • Karma: +0/-0
    • View Profile
Using UCC37322/21 with Steve Ward UD1.3b Driver
« on: December 19, 2020, 08:16:47 AM »
Hello everyone. I was wondering if I can use a pair of UCC37322/21 with the UD1.3B. I don't have any P channel MOSFETS on hand or any UCC27423 drivers. Thanks! :)
« Last Edit: December 20, 2020, 05:17:08 AM by WayneSeng »

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1656
  • Karma: +32/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #1 on: December 19, 2020, 11:45:14 AM »
Do you want to replace the entire MOSFET output stage and UCC27423 with a pair of UCC37322/21?

You already have one signal to the two AND-gates for the feedback signal inverted by IC1E. So you have to use either 2x UCC37322 or UCC37321, or remove the inverting gate IC1E and use one of each for the push-pull driving of a GDT directly from those drivers.

You just have to be sure that the UCC3732x can supply enough current to drive your IGBTs. So if depending on the type and voltage rating, I doubt you could drive much larger than a TO-247 bridge. You need to put an extra effort into making a good GDT and experiment with gate resistors for best performance.
https://kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics
https://www.youtube.com/KaizerPowerElectronicsDk60/join - Please consider supporting the forum, websites and youtube channel!

Offline WayneSeng

  • High Voltage Enthusiast
  • *
  • Posts: 6
  • Karma: +0/-0
    • View Profile
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #2 on: December 19, 2020, 09:01:59 PM »
Hi Mads, btw love your youtube channel. I do only plan on driving a half bridge of TO-247 IGBTS (FGH60N60SMD). I have attached an image of the updated schematic (please excuse the terrible drawing). I also plan on using an IFD-95T instead of the HFBR fiber receivers. Ive heard that the HDBR drivers are active low, so you need to invert the signal twice to get it to work. With the IFD-95T, I think you just need to invert the signal once as its active high. Also regarding removing the inverting gate IC1E, I have inverted the signal twice before feeding it into the AND gate. Should I remove IC1E and feed the signal straight into the AND gate through one inverter? Thanks :)

Offline Mads Barnkob

  • Administrator
  • High Voltage Expert
  • *****
  • Posts: 1656
  • Karma: +32/-0
  • Denmark
    • View Profile
    • Kaizer Power Electronics
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #3 on: December 19, 2020, 09:32:31 PM »
Thank you :)

If you are only driving one TO-247 at a time, you should go ahead and try with UCC3732x driving a GDT in push-pull, just by removing IC1E as you showed. I did not check your output stage, but just stake a look at Steve Wards Mini SSTC (SSTC5): http://stevehv.4hv.org/SSTC5/miniSSTCfnlsch.JPG

You can use the IF-D95T instead of HFBR, you only have to take IC1F out to switch the polarity depending on what type of optical receiver is used. You can NOT just delete the AND gate and OCD circuit! Without OCD you are going to blow your IGBTs up.
https://kaizerpowerelectronics.dk - Tesla coils, high voltage, pulse power, audio and general electronics
https://www.youtube.com/KaizerPowerElectronicsDk60/join - Please consider supporting the forum, websites and youtube channel!

Offline davekni

  • High Voltage Senior
  • *****
  • Posts: 960
  • Karma: +42/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #4 on: December 19, 2020, 09:37:19 PM »
As drawn the schematic has an issue.  When disabled, the UCC27231 output will be high and the UCC27232 output low.  Disabled should have both outputs low.

The easiest solution is to use the UCC2723x enable pins, wired to IC3A-5, and remove the AND gates.  Unless you are using the counterfeit Chinese UCC2723x chips that have non-functional enable pins.  If you are using chips with non-functional enables, then use two non-inverting parts (UCC27232) and put IC1E back in the circuit.  (The drivers that have FET buffers after the driver chips don't use enable because they want driver chip outputs to be high when disabled.)

