First drsstc issues

[Hari33]

Hello,I am an absolute beginner to drsstc. thanks to this forum I carried out the first test of my not-so-well made drsstc in my university laboratory.
The output is very bad, I only get small sparks when the secondary base wire gets close to the top load.
I gave feedback pulses from a signal generator because feedback doesn't work well.

These are the specifications of the coil
-half bridge of 60n60fd1 igbt
- 2000 secondary turns (36 AWG wire)
- 8 turns primary (14 awg) - cylindrical primary
- 80us on time of interrupter
-0.1uf tank capacitor 2.4kvdc
-javatc says the resonance frequency is 260 khz .
(I am getting those tiny sparks only when i am giving 260khz sine wave from signal generator to the feedback).

Gdt waveforms are good.
The gate waveforms are changing when the voltage input is increased.

I tried scoping the feedback ct when running the coil with feedback pulses from signal generator, those waveforms are a mess.the ratio of ct is 1:500.

primary coil draws almost 12 amps with only 50 volts AC

Could you suggest some troubleshooting tips?

Thank you.

[Mads Barnkob]

Hi Hari and welcome to HVF

When you run with a signal generator, you do not get the dynamic feedback of the changing resonant frequency, due to spark loading or other disturbances to the resonant system.

You risk blowing up your IGBTs from hard switching too high currents this way.

Did you check that phasing is correct of your GDT connections? Did you try to switch polarity of driver signal (what driver are you using?) with the jumpers or switching polarity on the primary coil?

The small sparks does resemble wrong phasing, as all spark you see is the small overlap in the dead-time.

[Hari33]

 Thank you so much for the reply.🙏

I am using Steve Wards 1.3b driver.

I tried reversing the CT secondary wires, but there is no difference.could you please tell how to check if phasing is right?.

I didnt try reversing the polarity of primary? Will update the result after doing it today.

Also could you please clear these beginner doubts?

-How does intially the feedback CT's give feedback to the driver if there is no current flowing in primary during startup in a drsstc?.

-what would be causing the change in gate waveforms when the voltage to the primary is increased?

-These tiny sparks that i get, i am getting this only at around 260khz from the signal generator, this means that 260khz is the resonant freqency of the secondary?( java tc says 260khz)

-Isnt this current draw abnormal for a small drsstc at this voltage?

Thank you

[Hari33]

Hi, I solved a lot of the issues with my first drsstc, now i am stuck with another one.
I was finally able to get the primary self oscillating by fixing some noise in the driver, but now the issue is the primary resonating at a different frequency(125khz) than what is specified by Javatc(250khz).
I tried fixing the issue by tuning it by adding more capacitors in series to mmc (mmc = 6×0.24uf mkp capacitors used in induction cookers) to increase the primary frequency to match the secondary frequency which is 260khz according to javatc.When the primary and secondary frequencies are matched, the primary feedback amplitude has reduced a lot ( primary current has decreased) also very small arcs, now the half bridge is drawing only 1.5 amp whereas it was drawing 5 amps when primary was resonating at 125Khz.

This is the problem i stuck with now.
Could this be because of the added esr in mmc?

In the pic i am attaching, blue is secondary current waveform and yellow is primary feedback.

[davekni]

In the pic i am attaching, blue is secondary current waveform and yellow is primary feedback.

200V on primary feedback indicates input diodes of UD1.3b are open-circuit.  After replacing those diodes, I'd start again back with JavaTC calculated MMC value.

[Hari33]

In the pic i am attaching, blue is secondary current waveform and yellow is primary feedback.

200V on primary feedback indicates input diodes of UD1.3b are open-circuit.  After replacing those diodes, I'd start again back with JavaTC calculated MMC value.

I replaced the diodes and today measured the primary feedback across a 3 ohm burden resistance, i am using a 1:500 ct.Now its measuring almost 0.9 volts peak across the resistor, so its 150 amps through the primary. I kept a big cylindrical topload and reduced the secondary frequency to 210 khz.I measured the primary resonance manually with a mmc of 60nf, and its around 205 khz.
At 300vdc on the bus,even though the sparks are bigger when i am bringing the earth wire near the top load, they are not breaking out on its own.

