Author Topic: Stan Resonance finally cracked! by Sebosfato  (Read 40732 times)

0 Members and 1 Guest are viewing this topic.

Offline Login to see usernames

  • Global Moderator
  • Hero member
  • ****
  • Posts: 3607
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #136 on: March 23, 2011, 16:58:04 pm »
Irf840 gate burned while working without snubber... The signal without the snubber were much better.. anyway not going to do this anymore...


I was thinking about using a 100k resistor between the gate and the variable dc voltage source, what you think about this? Could this help somehow? anyone?










Offline Login to see usernames

  • Sr. member
  • ***
  • Posts: 434
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #137 on: March 23, 2011, 17:21:05 pm »
Now seems that i could put the circuit to work better, the diode in parallel with the primary now has 1kohm resistor in series!!! without it the transistor didn't worked so well. And is not perfect yet, but at least can get up to 50 75 v without the 120hz behind. Now only 100ma flow...


My oscilloscope wasn't working again (burned resistor of one chanel) already changed.


so the primary now has the correct orientation in relation to the bobins.


I could get up to 200 volts in the tx5 (i have no secondary connected yet... i was applying around 50v 100ma to that primary, the coupling seems to be low.


I don't know what is happening, when i raise the voltage the frequency gets dirtier.


I'm now using irf840 it runs absolutely cool for now. Up to the point where the snubber capacitor blows... The good thing is that the mosfet now is not burning since i started using the diode in parallel with the primary, but it was not good cause was blocking the resonance, and also making the transistor to behave strange, so i added a resistor in series with this diode and so both resistor and diode are in series, wired across the primary. .
Why are you using 50V for input to the primary? Isn't 12V what Stan used?
The magnetic coupling could be low due too thick plates. Are they laminated iron plates?

Br,
Webmug

Offline Login to see usernames

  • Member
  • **
  • Posts: 184
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #138 on: March 23, 2011, 17:24:04 pm »
Sebo,
the resistor will help but try to keep the gate voltage at 5 volts or so.  I'm smelling lots of ozone from my tinfoil wrapped cell.  btw, do you have a user guide or instructions for using your pll?  I found some time and got it done and was wondering why not just pick up a signal from the 8xa coil by using another coil just sitting next to it?
I use a irfbc40 mosfet, the fastest one I could find and rarely burn them out.  11ns switching
your thoughts?
kb

Offline Login to see usernames

  • Global Moderator
  • Hero member
  • ****
  • Posts: 3607
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #139 on: March 23, 2011, 18:45:44 pm »

Hi webmug,
I'm using a variac with a full wave bridge rectifier as the source of the electricity, i found it is handy to adjust the input impedance up to the limit of the core saturation, for example, at different frequencies the core will saturate at different voltage per turn in the primary.. so you just adjust the voltage up to the noize in the core is clean, is where the max voltage is found in the induced coils...


I think that i was over saturating thats maybe why i was burning so many mosfets, now that i'm able to measure the secondary voltage i'm seeing what happens when i push up the volume..



I'm concerned about the fact that the source impedance must be high enough to allow the resonance to happen..


I think that this is the problem here, in my old experiments in italy the resonance wasn't in the transformer, i was using the transformer to drive the resonant tank in parallel, in series also worked and was much easier to design but in parallel worked better...


I had a vic like but only with the secondary and a small choke after the diode (i was using a toroidal ferrite core) the choke was about 30 turns..


The voltages went up to 10kv... that time...


I'm thinking about reducing the laminations number to try to improve the quality factor and reduce further the coupling of the resonance to the primary..


I'm thinking also about having a air core coil in series with tx5 to see if helps letting the resonance happens..



My laminations are oriented grain very thin. Those for audio transformer.. was the only i could get different from those thick ones intended for 60hz...


I understood but now need to think better about the reflected impedances of the secondary in the primary and tx5 coil, i think that now i know how to determine it, this would be handy to understand how much load is being felt by the tank. So how to determine the Q...


