Author Topic: Some questions about my first drsstc.  (Read 3655 times)

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #20 on: September 03, 2022, 05:53:29 AM »
Quote
Bad news :(:
Sorry about your bad news of IGBTs frying.

Quote
otherwise this is how i installed the new gdt:
Looks well-done.

Quote
y the way, I have 60N60SMD lying around, are they as good as the HGTG30N604AD?
Presuming FGH60N60SMD, they are probably as good.  Quite a few differences in spec details, but overall FGH60N60SMD looks as good, perhaps slightly better.  May need to tweak phase-lead when changing IGBT part types.  10-ohm gate resistors might also be good, as typical gate charge is slightly lower.

Quote
I may have a hypothesis, to test I used midi files with a lot of notes, I saw the files of bytecode77 and they have a lot less notes, so I think with all these notes, on times overlapped and maybe it made hard switching.
Hard switching is a possibility, or just too-high a duty cycle.  The hard-switching issue can be fixed with the UD2.7 self-oscillation mod I've posted about if also combined with a tweak to the enable FF:  Instead of using preset input for enable, use the D input (from the high-true version of enable gate output).  That way enable always starts on a zero-current edge.  If no current, self-oscillation provides edges to clock in enable.  (This is the way my DRSSTC works, though not based on UD2.7.)

Good luck with the repairs!
David Knierim

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #21 on: September 09, 2022, 12:19:06 PM »
I received my TVS diodes and resistors, I also ordered other more powerful fans because the fans installed are not powerful, really not powerful. I still buy other HGTG30N60A4D just in case.

I installed 27v TVS diodes and 10ohm gate resistors, and the FGH60N60SMD.


Here is the output of the gdt on the igbt, Vge (30v on the bus):








And here is the output of the bridge, I used several lipo in series for a total of about 80v :





We can also see that the spike is smaller :







I set the OCD to 250a for the first test, I think maybe increase it to 300a later.

I will wait to receive the new fans to do some tests with more power, the other fans are really not powerful, so I had a hypothesis: maybe the fans are so not powerful that the IGBTs have overheated and have fried.
Mathieu Thomé

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #22 on: September 09, 2022, 03:09:31 PM »
Well...........

I tested in low power, 100Vac, two Igbt died.
It demoralizes me very much, I understand nothing.

I'll try to replace the gdt with the old one and if that doesn't change anything, maybe it's better to switch to skm100 IGBTs or something like that ?
« Last Edit: September 09, 2022, 07:46:36 PM by mthome4 »
Mathieu Thomé

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #23 on: September 10, 2022, 06:16:36 AM »
Quote
tested in low power, 100Vac, two Igbt died.
It demoralizes me very much, I understand nothing.
Ouch!  Yes, failures are very frustrating.  Especially given how many different possible causes exist.  However, you are learning a lot!  (Sometimes it helps to take a break after a failure.  Both helps emotionally and allows time to think through possible issues and strategies for testing and/or design changes.)

One simple question:  I presume you remembered to disconnect scope ground before powering from AC.

Quote
Here is the output of the gdt on the igbt, Vge (30v on the bus):
There are short Vge spikes at each H-bridge output transition after enable ends.  Is this a low-side IGBT (emitter to Vbus-)?  If high-side, could be a scoping artifact.  Presuming low-side, could be a combination of GDT inter-winding capacitance (especially with shielded cable) and/or emitter lead inductance.

I can't quite tell from previous pictures:  How long are the IGBT leads from IGBT case to ECB?  Especially with your fast IGBTs, excess emitter lead or trace inductance causes Ice current transitions to induce spikes in the local Vge within IGBT devices.

Quote
I'll try to replace the gdt with the old one and if that doesn't change anything, maybe it's better to switch to skm100 IGBTs or something like that ?
The old GDT likely has lower inter-winding capacitance, so may reduce Vge spikes.  Or, add ferrite beads to high-side GDT outputs.

One possibility is to try slower TO247 IGBTs.  Phase-lead needs to increase to compensate for more delay.  Sensitivity to Vge spikes is less because slower IGBTs don't respond to such short glitches as much.  Switching loss will be a bit higher with slow IGBTs even under ZCS conditions.

Bricks are generally slower than TO247 IGBTs, but do have a few advantages.  One key advantage is separate emitter leads for GDT connection.  This mostly eliminates the issue of emitter lead inductance.  Ice current doesn't share the same interconnect path, so no common inductance with Vge connection.  TO247 IGBTs with single emitter leads require more careful ECB layout and lead length minimization to reduce the amount of emitter parasitic inductance that is common to both Ice and Vge.  Especially critical for fast IGBTs.

Another advantage of bricks is isolated case metal.  Avoids the need for insulating thermal pads externally, increases contact area with heatsink, and increases thermal mass within IGBT package.