[davekni]

I replaced the diodes and today measured the primary feedback across a 3 ohm burden resistance, i am using a 1:500 ct.Now its measuring almost 0.9 volts peak across the resistor, so its 150 amps through the primary.

Do you have a separate measurement CT for the 3 ohm burden resistance?  Or are you adding 3 ohms in parallel with UD1.3b feedback CT input?  If the latter, that is preventing proper UD1.3b operation.

[Hari33]

Do you have a separate measurement CT for the 3 ohm burden resistance?  Or are you adding 3 ohms in parallel with UD1.3b feedback CT input?  If the latter, that is preventing proper UD1.3b operation.

Yes, i have a separate ct for feedback measurement.

[davekni]

Is there now a +-5V square wave on UD1.3b feedback input with replaced diodes?

[Hari33]

Is there now a +-5V square wave on UD1.3b feedback input with replaced diodes?

Yes there is 👍

[Hari33]

I got my first light today!!.
I got it when the phase difference between primary and secondary was around 180 degrees.
Is this how its supposed to be?
Yellow is primary current and blue is secondary.

[davekni]

I got it when the phase difference between primary and secondary was around 180 degrees.
Is this how its supposed to be?

Presuming CTs are all the correct way around so current is truly 180 degrees, that indicates operation at upper pole frequency.  Likely primary is tuned above secondary.  I don't see any of the typical cyclic power transfer from primary to secondary and back.  If tuning is close (similar primary and secondary frequencies), energy cycles back and forth between primary and secondary.  Tuning may be quite a ways off.

[Hari33]

Hi, it took a while for me to restart the work of the drsstc.
Finally I got the 1.3b driver printed and the small drsstc is working fine, most of its issues disappeared once i started using pcb instead of the driver that i made on breadboards.
it's able to give out 30cm arcs easily.

I started working on a bigger coil this week, everything is going good except the feedback, its working fine giving out 50cm arcs with a signal generator but it's not oscillating on it's own, the ocd is also working good.
This is the first time I am making a full bridge with igbt half bridge modules. I am using 2 old bsm100gb60dlc which I salvaged from an old inverter.

I checked the output of the gdt and its showing the initial startup pulses given by the interrupter.
Some how that is not making the igbts work.
(No output from bridge at the time of startup pulses)

Gate voltage is 18volts, interrupter-140 bps( 80us on time).
Gate resistor - 5 ohms
Ct - 1:1000 (3 ferrite cores cascaded,10 turns each)

Any tips for debugging would of great help.
Thank you

[davekni]

Second paragraph of this post may help even though it was in an SSTC thread:
https://highvoltageforum.net/index.php?topic=840.msg5667#msg5667
There are several other discussions about startup issues on the forum.  Adding a power resistor across H-bridge outputs is the key take-away from above post, as recommended by Mads in his posts and web pages.

[Hari33]

Thanks for sharing this, will try this and see how it goes.

[Hari33]

Hi, i tried keeping a 100k 1W resistor across the bridge output,but it didnt work.
I also tried keeping a 100k resitor across the 1 and 2 pins of the 74hc14 to make the 1.3 driver self starting,this is working but the coil is having a lot of breaks in between.

There is no ct output when the interrupter is giving the startup pulses.

The coil is working well when its run with a signal generator.

And can someone please explain why making the drsstc using sine waves from a signal generator is a bad idea.
The feedback from ct is also a sine wave righ?, and wont the flip flop of 1.3b driver ensure soft switching the igbts?
Isn't it just the secondary frequency that gets affected by breakout and other factors, does the primary get affected too?

[davekni]

There is no ct output when the interrupter is giving the startup pulses.

Something is mis-wired or broken in your UD1.3.  I'd scope through UD1.3 from interrupter to driver outputs (GDT primary) to see where signal is getting lost.

[Egg]

And can someone please explain why making the drsstc using sine waves from a signal generator is a bad idea.
The feedback from ct is also a sine wave righ?, and wont the flip flop of 1.3b driver ensure soft switching the igbts?
Isn't it just the secondary frequency that gets affected by breakout and other factors, does the primary get affected too?