Also, when connecting a GDT directly to driver chips, I'd recommend adding schottky diodes (1n5819 or equiv.) from driver chip outputs to +15V supply.  The GDT magnetization current will attempt to pull driver chip outputs beyond supply rails after each output transition.  I've heard of driver chips frying if not clamped with diodes to supply rails.  Diodes to +15V are most important, but you can also add diodes from outputs to ground.
David Knierim

Offline WayneSeng

  • High Voltage Enthusiast
  • *
  • Posts: 6
  • Karma: +0/-0
    • View Profile
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #5 on: December 19, 2020, 10:23:38 PM »
Thanks davekni. Those counterfeit UCC37322/21 from Aliexpress are a pain. I remember spending hours troubleshooting why they aren't working only to realize that they may be fake. Buying genuine ones on digikey solve all my problems. Anyway, after using the ENABLE pins of the UCC chips and connecting pin 2 directly to the output of the 74hc14 pin 4, I was able to get output!!! The problem is the arcs are quite small even at a high input voltage. Now here is the weird part, when I move the current transformer, I sometimes get no output and sometimes I do. Maybe its interference? Im using a 2000:1 current sense transformer(image attached) and my homemade cascaded transformer does not work at all(image attached). Maybe the core permeability is too low for my cascaded transformer? I used ethernet cable to wind 32 turns on each core and looped one turn through the second core. The cores are the same ones I used to make my gate drive transformer.

Offline davekni

  • High Voltage Senior
  • *****
  • Posts: 960
  • Karma: +42/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #6 on: December 19, 2020, 11:30:32 PM »
Is this an SSTC or DRSSTC?  If DRSSTC, as Mads said, you are at risk of frying IGBTs or FETs without OCD.  Do you have a schematic of the entire design including H-Bridge and bulk and output capacitors, and which IGBTs or FETs are used in the bridge?  Pictures of your system will help too.  Long wires, especially within the H-Bridge, can cause problems.

I can't find good spec's on that DL-Ct1005A, but it may be for 60Hz, not kHz frequencies.  2000:1 in one transformer core makes high-frequency difficult due to winding capacitance.  Your home-built one is likely better, but I can't tell without knowing what core you are using.  A picture of your GDT will help too - how many turns and how are wires paired/twisted.

Do you have an oscilloscope?  Debugging is much easier with scope traces, even if using a cheap old analog scope and photos of the screen.
David Knierim

Offline WayneSeng

  • High Voltage Enthusiast
  • *
  • Posts: 6
  • Karma: +0/-0
    • View Profile
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #7 on: December 20, 2020, 12:57:59 AM »
This is a DRSSTC. I will add OCD once I get the coil working. I have attached some images of my setup as well as some waveforms of the output of the GDT to one of the IGBTS. The core was salvaged out of a power supply. I heard that the green cores are much better as they have a higher permeability. The core shown was also used to make the GDT. Its approximately 20 turns of 3 pairs of magnet wire. I didn't twist the magnet wire when winding the GDT. I simply used 3 separate pieces of magnet wire and wound it on the core. For some reason, my homemade transformer doesn't work. Do I need a burden resistor on the output of my homemade CT? I also have a problem at higher bus voltages. All of a sudden after I get past around 100VDC, the output makes loud banging noises and draws over 1amp from my variac. I think this is due to interference from the primary coil.

Offline davekni

  • High Voltage Senior
  • *****
  • Posts: 960
  • Karma: +42/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Using UCC27322/21 with Steve Ward UD1.3b Driver
« Reply #8 on: December 20, 2020, 04:16:26 AM »
How were those green cores wound in their original power supply use?  If they were common-mode chokes with two separate windings on opposite sides of the core, then it's ferrite and probably good.  If it had windings that extended most or all of the way around the core, then it's likely some form of powdered iron, not good for GDT or CT use.  Color may be an indication for cores from one manufacturer.  Toroid core color is not standardized globally.

Your GDT windings appear to be grouped, so that each winding extends all the way around the toroid adjacent the other two.  That's much better than three separate windings, although twisted pairs are even better.  Best is to use four windings, two twisted pairs.  One half of each pair drives a gate and the remaining two halves are in parallel for the primary.  BTW, your GDT primary leads are likely short enough that this isn't critical, but twisting two wires together for one GDT lead doesn't reduce parasitic inductance significantly.  Twisting the two GDT leads with each other is what helps.

Your GDT output waveform looks fine, so there's probably no need to change that.  It would help to see the gate waveform after the diode/resistor.  Don't push your IGBTs beyond their rated peak current.  Many designs run at higher current, but that generally requires higher gate voltage, usually +-18-24V, achieved with the FETs after the driver chips.