A related advantage is that bricks have mounting holes on both ends.  Allows more consistent thermal contact with heatsink.  TO247 packages that are mounted by their screw hole have pressure applied at only that one point.  Power dissipation (IGBT silicon die) is between hole and pins.  Pressure (thermal contact) under die can be poor.  More screw torque can actually make thermal contact worse.  Screw torque compresses the thermal pad under the mounting hole, which can pivot the lead-end of the TO247 package away from the heatsink.  Most commercial TO247 mounting uses spring clips or other mechanical methods to apply pressure to the package epoxy directly over the die.  The hole often isn't used at all.  It is quite possible that this latest failure was due to insufficient thermal contact between die and heatsink.  (My DRSSTC uses 40 TO247 IGBTs, each clamped over the die, not using screw hole.  Also has very-short IGBT leads.)

Good luck!  Keep learning.  That leads to expertise.
David Knierim

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #24 on: September 10, 2022, 10:33:42 AM »
Quote
Ouch!  Yes, failures are very frustrating.  Especially given how many different possible causes exist.  However, you are learning a lot!  (Sometimes it helps to take a break after a failure.  Both helps emotionally and allows time to think through possible issues and strategies for testing and/or design changes.)

Yes it's true that we learn a lot!
It's true that taking breaks helps, it helps not to do anything stupid!

Quote
One simple question:  I presume you remembered to disconnect scope ground before powering from AC.

yes I disconnected the oscilloscope from the bridge, I always check several times before powering up.

Quote
There are short Vge spikes at each H-bridge output transition after enable ends.  Is this a low-side IGBT (emitter to Vbus-)?  If high-side, could be a scoping artifact.  Presuming low-side, could be a combination of GDT inter-winding capacitance (especially with shielded cable) and/or emitter lead inductance.

It was on the high side but I also tested on the low side and it's the same.
For the gdt I finally made a "classic" gdt without shielding because I couldn't do anything clean (the shielding is very difficult to solder)

Quote
I can't quite tell from previous pictures:  How long are the IGBT leads from IGBT case to ECB?  Especially with your fast IGBTs, excess emitter lead or trace inductance causes Ice current transitions to induce spikes in the local Vge within IGBT devices.

Here is a picture showing the leads of the igbt :



Quote
The old GDT likely has lower inter-winding capacitance, so may reduce Vge spikes.  Or, add ferrite beads to high-side GDT outputs.

I do not understand, why these spikes make the igbt die?

Quote
One possibility is to try slower TO247 IGBTs.

The problem is that the igbt are difficult to solder, that's why I asked if bricks like the skm100 could solve the problem.(btw thank you for this information on the bricks!)

So the bricks can be more adapted to my configuration, there would be less stress compared to the igbt to247 ?

Quote
Most commercial TO247 mounting uses spring clips or other mechanical methods to apply pressure to the package epoxy directly over the die.  The hole often isn't used at all.  It is quite possible that this latest failure was due to insufficient thermal contact between die and heatsink.  (My DRSSTC uses 40 TO247 IGBTs, each clamped over the die, not using screw hole.  Also has very-short IGBT leads.)

It's true that it can only be better but it's rather difficult to do on my configuration.
I saw that your coil with 40 igbt works so well, well done!

Quote
Good luck!  Keep learning.

Thanks a lot !
Mathieu Thomé

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #25 on: September 11, 2022, 04:01:09 AM »
Quote
Here is a picture showing the leads of the igbt :
Looks great.  Nice short leads.

Quote
I do not understand, why these spikes make the igbt die?
May not be an issue even for fast IGBTs.  The only times I've experienced issues with short spikes due to GDT capacitance is with FETs, which are generally faster than even the fastest IGBTs.  In my FET cases, the spike sometimes gets large enough to turn FET on briefly, which causes another output voltage transition, which causes another spike, another transition, etc.  Result is high-frequency oscillation at each transition.  Causes excess power dissipation and damaged FETs.  Clearly you are not seeing such oscillation in scope captures.  Only possible concern is that such oscillation occurs at higher voltage where output voltage swings are larger and faster.  (Faster because IGBT capacitance is lower at higher voltage.)

Quote
The problem is that the igbt are difficult to solder, that's why I asked if bricks like the skm100 could solve the problem.
Soldering large devices such as IGBTs requires a large soldering iron.  Size and material of the tip are more important than wattage.  It is the stored heat within the tip that provides energy to solder one connection.  The heating element replaces that energy mostly between soldering operations.  Use a large tip and start with the tip a bit extra hot.  Then IGBTs will not be a problem to solder.

Quote
It's true that it can only be better but it's rather difficult to do on my configuration.
I saw that your coil with 40 igbt works so well, well done!
Here's one option that isn't too difficult.  Use a large washer (fender washer) bent to touch heatsink past IGBT case on one side and press on top of IGBT die at the other end.

Example parts for a single IGBT.  (This is actually an ancient BJT.  An old part that was easy to grab for a mechanical sample.)  The un-bent washer in the photo isn't used, just there to show the washer size used for the bent one.




Assembled to heat sink.  (Sorry for the very-fuzzy image.  Had a hard time getting my camera to focus close-up.)


Another variation is to add a plastic spacer to help define location of clamping force over IGBT and to soften contact line.  A short section of a nylon zip-tie works well.  Didn't cut the zip-tie for this picture.  Also a bit fuzzy.


My 40-IGBT build is a bit different and more complex.  Probably much more complicated than necessary.
« Last Edit: September 11, 2022, 04:28:37 AM by davekni »
David Knierim

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #26 on: September 11, 2022, 09:17:53 AM »
Thanks so much for all the advice!

