New projet OmegaDR

[Mugi30]

Hey all, i'm new on this forum, following the advice of Mads Barknob,
I signed up and so I make my first post hoping to be welcome 
i have for project since....5Years (yeah) to make a medium/large Dual resonant coil..

I already have some of the parts following the pictures.(Primary, MMC, bridge, i will give specs of them 

the bridge is made with SKM300GB123D, i try to make 40-50khz Resonant frequency, theses IGBT are not verry fast

2 MMC made by me

the grey mmc are arcotronics R76 1.2uf 630VDC, 12 serie of 5 parallel, 7560vDC and 500nF in total (each caps have 450 dv/dt,  540A pk,7.5A RMS, 2700 Apk, and 37.5A RMS total)

the white MMC use multicomp pro caps, 1200v 1uF each, 9 serie of 4 in parallel, 10800Vdc, 444nF in total, (each caps have 1000Apk, 25A RMS, 1000 Dv/dt, for 4kA Pk, and 100A RMS)



the oscilloscope with damped sine wave is a test when i connect a CT with a primary circuit on a DC bus voltage, just for see if the core is okey.
The scope with square wave is a test with GDT
all cores used are TDK N30 cores, so permeability are 4300, is it ok?


I will update as I go along
advice and help are welcome

Thanks!

[davekni]

It's looking good so far.

[Mugi30]

I hope so too  .
besides yesterday I did a bizzare test by testing my CT connected to the primary LC circuit, I noticed something weird but I think it's due to the inductance generated by the CT, I have a frequency of 40khz when the output of the CT is not short-circuited by a resistance (resistance load for example) but as soon as I add a resistance at the output, already I have a more stable start of oscillation, but the frequency that give my CT up to 50khz ??
an idea? would it be the CT which adds an inductance which causes it to lower my frequency of the LC circuit by 10khz? and in fact I have a frequency of 50khz at the base?

[johnf]

Your CT will add series impedance to your measured primary at the square of the turns ratio. this is why it is important to shunt the CT with a low R value ie 1 or 10 ohms to make measuring easier and not create added series impedance to your primary circuit.

[Mugi30]

thank you for your precious answer! it seemed to me that there was a history of impedance on the measure and on the level of the shunt!
suddenly my real resonant frequency will be rather 50khz suddenly because I shunt the CT

[Mads Barnkob]

Hi Mugi30 and welcome to HVF!

It is all looking real good and your white snubber capacitor MMC is just perfect for a SKM300 brick coil.

Seems you just need to settle on a secondary + topload, wire it all up and you are ready to make sparks! (do not feel back about the 5 years, I spend 4 years on my large DRSSTC )

[Mugi30]

Hey all, i just post an update, the pictures talk about them lol (debugging)
after soldered all, i made first test with function generator and oscilloscope
at the first test, the driver worked but at CW, I didn't know where it came from. after a small reflexion,
I thought of a pulling resistance to the ground which was missing, and it was the case looool!
in the pictures, my oscilloscope show directly the signal with a IGBT connected to the GDT, (here is a SKM bricks) with a gate resistor and just a 30v diode TVS.
i just have a priblem, i have a damped sine wave after each cycles, with 800mv peak... is it dangerous? i dont have it if i test the GDT imput of the driver..
so  i have it with the GDT output too.. is a really problem?

the feedback works well if the peak voltage of of the CT simulated to my function generator  is 3v minimum, ! if the voltage down, i ear a BRRRRBRBRB due to the instable signal.

pictures (damped sinewave is the picture when i measure it with curcors)


+also, the IC who driving the feedback is not a LT1016, TL3116 or MAX913, this is a AN8561ANZ, it works verry well!

[davekni]

Unless your driver circuit is self-oscillating, 3V minimum CT feedback is likely to be problematic.  Most driver circuits require the first H-Bridge half-cycle to generate enough current to trigger the next half-cycle.  The first half-cycle isn't likely to generate 3V on your CT.

The gate signal ring-down looks like the gate-drive has an emitter-follower output stage, or something other situation where the gate voltage isn't hard-clamped to 0V between bursts.

[Mugi30]

Can you explain more about this  damped sinewave/ gate signal ring down? how to repair it? i added a ferrite core for parasite and twisted output stage, but nothing changes... i don't understand very well about the "emitter follower output stage"

what's the minimum vshoult it read to the feedback circuit for starting  ? I've seen the circuit clamp the voltage to ~ 1.6v approx with resistors and diodes..
thanks! 

edit: the 3v was required because the burden 51 ohm of CT circuit resistor was connected !! i have the signal at 1v too now! and probably with more low voltage!
i just need to resolve this problem of damped sinewafe after each cycles... this is my single and last small problem, someone say me that's not a big problem for IGBT...