If you run the circuit through signal generator input than you lose the primary feedback which is the main benefit of the drsstc, where you have super rapid current growth. With no feedback you will not achieve that since being even a little off on the freq causes some impedance on primary circuit.

[Hari33]

There is no ct output when the interrupter is giving the startup pulses.

Something is mis-wired or broken in your UD1.3.  I'd scope through UD1.3 from interrupter to driver outputs (GDT primary) to see where signal is getting lost.

There is gdt output during the startup pulses, but its just that the bridge is not working at those pulses.
I am assuming that the bridge is not working because there is no feedback ct output at the time of those start up pulses

[Hari33]

And can someone please explain why making the drsstc using sine waves from a signal generator is a bad idea.
The feedback from ct is also a sine wave righ?, and wont the flip flop of 1.3b driver ensure soft switching the igbts?
Isn't it just the secondary frequency that gets affected by breakout and other factors, does the primary get affected too?

If you run the circuit through signal generator input than you lose the primary feedback which is the main benefit of the drsstc, where you have super rapid current growth. With no feedback you will not achieve that since being even a little off on the freq causes some impedance on primary circuit.

Thank you🙌
But what if the the system is tuned and signal generator is used to produce sine waves having frequency of the primary?

[AstRii]

But what if the the system is tuned and signal generator is used to produce sine waves having frequency of the primary?

That's never going to happen, as the primary frequency keeps ever so slightly changing with environment and load.

[davekni]

There is gdt output during the startup pulses, but its just that the bridge is not working at those pulses.

Sorry, I misread, thinking GDT rather than CT.  No CT output is what resistor is intended to fix.  Perhaps one or more of the IGBTs his high leakage current, more than 100k can overcome.  Measure bridge output voltage with a DMM while idle.  Ideally should be 0V.

If IGBT leakage current is the issue and more than 100k can fix, another option to fix:  Swap GDT primary leads.  Then opposite IGBTs will turn on during startup pulse, now the non-leaky IGBTs.  That should make a CT response.  If polarity is wrong, reverse CT leads to fix polarity.  (This fix presumes the same IGBT continues to have high leakage current, but that may be likely.)

[Hari33]

Hi, thanks for all the guidance.
I was able to get feedback today!!. The issue solved when i used a 1:50 feedback ct instead of 1:500 ct.
I am planning to keep ocd at 300amps so the feedback current is going to be around 6 amps, which if i am right is way above what 1.3b driver can handle.
Is there any way to tackle this?, or is it a feedback ct issue at all?is there anything else i should be concerned about?.

[Mads Barnkob]

But what if the the system is tuned and signal generator is used to produce sine waves having frequency of the primary?

That's never going to happen, as the primary frequency keeps ever so slightly changing with environment and load.

This happens

[davekni]

I was able to get feedback today!!. The issue solved when i used a 1:50 feedback ct instead of 1:500 ct.
I am planning to keep ocd at 300amps so the feedback current is going to be around 6 amps, which if i am right is way above what 1.3b driver can handle.
Is there any way to tackle this?, or is it a feedback ct issue at all?is there anything else i should be concerned about?.

You might get away with 50:1 if duty cycle is low enough and R4 is changed from 10 ohms to 0.5 ohm power resistor.
I'd recommend instead figuring out why 500:1 is failing.

Are you bypassing OCD now?   If so, how?  Is your OCD CT in the primary circuit with output unconnected?  That could be the cause of your startup issues.  Unconnected CT secondaries need to be shorted.

[Hari33]

But what if the the system is tuned and signal generator is used to produce sine waves having frequency of the primary?

That's never going to happen, as the primary frequency keeps ever so slightly changing with environment and load.

This happens

/>

I wasnt aware about how running drsstc using signal generator leads to hard switching😅.
Its clear now🙌.