Quote
Here's one option that isn't too difficult.  Use a large washer (fender washer) bent to touch heatsink past IGBT case on one side and press on top of IGBT die at the other end.

Thank you for this tip, I'll try all that in a few days, and I'll get back to you.
« Last Edit: September 11, 2022, 09:40:44 AM by mthome4 »
Mathieu Thomé

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #27 on: September 11, 2022, 07:39:19 PM »
I just realized that I forgot a question:

Quote
May not be an issue even for fast IGBTs.  The only times I've experienced issues with short spikes due to GDT capacitance is with FETs, which are generally faster than even the fastest IGBTs.  In my FET cases, the spike sometimes gets large enough to turn FET on briefly, which causes another output voltage transition, which causes another spike, another transition, etc.  Result is high-frequency oscillation at each transition.  Causes excess power dissipation and damaged FETs.  Clearly you are not seeing such oscillation in scope captures.  Only possible concern is that such oscillation occurs at higher voltage where output voltage swings are larger and faster.  (Faster because IGBT capacitance is lower at higher voltage.)

There is something that I do not understand, you say that the capacity of the gdt is not necessarily a problem even for fast igbt, with my old gdt (which probably has less capacity) I did not have fried igbt problem , but only after replacing the gdt, the igbt started to fried.
The waveform of the bridge is quite good (at least not worse than with the old gdt).
So I ask myself : maybe the problem comes from my new gdt? but I do not understand why, the difference is 10 turns instead of 18 turns and a capacity, which I suppose higher.
I'm not 100% sure but I think replacing the gdt with the old one might solve the problem.(because I had no fried igbt problem before gdt replacement.)
Maybe something escapes me but I do not understand how the new gdt burn the igbt's (if we based on my hypothesis of gdt problem)

Excuse me if my English has mistakes at times, I am not a pro in English :)
« Last Edit: September 11, 2022, 07:41:28 PM by mthome4 »
Mathieu Thomé

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #28 on: September 11, 2022, 09:19:30 PM »
Quote
There is something that I do not understand, you say that the capacity of the gdt is not necessarily a problem even for fast igbt, with my old gdt (which probably has less capacity) I did not have fried igbt problem , but only after replacing the gdt, the igbt started to fried.
The waveform of the bridge is quite good (at least not worse than with the old gdt).
So I ask myself : maybe the problem comes from my new gdt? but I do not understand why, the difference is 10 turns instead of 18 turns and a capacity, which I suppose higher.
I'm not 100% sure but I think replacing the gdt with the old one might solve the problem.(because I had no fried igbt problem before gdt replacement.)
Maybe something escapes me but I do not understand how the new gdt burn the igbt's (if we based on my hypothesis of gdt problem)
Figuring out why an electronic part failed is usually difficult.  It is rare to have a scope capture showing the exact failure event.  The strategy is to look for all possible causes of failure, choose which ones seem most likely to have caused problems, and improve those parts of the circuit.  It is still possible that the Vge glitch is related to a failure mechanism, but perhaps not.  I've seen other scope traces posted on the forum with glitches that didn't cause failures.  I've personally experienced failures, but only on FET H-bridges.
It is also possible that the latest failure was due to one IGBT having poor thermal contact with heatsink.  Each time an IGBT is replaced, there's a chance that the new part isn't quite as flat on the bottom or that the thermal pad picked up some grit or didn't compress as well.  (Most thermal pad manufactures specify that the pad needs replacing every time a part is removed.  I think most hobbyists don't bother replacing pads.  Replacement is less critical if force is applied to IGBT over die area and if the pad is kept clean.)  Quality of thermal pad makes a difference too.
The IGBT failure cause may also be something I haven't guessed yet.  Perhaps at high power some signal within UD2.7 is getting disturbed from magnetic or electrostatic coupling to external fields.  I think this is less likely in your build.  However, many builds place driver board inside a grounded metal box to minimize chances of such issues.
Another possibility:  The original 18-turn GDT likely has much higher leakage inductance than home-built ones.  Generally GDT leakage inductance is undesirable.  However, leakage inductance would slow Vge edge slew rates.  Slower Vge transitions translate to somewhat slower Vce transitions, which reduce sensitivity to H-bridge interconnect parasitic inductance.  Some DRSSTC builds include a smaller resistor in series with gate diodes, say 3 ohms for turn-off in series with diode and 10 ohms for turn-on across diode.  That is another way to slow Vge edges.  (In an ideal world, fast Vge edges are best.  Minimizes IGBT power dissipation.  However, if there is significant H-bridge interconnect inductance, fast edges may cause excess Vce spikes that fry IGBTs.)

Quote
Excuse me if my English has mistakes at times, I am not a pro in English :)
Your English is fine.  If I don't understand something, I'll ask.  If I misunderstand and answer the wrong question, please ask again.  (I'm in USA and know no other languages.  We get spoiled by so much of the world being willing to communicate in English.)
David Knierim

Offline Maju

  • High Voltage Enthusiast
  • *
  • Posts: 22
  • Karma: +3/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #29 on: September 12, 2022, 04:32:03 PM »
Hi

In my opinion your IGBT are still dying becouse of waaaaay to long pulses. TO247 really hate to be turned on for so long (based on your scopes) with so high currents.

