Help needed with SSTC. First TC build

[Maverikie]

Dear HV member,

I'm new in the TC coil business. I'm trying to build a SSTC with inspiration from SSTC 5 from Steve Ward and SSTC 2 from Loneoceans. Ofcourse I have deviated from the suggested schematics. 

So i've build the interrupter based on a ATTiny and the driver is based on two UCC37322P chips. One inverted by a SN74HC14N. It seems that the interrupter and driver are working fine. I think I've problems with the half bridge consisting of two IRGP50B60PD1.

So when I try to turn everything on, it's drawing a lot of current via the Variac. See below the schematic.



I used my scope to measure the following:

First pic.	1: Gate of Q1 2: Gate of Q2 3: Input signal generator sine wave 150 khz
Second pic.	1: Output UCC37322P without GDT 2: Output UCC37322P without GDT inverted 3: Input signal generator sine wave 150 khz.

Below pictures of the general setup:

[davekni]

First issue appears to be wrong core type for GDT (and likely CT too).  Cores must be high-permeability ferrite, not powdered iron.  Mads has a good GDT tutorial, with links to it in several of his posts.  (Other GDT and CT construction info is around too - do some searching.)  Besides core type, it is best to have a 1:1:1 turns ratio wound with twisted-pair wire for low parasitic inductance (with gate-driver chip input power at 15V for IGBT drive).  1:2:2 can be constructed with reasonably-low parasitic inductance, but that is a bit more tricky and increases gate-driver load.

Second likely issue:  I don't see any insulating pad under IGBTs.  They need either insulation or separate insulated heat sinks.  The bottom surface is electrically connected to collector.

Plug-board construction is often problematic, especially for gate-driver chips with their high switching current peaks.  However, your construction is the cleanest I've seen for plug-boards, so you may be fine with that.

If you haven't tried JavaTC, that is a good idea to check your coil parameters for reasonability.

Great documentation with detailed pictures and scope traces!  That makes assistance so much easier than for many initial requests.

[Mads Barnkob]

Hi Maverikie and welcome to HVF

Davekni is fast and catches it all before any other can even finish sleep to reply

Your IGBTs are not insulated, on page 10 you can see that the cooling connection on the back of it is assigned to Drain: https://www.infineon.com/dgdl/Infineon-IRGP50B60PD1-DataSheet-v01_00-EN.pdf?fileId=5546d462533600a401535656c5e9248a you you are effectively shorting out your two drain channels.

The GDT troubleshooting guide is written by Richie Burnett: http://www.richieburnett.co.uk/temp/gdt/gdt2.html

[Maverikie]

Thanks a lot! Stupid mistake .

I found out that I fried my IGTB's, but I had foreseen some rookie moves, so I already bought two extra ones.

When I replaced the originals I tested and got my first spark!!

I checked the waveform of my GDT and it seems good. See scope pic.

But right now I have problems with my heating boiler. Seems it doesn't like the RF noise on the mains ground. In my situation it's quite hard to have a separate RF grounding rod installed. I'm reading something about a line filter can be used to suppress the voltage spikes to mains ground. Anyone experience about this?

[davekni]

Davekni is fast and catches it all before any other can even finish sleep to reply

Yes, I live in Oregon (west coast of USA), so in one of the later time-zones of forum users.  My evening-time covers almost anything posted throughout the day.  You have many more tutorials and examples of coil builds - more helpful in many cases than my myopic focus on specific issues and questions.

I checked the waveform of my GDT and it seems good. See scope pic.

Yes, I agree.  Fried IGBTs is a reasonable explanation for your previous gate waveforms rather than inappropriate GDT core choice.

Congratulations on first sparks!

[Maverikie]

It's working finally! Full power.

The interrupter circuit based on Attiny performed not so well. At higher power the input pins of the drivers went self-oscillating. And the Attiny turned itself of. I tried to shield it but didn't succeed. Any suggestions?

I tried the NE555 circuit from Steve Ward and it worked wonderful!

[Maverikie]

My SSTC based on Steve Wards design is working good, but only with Antenna feedback.

I'm not succeeding in making it work with a CT feedback. When I scope the 'output' of the antenna I get a perfect Sine wave like expected, but when I scope the output of the CT I get a strange looking waveform. See photo's.


I used these components suggested in this Threat https://highvoltageforum.net/index.php?topic=840.msg5667#msg5667


Yellow antenna feedback
Blue CT feedback

[AstRii]

Looks like something is resonating, try to remove the blocking capacitor and scope it afterwards, see what you will get. Also maybe the CT material cannot keep up with the high frequencies.
In the worst case, the coil can run like this or can't it? The schmitt-trigger should clean it up to be a nice square wave even if the input sine wave is not perfect.

[davekni]

Looks like something is resonating, try to remove the blocking capacitor and scope it afterwards, see what you will get.

Don't bother removing C2 (blocking capacitor).  That would create a DC offset.  C2 is definitely not the resonate capacitance.  C2 has about 3 ohms reactance at the third-harmonic frequency of the undesired resonance, totally insignificant compared to the 1k series resistor.

Parasitic capacitance of CT leads + D1 + D2 + 74HC14 may be resonating with CT secondary inductance.  You could damp that resonance with a parallel resistor across CT secondary, perhaps around 10k.  The parallel resistor will add a bit of phase lag.  That bit of lag may actually help depending on delays through the driver from 74HC14 to bridge output.  If those delays are already 1/4 cycle, then the lag will hurt rather than help.  In that case, a better option may be to remove some turns from the CT to reduce its inductance and raise the resonant frequency to avoid 3rd harmonic.  (Adding 10k parallel resistance while running with antenna feedback would be a good test to see if this is the source of your 3rd harmonic resonance.)

One final (but more tricky to get right) option is to add capacitance across the CT until its resonance matches the SSTCs fundamental frequency.  Then small changes in capacitance can be used to adjust phase.  That is close to how my SSTC operates.

[Maverikie]

Just to close this threat with what I've done to overcome the problem above.

I think my problem was with the feedback transformer that I wounded myself. I've made a newer, lot smaller, transformer from a ferite bead that I scraped from some old electronics. I have wound it with >100 windings with 0.3 mm enameled wire. I then used the following scheme to connect it to the 74HC14 hex inverter:

Thanks to CircuitSalad for this schematic. The two first caps with variable resister enables you to 'fine' tune the drive signal. With the help of a oscilloscope you can aim for 90 degree Phase shift:

[davekni]

Thank you for the success post.  Always nice to see that things work in the end.

BTW, I like self-oscillating circuits like this version.  R4 makes it self-oscillating.  That helps with startup even when feedback signal is weak.  Of course, if the self-oscillating frequency is way off from resonance, it may not help.  If R4 is too low (frequency too high), it can fail to lock and operate at the high self-oscillation frequency, which may overheat driver or half-bridge.

[Maverikie]

Thanks, I sometimes hate it when threats die because there was no 'end post'. Other people can learn from these things...

I didn't understand the function of R4 before. But when I think about it, I can see that R4 is charging C12 upto trigger voltage to invert the output of the hex inverter. Is that right?

[davekni]

I didn't understand the function of R4 before. But when I think about it, I can see that R4 is charging C12 upto trigger voltage to invert the output of the hex inverter. Is that right?

Yes, that is correct.  Then after the inverter switches low, it discharges C12 through R4 until the output switches high again.  That continues indefinitely until the feedback signal is stronger than the current from R4.