Stan Resonance finally cracked! by Sebosfato

[Login to see usernames]

I think that i finally understood the idea behind the resistive coil of wire... Is only for limiting the Q factor of the resonance and localizing power dissipation... If the Q is too high you are not able to tune to resonance. so i'm going to add a resistive coil of wire, i have a wire here that has 30ohms per meter so maybe 3 meters is ok. this will be added between the tx5 and the cell, to reduce the Q factor.


I think that the secondary also act as a load to the resonant tank what should reduce the Q factor too and also the resistors in parallel with the primary...


I'm going to build every thing that i will use from resistors and capacitors for the snubber, cause i don't really have any other money to invest. In my previous pll i had a hand made capacitor that i had to make 300pf , very easy with pp film and copper foil that i have here..




My numbers here i have 1,1H and 330pf in the tank and around 30ohm so i will add 100 ohms in series to reduce the Q with was around 2000 and the frequency around 8300khz i'm not considering deviation in frequency cause of the load so is not an accurate value probably higher than that if you think well. So if i reduce the Q to 440 by adding more 100 ohms will be far easier to tune in. And from my calculations dissipating 93watts in the tank you would reach 50kv with a recirculating current of 850ma between the cell and the tx5 coil.



I'm also not considering the capacitor losses..

I'm not considering also the reflected load of the secondary witch should also reduce the Q.


I'm thinking about winding up a new secondary with thicker wire so i have more current available for the cell but i'm not sure if is actually necessary. The resonance should induce in the tx4 coil already a nice impressed voltage. I think that i will need to add also coils for the positive side on the other core too to have more load in form of voltage to the cell.

[Login to see usernames]

I made the snubber components today, i made two 450pf capacitors and two non inductive wire wound 100 ohms resistors. And also a resistive choke having 200 ohms.. These resistances can be easily varied.


Now i need to test to see if the resistance wire will suport the power dissipation.

[Login to see usernames]

Fabio,

I seen schematics adding zener diodes across the fets.
That is done to protect the fet against HV surges.
Have you looked into that?

Steve

[Login to see usernames]

Hi Steve

I thought yes but I could not find Zener that bear more than 200v, seems in the irfp840 that there is already a zener between drain and source. But I think the problem is happening cause the rate of change in voltage dv/dt is too high and together with heat this may cause a spark inside the mosfet, so i need to get around the transients, for this i will put a diode in parallel with the primary witch is one way to prevent that when the pulse is interrupted that the coil charge the drain source capacitance of the mosfet as result of the kick back effect... Actually when resonance happens i think there is not going to be any hard switching, cause the resonance will prevent the kickback (only at resonance) in the primary if you add an oscilloscope you will see the double frequency, also in the secondary and the tx4 coil, the tx5 coil will still have the same frequency you a re applying but ac.. 


I will also add a 450pf capacitor (hand made) in series with a 100ohms non inductive resistor (hand made) all this in parallel with drain and gate, all this with the hope to  stop burning mosfets.


I will try this simple snubber, if don't work I still have one other 840 so I will try another snubber diagram that uses yet a diode in it. Is called RCD snubber. Not sure if will be needed.


Other thing that concerns me is the fuse blow of the variac that is blowing also the mosfet. I'm going to add a resistor in parallel with the fuse but i'm not sure if the fuse develops a voltage in it before blowing. So i'm going to try this to see how much voltage it needs to blow. Than I will add a resistor that don't let to much current to flow.. . Is just one way i thought to be able to not have all that interrupted energy to discharge all in once in the mosfet


I'm wiling to apply up to 250v 2amps in the primary so I need to be 100% safe that the circuit will not interrupt functioning, so tests can't be stopped cause of this. This is important cause if I go to the university make the tests I cannot stop the experiment many times to change the mosfets or fuses... or the teacher will laughs me. 


Steve do you have any idea of the kind of zener? part number? I could try to find if you tell me...


Thanks


Sebs


(http://www.softwareforeducation.com/electronics/notes/AS/mosfet/switch.gif)
 

[Login to see usernames]

Hello guys, i'm working on the thing here and i have the following situation.


The snubber worked ok, the diode never gets hot, i burned a irf840 cause i wanted to try lowering to the maximum the gate voltage but it got burn so always should be driven at 10v min, but again seems to work ok now with the snubber.


Except that the mosfet seem to not completely switch block the 120hz from the mains rectified, like i also pointed before during 0% PW the mosfet still conducts..


I don't really know how to solve this. But i guess the only answer would be to try bipolar transistors.


Anyone any idea?


I now have only a 20ohm resistor between the driver and the gate. And seems to work fine. but still have the superimposed 120hz current.


Maybe PNP?


Any idea?


Please?










 

[Login to see usernames]

Why don't you just use a mosfet driver? I use a FAN3121T.It will need to be mounted on a breakout board because it is a small surface mount IC.A place called sparkfun.com sells the board.
 
Why aren't you filtering out the 120 hz pulses from your vari-ac? Making it a clean DC voltage.You don't need the pulses on the positive side of your primary coil.
 
Don

[Login to see usernames]

Sebo,
agreed...
Use a car battery or something like it to power your pulser seperatley and connect it to your 120v variac ground after the bridge to the battery negative and your pulser negative.  A mosfet driver is useful as well and saves the rest of the circuit should things go bad.  The car battery works well for me.

[Login to see usernames]

Hy don,


I'm using IR4427 driver but is not turning off the mosfet anyway maybe i will try to find another one with more current peak,,,


 Hi kickbackemf


thanks for the idea, the battery should filter the 120hz i think... Good idea.


Thanks don and kick