There is a glitch in your driver at the end of the enable pulse.  Enable goes away for 4us, then goes true again for 0.8us, then finally stays off.  It's worth scoping through your driver to see what is causing the glitch.  That one glitch may become more of a problem at different operating conditions.

With that driver design you don't want any additional burden resistor across the CT output.  The input zener diodes provide the burden.  I suggest adding equal high-value resistors in parallel with D5 and D6, 100k-1meg each.  Some versions of this driver use germanium or schottky diodes for D5 and D6, which often have enough leakage current to keep the input to IC1D centered.  With low-leakage 1N4148 diodes, it helps to add resistors across them.  That may be enough to allow operation with your two-core CT, which is likely better than the single DL-CT1005A part.  Remember to try both polarities of connection.

For scoping, it's useful to add one more core to your CT, a copy of the second core that shares the one-turn loop from the first core.  The second core can be eventually for OCD.  In the meantime, add a 10-ohm burden resistor and use it to scope current.  Scoping current and half-bridge output voltage will help.
David Knierim

Offline WayneSeng

  • High Voltage Enthusiast
  • *
  • Posts: 6
  • Karma: +0/-0
    • View Profile
Re: Using UCC37322/21 with Steve Ward UD1.3b Driver
« Reply #9 on: December 20, 2020, 06:46:29 AM »
I remember those green cores had some black tar inside with 2 windings, so I guess it was a common mode choke.

I did some research into GDT's and apparently you can make a 1:2:2 GDT to double the output voltage. In this case, I can use 12V instead of 15V and after the 1:2:2 GDT I can get 24V. Ive never had any experience with 1:2:2 GDT's and don't really know how to make one. Do you have any suggestions?

The glitch seems to be gone after using the cascaded transformer. The problem was actually with the transformer itself and not the driver board. When I was soldering the winding that passes through one of the transformers, I didn't make a good connection. The insulation actually melted the 2 wires together during soldering, so I thought it was making a connection while in reality they weren't connected at all. Now the driver board works with the cascaded transformer. :)

Now I have another problem. At higher bus voltages, I get loud bangs on the output and the interrupt light starts to flicker. At voltages below 100V, everything is working fine. I have grounded all unused inputs of the 74hc14, but the problem persists.

BTW, the original post said I was using UCC27322/21 drivers. That was a typo. I am actually using UCC37322/21 drivers.

Offline davekni

  • High Voltage Senior
  • *****
  • Posts: 960
  • Karma: +42/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Using UCC37322/21 with Steve Ward UD1.3b Driver
« Reply #10 on: December 20, 2020, 08:00:39 PM »
Yes, sounds like the cores are from common-mode chokes, so likely fine.

Where is the "interrupt light"?  Is that on an interrupter board that is feeding the optical fiber to your driver board?  If so, it is a problem with that interrupter, perhaps being too close to the coil and getting confused by stray fields.

If the issue isn't from a separate interrupter, then scope waveforms while running correctly just under 100V, and again when just starting to misbehave just over 100V.  Of course, don't run for long when misbehaving, as that may be stressing parts.

Concerning 1:2:2 GDT, that makes 4x load on driver chips, as impedance goes as turn-ratio squared.  1:1.5:1.5 is even trickier to construct with low leakage inductance, but would be easier on driver chips.  I suggest staying with 1:1:1 until everything is operating correctly within the IGBT's rated current.  Then consider changing if you want to push current higher.
David Knierim

Offline WayneSeng

  • High Voltage Enthusiast
  • *
  • Posts: 6
  • Karma: +0/-0
    • View Profile
Re: Using UCC37322/21 with Steve Ward UD1.3b Driver
« Reply #11 on: December 21, 2020, 06:09:28 AM »
Yay!!! First Light!! ;D After scoping the output of the GDT, realized that the frequency was way too low for my secondary(Around 220khz). After swapping my capacitor 68nf to 44nf, the resonance frequency of the primary was brought up to around 250khz, right where it should be.

The problem with the random banging noises on the output past 100vdc was one of the primary output wires was too close to the GDT. This introduced noise into the driver board and IGBTS, which made the weird noises on the output. After moving the wire far away from the GDT, all the problems with noise have been resolved. I also added a bunch of decoupling capacitors on the driver board to reduce noise on the chips.