To clarify, you should trying to achieve OCD goal ASAP - in just fiew cycles! With your frequency it should be possible in under 40-50us (lesser = better).
So if you still want experimenting with TO247, increase capacity of your resonant cap and decrise number of turns in primary (even by half or more) to achieve the same resonant freq. with lesser on time.

P.S. SKM100 are great if you're gonna done with TO247

Good luck!

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #30 on: September 12, 2022, 08:09:42 PM »
Hi

Quote
In my opinion your IGBT are still dying becouse of waaaaay to long pulses. TO247 really hate to be turned on for so long (based on your scopes) with so high currents.
Thank you for this information, but the problem is that before I replaced the gdt, the coil was working fine, with about 190Vac and 100us of on time.I will test with the old gdt to see if everything works again.

Quote
So if you still want experimenting with TO247, increase capacity of your resonant cap and decrise number of turns in primary (even by half or more) to achieve the same resonant freq. with lesser on time.
The problem is how i made the MMC it's hard to modify it to make something clean.

Quote
P.S. SKM100 are great if you're gonna done with TO247
If I can't do it with TO 247, I think I'll switch to skm100, it will be more reliable than with TO 247, from what I've seen there is also much less interconnect inductance on the H-bridge , which is much better I think.







Quote
It is also possible that the latest failure was due to one IGBT having poor thermal contact with heatsink.
it is likely that the problem comes from there, so I will do the washer technique (thank you for this technique ;D) and replace the gdt with the old one to see if there are any differences.

Quote
Perhaps at high power some signal within UD2.7 is getting disturbed from magnetic or electrostatic coupling to external fields.  I think this is less likely in your build.  However, many builds place driver board inside a grounded metal box to minimize chances of such issues.
I agree that the problem does not come from there, I put the ud2.7 in a metal box connected to the ground.


Quote
Another possibility:  The original 18-turn GDT likely has much higher leakage inductance than home-built ones.  Generally GDT leakage inductance is undesirable.  However, leakage inductance would slow Vge edge slew rates.  Slower Vge transitions translate to somewhat slower Vce transitions, which reduce sensitivity to H-bridge interconnect parasitic inductance.  Some DRSSTC builds include a smaller resistor in series with gate diodes, say 3 ohms for turn-off in series with diode and 10 ohms for turn-on across diode.  That is another way to slow Vge edges.  (In an ideal world, fast Vge edges are best.  Minimizes IGBT power dissipation.  However, if there is significant H-bridge interconnect inductance, fast edges may cause excess Vce spikes that fry IGBTs.)
Thank you for these explanations, it's always a pleasure to learn more.



Of course, I will keep you informed of the progress.

btw,I also received and installed the new fans which are much more powerful than the fans I got from a computer.


« Last Edit: September 12, 2022, 08:24:58 PM by mthome4 »
Mathieu Thomé

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #31 on: September 13, 2022, 03:43:58 AM »
Quote
    In my opinion your IGBT are still dying becouse of waaaaay to long pulses. TO247 really hate to be turned on for so long (based on your scopes) with so high currents.

Thank you for this information, but the problem is that before I replaced the gdt, the coil was working fine, with about 190Vac and 100us of on time.I will test with the old gdt to see if everything works again.
The definition of too-long for TO247 depends on the power dissipation and thermal resistance.  Many (I think most) QCW coils use TO247 devices and run ~10ms on-time.  My TO247 DRSSTC typically runs 100-200us when playing music, and up to 1ms when tuned for longest arcs.

Quote
Of course, I will keep you informed of the progress.
Yes, please do.  Definitely interested.

Quote
btw,I also received and installed the new fans which are much more powerful than the fans I got from a computer.
Great.  Good thermal conductivity to heat sink and from heat sink to air is likely to make IGBTs more robust.  Of course, even perfect thermal management won't protect IGBTs if spikes exceed Vce rating for long or high-frequency oscillations occur at transitions.

Good luck with this round of repairs!
David Knierim

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #32 on: September 16, 2022, 11:05:46 PM »
After a little break, here I am again in the build of my drsstc.

I followed your tips of washers on the IGBTs but I did not take washers but pieces of metal.
Here is the result :






I also have a question, I have 400V TVS diode on each IGBT (connect to CE), I saw that it can be annoying, I understood
 that if a spike exceeds the voltage of the TVS, it goes shorted which also shorts an IGBT, the IGBT can short itself, and then short the other IGBT next to it.
My question is, do I remove the TVS diodes?
I could see several things on it but good and bad but often it is recommended to remove them, if I'm not mistaken.

