No change in Phase Lead

[Admiral Aaron Ravensdale]

Hi all,

still no progress with my bigger coil https://highvoltageforum.net/index.php?topic=2821.0
I can't get any change in phase lead.
I already try a second UD2.7 driver board both with C33 in 1000pF
and ALL Slot-7 Inductors 7M3-123 (9 - 15uH; works well with TO247 IGBTs), 7M3-153 (11 to 19uH), going up through 7M3-223 (17-28uH), 333 (25-41uH), 393 (29-49uH; works well with CM200/300s), 563 (42 - 70uH).
When I use the 393 and screw them in I notably here that the coil runs really bad but also there is no change on the scope!
For measurement I use a Micsig DP-20003 and a LEM 1005 and a Rigol DS1102
l also try to put the HV Probe on both CM200 outputs and also on on out and one on negative rail.
What happens is that I can  reduce the spikes on the IGBT (blue) but no change on the bad timing.
Today I want to order a new GDT Module for another coil and see that tey use different resistors on the Module for the -12 and the -24 version of the IGBT and now I wonder if this is the issue!
Currently I use this Module for my CM200DY-24NF with 4.7 Ohm resisators


But from the picture they use for the -24 Version 7.5 Ohm resistors



and now a 4.7 for -12 Version




Could this be the reason?
Or I messed up something with my scope and do a wrong mesurement?
If anyone here speak German or live near Wuppertal and can help me I would prefer because of 3 Month of useless testing and frustration

[Mads Barnkob]

Gate resistors can vary a lot from a few Ohms to up in twenties, all depends on how slow you want to turn on, or what oscillations/overshoot your are trying to fix. I doubt the problem lies with any of those shown GDT boards.

If you can see spikes disappear when adjusting the phase lead, you are actually hitting ZCS when switching spikes are the least. Or at least you are delaying ZCS to avoid excessive turn-off losses (from the spikes).

Maybe you need to check your measurement setup, Tektronix did a great SMPS guide some years back: https://highvoltageforum.net/index.php?topic=111.0

In your build thread, you mention that your coils specs are:
Specs:
Secondary:
110mm x 660mm
0.2mm AWG32 Wire
146.3 kHz

Toroid:
500mm x 60mm

Running at 93.3kHz with Toroid

Primary:
6mm tubing
tap at 79kHz

About 10% detuning on the primary would be expected for a large coil, I think you have tapped it too low, try to start near 5% detuning and move towards 10% as you can get longer and longer sparks.

Your scope shots says 48 kHz, that is pretty far from your coil specs?

[Admiral Aaron Ravensdale]

Thanks Mads,
yes you are right this is my first time I use a Scope and it is possible I messed something up!
For the scope with 48kHz, yes it shows less when it is zoomed out.
I just take a minute and make a new scope picture an zoom in then i get 97kHz

You see the spikes are nearly gone but the Phase lead is a nightmare!
When I put the probe in the air and measure the secondary wave it is nearly perfect, is this a hint for you?

[Admiral Aaron Ravensdale]

Ohh missed the note for the tuning.
I used a spare OneTesla board without original HV cap and link the HV caps from the big coil to the board.
I run it with max power and pulswide and tune in this way the primary to the sweet spot.
With the OneTesla board I got nearly 50cm Sparks.

[davekni]

yes you are right this is my first time I use a Scope and it is possible I messed something up!

Looks like scope readout labels are swapped.  Cyan (light blue) trace lists 20A/div but looks like H-bridge output.  Yellow trace lists 10V/div.  Looks like perhaps output of current transformer missing a burden resistor.  Current transformers (CTs) output current and thus require a load resistor to convert that current to voltage.  BTW, for future posts, it would help to label scope traces in text just above or below image.

Good luck with debug.

[Admiral Aaron Ravensdale]

Ok maybe I got it...
I bought this terrible expensive LEM1005 but it seems it dosen't work right.
After some research I found a hint that I can get the feedback also over the 51Ohm resistor on the UD2.7.
Now I was able to get a good scope picture and was also able to tune it.
Can someone verify that I do it right?
Her how I measure it on the UD2.7


The scope picture before tuning ( before I was only able to reduce the spikes on the IGBT output)


And after tuning

[davekni]

I bought this terrible expensive LEM1005 but it seems it dosen't work right.

Your LEM1005 is probably good.  It outputs current, so needs a burden resistor.  LEM1005 specification lists minimum and maximum allowed burden resistor values based on supply voltage used and operating temperature range.

Can someone verify that I do it right?

Your first scope capture looks better, with minimal spikes.  H-bridge output should switch slightly before zero current.
Best to do another final adjustment after H-bridge is running at full voltage.  IGBT timing changes with bus voltage.

[flyingperson23]

Can hall effect sensors measure fast enough for drsstc feedback? It looks like this one is managing but I tried a hass 300-s and it wasn't anywhere close to fast enough.

[davekni]

Can hall effect sensors measure fast enough for drsstc feedback?

Good point.  LEM1005 would be marginal for most TCs.  There are combined Hall-effect/CT sensors, common for scope current probes.  Combining is actually quite simple as in the original Tektronix patent.

[Admiral Aaron Ravensdale]

Can hall effect sensors measure fast enough for drsstc feedback? It looks like this one is managing but I tried a hass 300-s and it wasn't anywhere close to fast enough.

Data Sheet say up to 150kHz

https://www.lem.com/sites/default/files/products_datasheets/lf_1005-s_sp33.pdf

[Admiral Aaron Ravensdale]

I bought this terrible expensive LEM1005 but it seems it dosen't work right.

Your LEM1005 is probably good.  It outputs current, so needs a burden resistor.  LEM1005 specification lists minimum and maximum allowed burden resistor values based on supply voltage used and operating temperature range.

I use it with a 50 Ohm burden like in the data Sheet but I don't ge a sinus waveform more a Square signal.
I think my LEM is damaged.

Can someone verify that I do it right?

Your first scope capture looks better, with minimal spikes.  H-bridge output should switch slightly before zero current.
Best to do another final adjustment after H-bridge is running at full voltage.  IGBT timing changes with bus voltage.

Ok Thanks turn it back a little... and check it also with higher voltage!
I only use 100V DC on the bridge with 3/4 pulswidh from the OneTesla interupter...

[flyingperson23]

Data Sheet say up to 150kHz

The hass 300-s lists up to 240khz and the signal resembled a mangled square wave at 200khz. Any particular reason you're using the lem? It seems like a standard CT would be cheaper and more reliable.

[Anders Mikkelsen]

Can hall effect sensors measure fast enough for drsstc feedback? It looks like this one is managing but I tried a hass 300-s and it wasn't anywhere close to fast enough.

Data Sheet say up to 150kHz

https://www.lem.com/sites/default/files/products_datasheets/lf_1005-s_sp33.pdf

Sure, but what's the phase behavior at tesla coil frequencies? They don't specify it, but it could easily be problematically high.