The next step is to add OCD and maybe a 1:2:2 or a 1:1.5:1.5 GDT to increase the gate voltage. You said a 1:2:2 GDT makes 4x load on the driver chips. Do you think UCC37322/21 chips can handle this extra current? Do you have any experience using 1:2:2 GDT's with UCC37322/21 drivers or something similar? Thanks :)

Offline davekni

  • High Voltage Senior
  • *****
  • Posts: 960
  • Karma: +42/-0
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Using UCC37322/21 with Steve Ward UD1.3b Driver
« Reply #12 on: December 21, 2020, 10:19:20 PM »
Congratulations on first light!

To be clear, 1:2:2 GDT will make the driver load impedance drop to 1/4th.  Because driver voltage is dropping by 12/15, power won't go up 4x, but rather by 4 * (12V/15V)^2 = 2.56x.  With only a half-bridge, it's possible the drivers will be OK.  Looking at the TI data sheet, thermal numbers seem inconsistent.  Maximum power for the plastic DIP "P" package is listed as only 350mW.  However, thermal resistance numbers are very close to the same as the "DGN" package that is rated for 3W.  Given the low-duty-cycle DRSSTC use, I'm going to guess it will work.  I have no personal experience driving GDTs with driver chips - only with discrete transistors.  If the driver chips get hot, glue a small heat-sink or scrap of aluminum to the top.

For 1:2:2 GDT, I suggest winding with 4 twisted pairs (ie. CAT5 cable).  For a given construction, magnet wire's thinner insulation will reduce leakage inductance.  However, unless you are using double-coated magnet wire, the thin insulation that is easily nicked can cause breakdown between primary and secondary.  So, if you are confident in your magnet wire and careful construction, you can twist pairs of magnet wire.

Once the 4 pairs are wound around the core, connect one wire of each pair to the driver chips.  Four wires in parallel form the primary.  Two of the remaining wires connect in series to one IGBT.  The remain two in series to the other IGBT.  For the series connections, cut the leads short and solder close to the core.

If you have added schottky diodes from driver chip outputs to supply, then you can likely get away with fewer turns on the GDT.  Five is likely enough, ten certainly plenty.  Fewer turns reduces parasitic inductance.  Also reduces parallel inductance, which is why the diodes are needed.

Yes, decoupling caps are great.  Hard to have too many.
« Last Edit: December 21, 2020, 10:24:40 PM by davekni »
David Knierim

High Voltage Forum

Re: Using UCC37322/21 with Steve Ward UD1.3b Driver
« Reply #12 on: December 21, 2020, 10:19:20 PM »

 