[Hari33]

I was able to get feedback today!!. The issue solved when i used a 1:50 feedback ct instead of 1:500 ct.
I am planning to keep ocd at 300amps so the feedback current is going to be around 6 amps, which if i am right is way above what 1.3b driver can handle.
Is there any way to tackle this?, or is it a feedback ct issue at all?is there anything else i should be concerned about?.

You might get away with 50:1 if duty cycle is low enough and R4 is changed from 10 ohms to 0.5 ohm power resistor.
I'd recommend instead figuring out why 500:1 is failing.

Are you bypassing OCD now?   If so, how?  Is your OCD CT in the primary circuit with output unconnected?  That could be the cause of your startup issues.  Unconnected CT secondaries need to be shorted.

All the unused ct's were shorted👍.
I tried using a 1:100 ct, but the start up oscillations are only having an amplitude of 1 volts max, and its not increasing even with a bus voltage of 100V.
What all could be the reasons for such a weak feedback signal?

[Hari33]

I tried scoping the igbt gates today.
The gate signals at the time of interrupter pulses are very low, in the range of millivolts.
When i gave feedback input from a signal generator without any voltage on bus,the gate signals are good.

Is this the reason for low feedback signal?


I am not sure about the gdt core material.
Dimensions are
34mm outer dia
22mm inner dia
12mm width
7 turns.

Gate resistance is 5 ohms.

Attaching the pics of waveforms with this

[Mads Barnkob]

I tried scoping the igbt gates today.
The gate signals at the time of interrupter pulses are very low, in the range of millivolts.
When i gave feedback input from a signal generator without any voltage on bus,the gate signals are good.

Is this the reason for low feedback signal?


I am not sure about the gdt core material.
Dimensions are
34mm outer dia
22mm inner dia
12mm width
7 turns.

Gate resistance is 5 ohms.

Attaching the pics of waveforms with this

Feedback signal is needed for it to run. IC2C and IC2D are AND-gates which combines feedback and interrupter signal, to the gate driver ICs. This AND-gate is what insures ZCS and prevents hard switching as when you ran it with signal generator, this feedback needs to come from the primary LC circuit.

https://kaizerpowerelectronics.dk/tesla-coils/universal-driver-1-3b/

[Hari33]

Okay,got it👍.
Shouldnt the gate startup pulses at igbt gates have the same amplitude of gate drive voltage?
In my case those pulses are very weak.
Is this normal?

Used 18v zener diodes for gate protection.

[davekni]

Shouldnt the gate startup pulses at igbt gates have the same amplitude of gate drive voltage?
In my case those pulses are very weak.
Is this normal?

Yes, start amplitude should be same as run amplitude, one Vge at +18V and other at -18V.  Low amplitude is not normal.
My guess is that your driver chip is the wrong part, not inverting.  Should be UCC27423, not UCC27424.
Either that or 74HC08 is inverting (74HC00), but I think that would cause other issues too.
Measure voltage of driver chip outputs when idle (not enabled).  They should be +9V.

[Hari33]

Shouldnt the gate startup pulses at igbt gates have the same amplitude of gate drive voltage?
In my case those pulses are very weak.
Is this normal?

Yes, start amplitude should be same as run amplitude, one Vge at +18V and other at -18V.  Low amplitude is not normal.
My guess is that your driver chip is the wrong part, not inverting.  Should be UCC27423, not UCC27424.
Either that or 74HC08 is inverting (74HC00), but I think that would cause other issues too.
Measure voltage of driver chip outputs when idle (not enabled).  They should be +9V.

Hi, i am using TC4427 instead of UCC27423.
Mosfets are irf540 and 9540.
When i connect sg to the feedback, everything works well.
Then the outputs of driver IC is inverting too (i scoped them).

It looks like the bridge is not working well for the startup pulse from interrupter.The whole thing is working well when there is a x frequency pulses at the gates(when sg is connected to feedback input).

[davekni]

Hi, i am using TC4427 instead of UCC27423.

Scope through driver as I'd suggested in reply #16.  Given your scope capture of Vge startup pulse missing, the issue is in the driver.  (Hint:  TC4427 data sheet says it is non-inverting.  I think you need to scope carefully input and output of every chip and FETs on driver path from enable to GDT primary leads.)