Here is a pic of the TVS:



I am also seriously starting to look at skm100gb125, I think it could be more suitable, during my tests I could see that I could not go up to more than 190Vac, I would like to be able to supply it with 230v directly.
The problem is that I think that the 600v limit of IGBT TO-247 will be quickly reached whereas with the skm100 I would have much more margin.
I prefer to put 325 vdc in a skm 100 with more margin rather than in a to-247 which I think, with my configuration, is not powerful enough.
I have heard only good things about IGBT bricks and I would also like experiment with bricks :)
« Last Edit: September 16, 2022, 11:50:50 PM by mthome4 »
Mathieu Thomé

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #33 on: September 17, 2022, 01:50:27 AM »
Quote
I followed your tips of washers on the IGBTs but I did not take washers but pieces of metal.
Here is the result :
Very nicely implemented!  I expect your mounting to work quite well.  Thermal resistance will be lower than before, perhaps much lower.

Your metal tabs would be a great example for some sort of TO247 tutorial page.  I'm not sure where to put it so it would get found by new coil builders looking for guides.  Perhaps somewhere in the "beginners" page.  Would love for your example to get more visibility than buried in the middle of this one thread.

Quote
I also have a question, I have 400V TVS diode on each IGBT (connect to CE), I saw that it can be annoying, I understood
 that if a spike exceeds the voltage of the TVS, it goes shorted which also shorts an IGBT, the IGBT can short itself, and then short the other IGBT next to it.
My question is, do I remove the TVS diodes?
I could see several things on it but good and bad but often it is recommended to remove them, if I'm not mistaken.
Difficult to know for sure in your situation.  HGTG30N60A4D has no specification for avalanche energy.  There is no way to know if IGBT avalanche energy capability is higher than TVS diode or not.  Hopefully you'll get other opinions.

My suggestion is to lift one end of each TVS diode and add a capacitor in series, with a resistor in parallel with the capacitor.  I used this configuration for my resonant isolated PFC.  Allows TVS diodes to clamp short transients without excess total power dissipation.  For values, something around 0.1uF paralleled with 1k.  (Avoid carbon film for 1k resistor.  In case of over-temperature, carbon film resistors tend to fail towards low resistance.)

Quote
I am also seriously starting to look at skm100gb125, I think it could be more suitable, during my tests I could see that I could not go up to more than 190Vac, I would like to be able to supply it with 230v directly.
The problem is that I think that the 600v limit of IGBT TO-247 will be quickly reached whereas with the skm100 I would have much more margin.
I prefer to put 325 vdc in a skm 100 with more margin rather than in a to-247 which I think, with my configuration, is not powerful enough.
I have heard only good things about IGBT bricks and I would also like experiment with bricks :)
If your H-bridge ECB layout has low parasitic inductance, then 600V IGBTs will be fine at 325Vbus.  I use 650V IGBTs for 450Vbus.  Here's a recent example of a great low-inductance TO247 H-bridge ECB layout:
https://highvoltageforum.net/index.php?topic=2183.msg16087#msg16087
And my tutorial on such layout, starting with a half-bridge version:
https://highvoltageforum.net/index.php?topic=1324.msg9795#msg9795
Parasitic inductance does not necessarily need to be that low.  However, HGTG30N60A4D are quite fast, making low inductance more important.

Yes, bricks are generally more robust, especially going to 1200V bricks.  Even this "fast" SKM100 brick is about half the speed and twice the gate charge of your HGTG30N60A4D IGBTs.  Shouldn't be an issue with your ~200kHz coil.  More of a concern with ~400kHz QCW coils, especially if phase-shift is used for ramping.
« Last Edit: September 17, 2022, 01:56:00 AM by davekni »
David Knierim

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #34 on: September 17, 2022, 11:56:26 AM »
Quote
Very nicely implemented!  I expect your mounting to work quite well.  Thermal resistance will be lower than before, perhaps much lower.

Your metal tabs would be a great example for some sort of TO247 tutorial page.  I'm not sure where to put it so it would get found by new coil builders looking for guides.  Perhaps somewhere in the "beginners" page.  Would love for your example to get more visibility than buried in the middle of this one thread.

Thanks ! If you want I can do a little tutorial on that, I also think the beginner page would be suitable.


Quote
Difficult to know for sure in your situation.  HGTG30N60A4D has no specification for avalanche energy.  There is no way to know if IGBT avalanche energy capability is higher than TVS diode or not.  Hopefully you'll get other opinions.

My suggestion is to lift one end of each TVS diode and add a capacitor in series, with a resistor in parallel with the capacitor.  I used this configuration for my resonant isolated PFC.  Allows TVS diodes to clamp short transients without excess total power dissipation.  For values, something around 0.1uF paralleled with 1k.  (Avoid carbon film for 1k resistor.  In case of over-temperature, carbon film resistors tend to fail towards low resistance.)

I found four capacitors in my stock but I'm not sure if they will fit, they are 0.1uf but they are different and have voltages of 250Vac and 275Vac.




Quote
If your H-bridge ECB layout has low parasitic inductance, then 600V IGBTs will be fine at 325Vbus.  I use 650V IGBTs for 450Vbus.  Here's a recent example of a great low-inductance TO247 H-bridge ECB layout:
https://highvoltageforum.net/index.php?topic=2183.msg16087#msg16087
And my tutorial on such layout, starting with a half-bridge version:
https://highvoltageforum.net/index.php?topic=1324.msg9795#msg9795
Parasitic inductance does not necessarily need to be that low.  However, HGTG30N60A4D are quite fast, making low inductance more important.