* Recent Topics and Posts

post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
February 27, 2021, 11:30:47 PM
post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ChrisBlis
February 27, 2021, 10:44:59 PM
post Re: Rocket engine questions
[General Chat]
Twospoons
February 27, 2021, 10:41:00 PM
post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
February 27, 2021, 10:32:43 PM
post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ChrisBlis
February 27, 2021, 09:59:59 PM
post Re: GDT resources
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
February 27, 2021, 09:55:40 PM
post Re: Gate drive push pull(FDS8958) heating up
[Solid State Tesla Coils (SSTC)]
davekni
February 27, 2021, 09:18:10 PM
post Re: Rocket engine questions
[General Chat]
plasma
February 27, 2021, 09:11:08 PM
post Re: Very old flyback
[Transformer (Ferrite Core)]
Alberto
February 27, 2021, 08:50:27 PM
post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
February 27, 2021, 08:01:39 PM
post Re: Rocket engine questions
[General Chat]
TMaxElectronics
February 27, 2021, 07:59:26 PM
post Re: Search for opinions concerning the use or test of a yunbao / Ebay drsstc skp
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Netzpfuscher
February 27, 2021, 07:06:07 PM
post Re: CPU Stock Cooler Failure on My i7-4970K Rendering PC
[Computers, Microcontrollers, Programmable Logic, Interfaces and Displays]
Mads Barnkob
February 27, 2021, 05:13:55 PM
post Re: Search for opinions concerning the use or test of a yunbao / Ebay drsstc skp
[Dual Resonant Solid State Tesla coils (DRSSTC)]
JCF
February 27, 2021, 03:28:25 PM
post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ChrisBlis
February 27, 2021, 01:44:13 PM
post Re: Gate drive push pull(FDS8958) heating up
[Solid State Tesla Coils (SSTC)]
prabhatkumar
February 27, 2021, 12:38:40 PM
post Re: Rocket engine questions
[General Chat]
plasma
February 27, 2021, 08:03:49 AM
post Re: Tesla Coil Show Controller Project
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Mads Barnkob
February 27, 2021, 07:07:16 AM
post Re: Very old flyback
[Transformer (Ferrite Core)]
davekni
February 27, 2021, 06:11:37 AM
post Re: Gate drive push pull(FDS8958) heating up
[Solid State Tesla Coils (SSTC)]
davekni
February 27, 2021, 06:03:00 AM
post Re: Rocket engine questions
[General Chat]
Twospoons
February 27, 2021, 05:53:55 AM
post Re: Very old flyback
[Transformer (Ferrite Core)]
MRMILSTAR
February 27, 2021, 05:53:52 AM
post Rocket engine questions
[General Chat]
plasma
February 27, 2021, 05:27:55 AM
post Re: Help for people buying the "12-48 Volt 1800/2500 Watt ZVS induction Heater"
[Electronic Circuits]
hightemp1
February 27, 2021, 04:50:48 AM
post Re: Gate drive push pull(FDS8958) heating up
[Solid State Tesla Coils (SSTC)]
prabhatkumar
February 27, 2021, 04:48:49 AM
post Re: Gate drive push pull(FDS8958) heating up
[Solid State Tesla Coils (SSTC)]
davekni
February 26, 2021, 10:55:04 PM
post Re: Very old flyback
[Transformer (Ferrite Core)]
Alberto
February 26, 2021, 10:54:13 PM
post Re: Very old flyback
[Transformer (Ferrite Core)]
davekni
February 26, 2021, 10:10:11 PM
post Gate drive push pull(FDS8958) heating up
[Solid State Tesla Coils (SSTC)]
prabhatkumar
February 26, 2021, 09:56:58 PM
post Re: Search for opinions concerning the use or test of a yunbao / Leopard drsstc skp
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Maju
February 26, 2021, 09:07:13 PM
post Re: Very old flyback
[Transformer (Ferrite Core)]
MRMILSTAR
February 26, 2021, 08:30:42 PM
post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
February 26, 2021, 08:01:54 PM
post Re: Very old flyback
[Transformer (Ferrite Core)]
davekni
February 26, 2021, 06:57:00 PM
post Search for opinions concerning the use or test of a yunbao / Ebay drsstc skp
[Dual Resonant Solid State Tesla coils (DRSSTC)]
JCF
February 26, 2021, 04:15:39 PM
post Re: GDT resources
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Jesperb123
February 26, 2021, 09:26:09 AM
post Re: DRSSTC Performs only with 1cm Arcs - Troubleshooting
[Dual Resonant Solid State Tesla coils (DRSSTC)]
ChrisBlis
February 26, 2021, 09:07:40 AM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
fh89
February 26, 2021, 07:20:13 AM
post Re: Mains Fullwave CWDRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Phoenix
February 26, 2021, 06:36:58 AM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
fh89
February 26, 2021, 12:18:08 AM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
February 25, 2021, 11:26:16 PM
post Re: Very old flyback
[Transformer (Ferrite Core)]
Alberto
February 25, 2021, 11:14:19 PM
post Re: Mains Fullwave CWDRSSTC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
AeraCura_
February 25, 2021, 09:54:08 PM
post Re: 8 ft spark french tesla coil
[Spark Gap Tesla Coils (SGTC)]
paulj
February 25, 2021, 09:25:11 PM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
fh89
February 25, 2021, 09:09:45 PM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
February 25, 2021, 08:37:26 PM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
fh89
February 25, 2021, 08:29:10 PM
post Re: Very old flyback
[Transformer (Ferrite Core)]
davekni
February 25, 2021, 07:54:08 PM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
February 25, 2021, 06:49:00 PM
post Re: Large coil experimenting with long on-times.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
fh89
February 25, 2021, 06:01:13 PM
post Re: Try to make a full Bridge with IGBT and GDT
[Dual Resonant Solid State Tesla coils (DRSSTC)]
JCF
February 25, 2021, 04:49:19 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