micro sstc

[niels]

Hi Everyone!
I’ve been trying to build a sstc for a school project for awhile now but it didn’t work yet.
I tried to build Steve ward’s micro sstc only with the icl7667 as a mosfet-driver. After powering it up the coil draw quite a few current and there were now arcs not even a fluorescent tube lid up. For safety I installed a 8oms resistor parallel to the coil to limit the current. Firstly I tried to measure the gate voltage at the Mosfet without any load attached  [ Invalid Attachment ] but I can’t tell weather the oscillations mind ore not. Then I did exactly the same but with load [ Invalid Attachment ]  I was expecting the frequency to change because the driver is feed by an antenna which should pick up the resonate frequency. Instead I just get some weird noise signals. The question now is: could the driver be the problem because it is not providing enough current? I tried switching to a smaller Mosfet as it should have a smaller gate capacitance but the same problem accrued again. The second question is where is all that excess power going? Im driving the coil with a bench power supply at 24volt and max current of 1,4 amps when driven directly without the resistor the voltage drops to about 4 volts ad 1,4 amps but the mosfet stays cool. I already tried to adjust the the primary coil(diameter and position) but without success
can anyone help me with this problem ? Thanks in advance !

[davekni]

Good questions to be asking.  4V at 1.4A shows that 5.6 watts is being dissipated somewhere.  My first guess would be U1 12V regulator and/or driver chip.  See if they are getting hot.

The antenna is much too close to the secondary.  If it did start to work, the secondary will arc to the antenna, which may damage the circuit.  That would be my first thought, that it worked momentarily, then the driver chip latches up due to excess input current.  Could be just temporary latch-up until power is removed or could have damaged the chip and/or D1 and D2.

Is the unused driver chip input tied to ground as the data sheet says to do?

The gate waveform ring may be enough to cause problems.  Does that ring show up on 12V supply?  Is C2 close to the driver chip with short wiring?  A separate capacitor across driver chip supply (+12V to ground) is a good idea either way, especially if C2 isn't very close.

Yes, IRFP460 is a large device for that driver, but I think it will work well enough.  I'd stay with IRFP460.  Good insight to check that.  You could reduce the value of R3 a bit to compensate for the driver's typical 4-ohm output impedance.

How close is Q1 to the driver?  Is Q1 source wired as the schematic is drawn, with one wire from the source to driver ground and a separate wire from source to C3?  Is C3 close to Q1 with short wires?

A couple close-up pictures of your board, top and bottom, would help if the above isn't enough to get things working.

Good luck!

[niels]

Thanks For the fast answer!
I replaced the chip and the two diodes to be sure thy aren't faulty after reducing r3 to 6 ohm the oscillating at the gate disappeared. but this measurement was taken with the source of the mosfet floating. after connecting it to ground as well the gate voltage looked like this. in the datasheet i found this could the grounding be the problem? I'm not sure how close to the chip q1 has to be, in my case there is only 1 to 2 cm of wire in between. I think the heat goes into the driver chip. in the first test i had 2 poversupplys with the negatives connected so i could remove the voltage regulator. in the second test i connected everything to one power source with 12 volts only to get the same issues and the mosfet driver got sigificantly warm. i now wonder if the coupling of primary and secondary coil in good enough because the antenna is not picking up any signal. could the problem be that the secondary coil isn't oscillating?(i already tried reversing the primary coil with no success)
Thanks for your help i'm  really appreciating it!

[davekni]

Are those scope plots on the driver chip output before the series resistor to the gate?  That is what they look like.  (If you are probing directly on the gate, I have a hard time explaining the shape - would imply significant internal gate resistance on the FET.)  Is the scope vertical channel set to "AC-coupled"?  The actual gate waveform should go from 0V to 12V, not -6V to +6V as shown.  Changing the scope setting to DC-coupled will show the actual waveform.  AC-coupled is useful for a few special cases only - measuring small AC signals on top of a large DC voltage.

The capacitor you have across the FET source-drain looks like one from a microwave oven, which is MUCH larger than the 0.001uF of C3 in the schematic.  Typical microwave oven capacitors are 0.5 to 1.0uF.  I suggest trying a value much closer to 0.001uF and using short wires to the FET.

Have you tried using JavaTC?  It's targeted to DRSSTC, but can be used to calculate your secondary resonant frequency based on number of turns and geometry (diameter, height, position and size of top-load).  Knowing approximate frequency will help, as you can then adjust 555 frequency to match.

I noticed that the two halves of icl7667 are wired in parallel.  That's great for drive capability.  It could cause an issue if the two halves don't switch at close enough to the same time.  Probably fine, but if you later end up with driver chips frying, that is a place to look.

[Mads Barnkob]

Hi Niels and welcome to HVF!

Microwave capacitors also have a built-in bleeder resistor and are rated at 50/60 Hz, they have terrible dissipation factors and impedance at high frequency. It is unusable in this application, you need a metalized polypropylene capacitor (MKP) in that place.

If this circuit gives you too much of an head-ache, you should build Steve Wards mini SSTC instead, his SSTC5 is the one he recommends as all SSTC1-4 are his development towards the "stable" SSTC5, all the previous circuits have their quirks and bugs that can make them hard to get going if the physical dimensions of the coils and wires are different from what Steve used. The SSTC5: http://stevehv.4hv.org/SSTC5.htm

The weird waveform... you are clearly hitting some kind of capacitance that presents itself as a short and takes a long time to charge, it does not correspond to any of the "normal" wrong GDT waveform shapes, so I think your problem is outside of the driver/GDT part of the circuit. Use these as references: http://www.richieburnett.co.uk/temp/gdt/gdt2.html

[niels]

Thanks For the Help you really helped me out!
didn't knew that microwave oven caps are not suitable
changing the cap to the right one did the trick now i get a few centimetres of sparks although i noticed that the position of the primary coil is very  critical. thanks for the tip with the sst5 design i will definitely try out this design!