Yes, bricks are generally more robust, especially going to 1200V bricks.  Even this "fast" SKM100 brick is about half the speed and twice the gate charge of your HGTG30N60A4D IGBTs.  Shouldn't be an issue with your ~200kHz coil.  More of a concern with ~400kHz QCW coils, especially if phase-shift is used for ramping.

I am impressed by your gdt, he has so few turns!

I saw somewhere that Mads said that he had stopped with the TO-247 because he found them too fragile, I'm not far from the limit of the TO-247, what I like with the skm100 is that there are many advantages, as you said with the two dedicated wires for the gate drive, the better heat dissipation, the metal plate is isolated and many other advantages.
I also like that the primary current will have more margin, I saw the skm100 works well with 550a OCD,
if I put my OCD at 400a, I would have more margin.
I like the fact of not pushing the IGBTs to the maximum but keeping margin.
Mathieu Thomé

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #35 on: September 17, 2022, 06:37:23 PM »
Quote
Thanks ! If you want I can do a little tutorial on that, I also think the beginner page would be suitable.
I think that would be great.  Even just a copy of those images moved to a beginner page would be useful.  Perhaps a new thread about TO247 mounting or some such title.  Whenever you get to that (no hurry), I'll add a link to your thread from my low-inductance bridge tutorial.

Quote
I found four capacitors in my stock but I'm not sure if they will fit, they are 0.1uf but they are different and have voltages of 250Vac and 275Vac.
Electrically fine.  Ceramic capacitors are likely smaller. 100V or so would probably suffice, though 200V would be prefered.  Down to 0.047uF would likely be fine, certainly no worse than leaving off TVS diodes.

Quote
I am impressed by your gdt, he has so few turns!
Most GDTs have more turns than necessary, which increases parasitic inductance in gate drive.  However, too few turns causes more serious issues than too many turns.  Thus most GDTs have extra turns just for margin.  With proper calculations and/or measurements, there is no need for extra turns.  Some day I want to add a section about such measurements and calculations to my GDT tutorial page.  Still need to figure out how to explain it simply and clearly.
David Knierim

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #36 on: September 19, 2022, 09:05:35 PM »
Quote
I think that would be great.  Even just a copy of those images moved to a beginner page would be useful.  Perhaps a new thread about TO247 mounting or some such title.  Whenever you get to that (no hurry), I'll add a link to your thread from my low-inductance bridge tutorial.
It's done :)

Quote
Electrically fine.  Ceramic capacitors are likely smaller. 100V or so would probably suffice, though 200V would be prefered.  Down to 0.047uF would likely be fine, certainly no worse than leaving off TVS diodes.

I found 0.047uf capacitors, I'll let you know soon

I just thought something, and if you can confirm what I think would be great.
When the drsstc was working, I noticed that with low on times it produces small streamers (normal), when I increase the on time, almost nothing happens until the streamers suddenly become large, I therefore think that it is due to the detune, when the streamers are too small, the frequency was too much detune and when the on time is large enough, the streamers grow to "better" be in tune with the primary, which means that the streamers suddenly becomes large.
Sorry if that's not fully understandable, but is that what i think?
« Last Edit: September 19, 2022, 09:10:51 PM by mthome4 »
Mathieu Thomé

Offline davekni

  • High Voltage Expert
  • ******
  • Posts: 1895
  • Karma: +95/-1
  • Physicist, engineer (electronic), and hobbiest
    • View Profile
Re: Some questions about my first drsstc.
« Reply #37 on: September 20, 2022, 05:20:07 AM »
Quote
I just thought something, and if you can confirm what I think would be great.
When the drsstc was working, I noticed that with low on times it produces small streamers (normal), when I increase the on time, almost nothing happens until the streamers suddenly become large, I therefore think that it is due to the detune, when the streamers are too small, the frequency was too much detune and when the on time is large enough, the streamers grow to "better" be in tune with the primary, which means that the streamers suddenly becomes large.
Sorry if that's not fully understandable, but is that what i think?
Perfectly understandable, and normal.  I believe your explanation is exactly correct.

That level of detuning is ideal for making longest-possible arcs.  A little less detuning is better for music (MIDI interrupter).  At least that has been my experience.  Provides more linear response of arc size to pulse width, so better sound reproduction.
David Knierim

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #38 on: September 23, 2022, 10:55:21 PM »
I soldered the 0.047uf capacitors and their 1k resistor :



I'll adjust and test it probably tomorrow.

Quote
That level of detuning is ideal for making longest-possible arcs.  A little less detuning is better for music (MIDI interrupter).  At least that has been my experience.  Provides more linear response of arc size to pulse width, so better sound reproduction.