[Hydron]

If the damped sine wave is on the igbt bridge output after switching is stopped then that's normal - almost no current is flowing, you're just seeing the last dregs of energy bouncing around the parasitics of the circuit.

Just a note about the CT - the general rule is to never use a CT without a burden resistor - it is quite capable of also acting as a voltage transformer (up until it saturates anyway) so you can end up with significant voltages on the output!

[Mugi30]

hey Hydron!
the damped sine wave is at the outpout of the GDT. i dont checked the output of the fullbridge of SKM300GB.

[davekni]

Do you have schematics for your gate driver ECB, or at least the output stage?  That would likely make the CT output damped ringing more clear.  The 0.8V signal by itself isn't an issue, as it's not enough to turn on an IGBT gate.  My only concern would be if that damped sine wave is an indication of a more serious issue.  If the driver is designed in a way to clamp gate voltage only to +- a diode voltage (ie. emitter follower), then things should be fine.  If the driver is supposed to be clamping with a FET source-drain (as is more typical), then the damped sine wave indicates a problem.

[Mugi30]

here is the schematic of the driver.
That's from ProfDC9 ud27-schematic.pdf

the difference i dont think that's from that, but, mine for C34 is not a 1uF, but a 4.7uf film cap. with its 15ohm serie resistor (R24)

At the stage of power supply with LM7824, here is at output a 4.7uf with a serie resistor, i dont put them on the pcb too (R1, C4)

[Hydron]

hey Hydron!
the damped sine wave is at the outpout of the GDT. i dont checked the output of the fullbridge of SKM300GB.

I think this will be parasitics again, probably the leakage inductance of the GDT (which is shorted on it's primary by the driver FETs) resonating with the IGBT gates, with damping provided by the gate resistor. Should still not be an issue given the low amplitude (not even close to turning on an IGBT).

[davekni]

Are you using the non-inverting driver chip UCC27424?  If so, that explains the ring, which would be GDT inductance with C33+C34 capacitance.  I think that UD2.7 design is intended for the inverting driver chip UCC27423.  That way the preceding AND gates will force UCC27423 outputs high between bursts, which forces both GDT terminals low.  You could scope directly on one of the GDT drive terminals to see if quiescent is low or high.  If high, the PFET (Q1 and Q3) gate drives are on (low), but will charge high (off) since they are AC coupled.  That will make start-up of the next burst problematic.  (I think there is a typo in the schematic you linked, listing UCC27243, which doesn't appear to exist.)  If you are using inverting drivers (UCC27424), you could change the AND gates to NAND gates (and bypass inverter U8D).

[Mugi30]

i use TC4423 IC, that's 2 inverting driver

i think he mean UCC27423 and not 27243 yeah,

[davekni]

The only reason I mentioned "UCC27243" is because that is the label for U9 in the schematic link you posted in reply 13.  Yes, I believe that label in the schematic is an error.

TC4423 should work fine.

I was thinking incorrectly.  C33 and C34 are still in the GDT path even with both driver outputs low, so that ring would be normal, nothing to worry about.

[Mugi30]

Nice! i just need to connect each component together for my dual resonant coil, i just need to finish my secondary coil. Basicely, i wanted to make my secondary coil with a 200mm diameter.
I ended up finding only 160mm in diameter. 200MM and more is difficult to find.
is it serious to have a secondary length 6 times its diameter?

it is said that a dimension of secondary coil must need in maximum length 5 times the diameter ...


I saw Tesla coils break this rule and still work well

Mugi

[davekni]

My DRSSTC secondary is 157mm diameter and 1100mm long and works fine.  Larger diameter would be nice for higher coupling factor, but I managed to get 0.14 with this coil.

[Mads Barnkob]

Nice! i just need to connect each component together for my dual resonant coil, i just need to finish my secondary coil. Basicely, i wanted to make my secondary coil with a 200mm diameter.
I ended up finding only 160mm in diameter. 200MM and more is difficult to find.
is it serious to have a secondary length 6 times its diameter?

it is said that a dimension of secondary coil must need in maximum length 5 times the diameter ...

I saw Tesla coils break this rule and still work well

Mugi

Nothing is set in stone!

My recommandations in the DRSSTC design guide ( http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/secondary-coil/ ) follow the best practise, known to work designs and is a guide to build something that is based on experience from many different builders. It is not an absolute list of solutions and that is also why I write "The color coding of the secondary system impedance should be understood as a guide line towards the magic 50k which was explained further up. I made a spreadsheet with all these and more example calculations. The magic 50k is thought to be too high for large or very large coils that will experience serious spark loading, here it properly is more suitable with 40k or even as low as 30k. Experimentation is needed to confirm these assumptions."

what about wire size? what is your estimated impedance of the secondary coil? If it lands inbetween 30-70K I would not worry about changing anything