Thank you for your test feedback, I will test to see what is the best.
Mathieu Thomé

Offline mthome4

  • High Voltage Enthusiast
  • *
  • Posts: 36
  • Karma: +2/-0
    • View Profile
Re: Some questions about my first drsstc.
« Reply #39 on: September 25, 2022, 09:43:24 PM »
I installed the bridge and I can feel that with the new cooling and the IGBTs mounted better, cooling is much better.
I tested the bridge with lipo in series for a total of about 110Vdc on the bus, here is what it gives :


We agree that the max voltage Vce of a single IGBT is not the "peak to peak" value on the oscilloscope but the "max" value which is around 236v, in theory I can have 1200v peak to peak to have a max of 600v on each IGBT ?
« Last Edit: September 25, 2022, 09:47:44 PM by mthome4 »
Mathieu Thomé

High Voltage Forum

Re: Some questions about my first drsstc.
« Reply #39 on: September 25, 2022, 09:43:24 PM »

 


* Recent Topics and Posts

post Re: IF D95T vs IF D95OC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
RoamingD
Today at 06:56:51 AM
post Re: IF D95T vs IF D95OC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
alexhanyuan
November 25, 2022, 11:28:15 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 25, 2022, 12:05:50 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 25, 2022, 07:53:52 AM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 24, 2022, 08:08:48 AM
post Re: EEFL backlight circuits
[Light, Lasers and Optics]
klugesmith
November 24, 2022, 07:26:49 AM
post Re: Air cored inductors proximity to conductive objects
[General Chat]
davekni
November 24, 2022, 03:02:11 AM
post Re: Solid State “rheostat” that will withstand high electrostatic fields
[General Chat]
AkashaStar
November 24, 2022, 02:25:42 AM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 24, 2022, 01:45:33 AM
post Re: Air cored inductors proximity to conductive objects
[General Chat]
John123
November 23, 2022, 11:27:18 PM
post Stepper motor drive without switching
[Electronic Circuits]
klugesmith
November 23, 2022, 10:23:07 PM
post Re: Air cored inductors proximity to conductive objects
[General Chat]
Twospoons
November 23, 2022, 09:29:01 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Alberto
November 23, 2022, 08:45:37 AM
post Re: Air cored inductors proximity to conductive objects
[General Chat]
davekni
November 23, 2022, 04:44:20 AM
post Air cored inductors proximity to conductive objects
[General Chat]
John123
November 23, 2022, 01:10:17 AM
post Re: Post-Soviet Tesla Coil Teardown and Test
[Solid State Tesla Coils (SSTC)]
Mads Barnkob
November 22, 2022, 08:40:19 PM
post Re: Ericsson RBS 6601 RRUS 12 analysis(kind of)
[Electronic Circuits]
Da_Stier
November 22, 2022, 08:22:09 PM
post Re: Ericsson RBS 6601 RRUS 12 analysis(kind of)
[Electronic Circuits]
Kirby4life
November 22, 2022, 08:45:06 AM
post Ericsson RBS 6601 RRUS 12 analysis(kind of)
[Electronic Circuits]
RickNitro
November 22, 2022, 08:36:11 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
davekni
November 22, 2022, 06:42:31 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Twospoons
November 22, 2022, 06:40:32 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
klugesmith
November 22, 2022, 04:53:40 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Twospoons
November 22, 2022, 02:17:08 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
klugesmith
November 21, 2022, 11:03:04 PM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Alberto
November 21, 2022, 09:55:02 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 21, 2022, 03:09:26 PM
post Re: Post-Soviet Tesla Coil Teardown and Test
[Solid State Tesla Coils (SSTC)]
LoOdaK
November 21, 2022, 09:38:22 AM
post Re: Trichel pulses
[Electronic Circuits]
klugesmith
November 20, 2022, 10:05:15 PM
post Trichel pulses
[Electronic Circuits]
haversin
November 20, 2022, 09:48:37 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
klugesmith
November 20, 2022, 07:17:36 PM
post Re: Some questions about my first drsstc.
[Dual Resonant Solid State Tesla coils (DRSSTC)]
mthome4
November 20, 2022, 02:28:22 PM
post Post-Soviet Tesla Coil Teardown and Test
[Solid State Tesla Coils (SSTC)]
Mads Barnkob
November 20, 2022, 02:03:49 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 20, 2022, 10:00:13 AM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
davekni
November 20, 2022, 05:13:12 AM
post Re: Plasma Toroid
[Solid State Tesla Coils (SSTC)]
davekni
November 20, 2022, 05:08:31 AM
post Re: Plasma Toroid
[Solid State Tesla Coils (SSTC)]
alan sailer
November 20, 2022, 01:27:00 AM
post Re: Basic Gate driving doubts
[Solid State Tesla Coils (SSTC)]
davekni
November 19, 2022, 10:12:14 PM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
John123
November 19, 2022, 09:57:35 PM
post Re: Basic Gate driving doubts
[Solid State Tesla Coils (SSTC)]
prabhatkumar
November 19, 2022, 09:50:23 PM
post Re: Basic Gate driving doubts
[Solid State Tesla Coils (SSTC)]
davekni
November 19, 2022, 08:26:45 PM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
davekni
November 19, 2022, 06:53:17 PM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
klugesmith
November 19, 2022, 06:05:47 PM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Alberto
November 19, 2022, 05:34:46 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 19, 2022, 02:30:04 PM
post Re: Welcome new members, come say hello and tell a little about yourself :)
[General Chat]
AkashaStar
November 19, 2022, 01:56:10 PM
post Re: Buck converter question
[Dual Resonant Solid State Tesla coils (DRSSTC)]
dru
November 19, 2022, 08:03:56 AM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 19, 2022, 06:38:00 AM
post Basic Gate driving doubts
[Solid State Tesla Coils (SSTC)]
prabhatkumar
November 19, 2022, 05:37:09 AM
post Re: ZVS driver won't oscillate
[Electronic Circuits]
davekni
November 19, 2022, 03:59:56 AM
post Re: Non-laser pointer
[Light, Lasers and Optics]
klugesmith
November 19, 2022, 03:56:32 AM
post Re: Buck converter question
[Dual Resonant Solid State Tesla coils (DRSSTC)]
davekni
November 19, 2022, 03:36:20 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
davekni
November 19, 2022, 03:26:48 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Alberto
November 19, 2022, 12:40:33 AM
post Re: ZVS driver won't oscillate
[Electronic Circuits]
AstRii
November 18, 2022, 09:32:33 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
klugesmith
November 18, 2022, 05:48:00 PM
post Re: Audio Modulated CRT Flyback Transformer Loud Sparks
[Transformer (Ferrite Core)]
John123
November 18, 2022, 03:49:12 PM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
John123
November 18, 2022, 02:56:49 PM
post IF D95T vs IF D95OC
[Dual Resonant Solid State Tesla coils (DRSSTC)]
RoamingD
November 18, 2022, 01:44:38 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 18, 2022, 11:38:44 AM
post Non-laser pointer
[Light, Lasers and Optics]
klugesmith
November 18, 2022, 10:05:29 AM
post Re: Induction cooker without electronics
[Electronic Circuits]
klugesmith
November 18, 2022, 06:54:55 AM
post Re: Induction cooker without electronics
[Electronic Circuits]
Twospoons
November 18, 2022, 06:29:23 AM
post Buck converter question
[Dual Resonant Solid State Tesla coils (DRSSTC)]
dru
November 18, 2022, 06:10:52 AM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
davekni
November 18, 2022, 05:53:15 AM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 18, 2022, 05:21:52 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
davekni
November 18, 2022, 04:00:04 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
alan sailer
November 17, 2022, 10:24:35 PM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
John123
November 17, 2022, 08:30:00 PM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
klugesmith
November 17, 2022, 07:28:16 PM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
John123
November 17, 2022, 03:38:00 PM
post Re: TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
davekni
November 17, 2022, 04:03:51 AM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
davekni
November 17, 2022, 03:57:40 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
davekni
November 17, 2022, 03:26:32 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Twospoons
November 16, 2022, 11:23:37 PM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
alan sailer
November 16, 2022, 08:00:54 PM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
John123
November 16, 2022, 05:09:00 PM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
MRMILSTAR
November 16, 2022, 04:57:27 PM
post TV flyback number of primary turns rule of thumb
[Transformer (Ferrite Core)]
John123
November 16, 2022, 03:27:06 PM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
John123
November 16, 2022, 03:16:13 PM
post Bluetooth to aux for plasma speaker circuits
[Electronic Circuits]
John123
November 16, 2022, 03:00:54 PM
post Re: Induction cooker without electronics
[Electronic Circuits]
Solhi
November 16, 2022, 11:43:49 AM
post Re: Electrical resistance cooking
[Laboratories, Equipment and Tools]
klugesmith
November 16, 2022, 09:00:46 AM
post Re: Electrical resistance cooking
[Laboratories, Equipment and Tools]
klugesmith
November 16, 2022, 08:53:42 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
davekni
November 16, 2022, 05:45:24 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
alan sailer
November 15, 2022, 08:17:46 PM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Alberto
November 15, 2022, 07:25:04 PM
post Re: Yet another UD2.x pulse skipper
[Dual Resonant Solid State Tesla coils (DRSSTC)]
hammertone
November 14, 2022, 12:26:19 PM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
klugesmith
November 14, 2022, 04:57:41 AM
post Re: Vacuum pump
[Laboratories, Equipment and Tools]
Twospoons
November 14, 2022, 12:32:54 AM
post Vacuum pump
[Laboratories, Equipment and Tools]
Alberto
November 13, 2022, 11:06:55 PM
post Re: QO-100 / EShail 2 Uplink amplifier development adventures
[Radio Frequency]
Da_Stier
November 13, 2022, 06:32:11 PM
post Re: Yet another UD2.x pulse skipper
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Hydron
November 13, 2022, 04:41:20 PM
post Re: QCW questions
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Rafft
November 13, 2022, 02:38:41 PM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
MRMILSTAR
November 13, 2022, 05:23:09 AM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
thedark
November 12, 2022, 10:39:21 PM
post Re: Yet another UD2.x pulse skipper
[Dual Resonant Solid State Tesla coils (DRSSTC)]
hammertone
November 12, 2022, 12:45:16 PM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
klugesmith
November 12, 2022, 05:44:35 AM
post Re: what type of diode is better for make high voltage diode?
[Voltage Multipliers]
thedark
November 12, 2022, 05:30:10 AM
post Re: Yet another UD2.x pulse skipper
[Dual Resonant Solid State Tesla coils (DRSSTC)]
Hydron
November 12, 2022, 01:11:31 AM
post Re: Induction cooker without electronics
[Electronic Circuits]
Twospoons
November 12, 2022, 12:42:54 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