I'm using pure water now and when i measure the voltage for example the other electrode non connected in relation to the outside




Hi kickbackemf,

I had two tip35 transistors here and i have now burned one but i have a good news, it now works very solid with the last one having the diode in parallel with the primary having a 1kohm resistor in series than all across the primary like i described before... Oh it also need a diode between the emiter and colector, after i added this it didn't burned anymore.


I also have a variable 10 turn 1kohm resistor between the base and the emiter, and i noticed that if i take it almost to 0 the voltage induced is greater. So i'm thinking about adding a 50 ohms resistor in series with the primary to see what happens. To change the input source impedance....
 
I think that a high voltage bipolar transistor would be nice for this project. The tip35 suport 100v 25 amps, works a bit warm, buts ok, one thing i noticed is that when it get higher in frequency more current is flowing instead of reducing... this is somewhat strange but seems to be better to work at high frequencies you just need to turn the variac down a bit to control the saturation of the core...


You mean ozone in the tin outside the cell? Is it happening to you? the resonance? what are your voltage levels?


For the pll you can use this last diagram i just changed it again... The only thing you need is to add some capacitors in the boar that i din't represented in the drawing. you could use a fast prototype board is very easy...


Yes you are right just need to to pick that signal up and close the resonance loop. 




I will try adding more resistance to the base of the tip to try changing the impedance... The bipolars are very nice cause they will turn on at low base current but will only let a current proportional to the gate current to pass to the primary.- this is very nice. So you can adjust the impedance. I think we could use one to control the current in the positive side and the other just to switch. Like his analog voltage generator.

Offline Login to see usernames

  • Member
  • **
  • Posts: 184
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #140 on: March 23, 2011, 19:43:07 pm »
Sebo,


Thanks for the updated circuit.  Let me ask a silly question.  Since I run at 110 volts all the time through the variac would there be a problem I'm not thinking about in just connecting my bridge rectifier directly to a 110 volt wall plug?  The stong ozone smell is coming off the water bath.  I'm running 1700 volts as peak to peak on the oscope.  My latest winner in my assortment of coils is an iron powder three inch toroid reading 2.52H using 21 awg wound per 8xa.


kb, snow blizzard today so I'm home with the kids.

Offline Login to see usernames

  • Global Moderator
  • Hero member
  • ****
  • Posts: 3607
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #141 on: March 24, 2011, 03:23:18 am »
Hey kick, nice to have ozone water bath =) I'm using the variac 1° cause it allow you to adjust the impedance of the input impedance of the circuit, 2° cause it kind of limits the current to the maximum it can handle...


Do you know the current that is flowing there?


Does it behave this way only at a specific frequency? or voltage is high at all frequencies?




I just came back from the universe city =) and the other tip 35 blowed strangely... =(


Tomorrow i'm going to buy some more transistors and resistors capacitors and stuff.


I'm going to make a sculpture of the so many transistors i'm blowing hehe




Offline Login to see usernames

  • Global Moderator
  • Hero member
  • ****
  • Posts: 3607
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #142 on: March 25, 2011, 05:01:51 am »

Today i went to the engineering school, and talked to a doctor in power electronics, for about 4 hours.


He corrected my snubber diagram, gave me a snubber capacitor (68nf 1.6kv), told me that bipolar transistors would be worst than mos or igbts, and suggested me to use igbt since i work under 20khz cause as it is a combined mos and bipolar it has the best from both worlds it only need a field to saturate and have low conduction losses. He said that the bipolar transistors is not very good at high frequencies again. 


He explained me about the magnetizing inductance and leakage, and so i understood that the energy of the leakage inductance must go to the snubber, the diode in parallel with the primary should have no more than 10 or 20 ohm in series (not 1kohm like i had).


The snubber will than have a high speed diode, the capacitor and a 100ohm non inductive resistor. (this will not let the leakage inductance during off pulse to develop high destruction voltage between gate and source)


In parallel with the primary there will be the diode having a 10 ohm resistor in series.... (this is to prevent the demagnetization current of the primary)


About the diode between the primary and the transistor he said is useless unless you are going to make bipolar switching( half wave, full wave drive, to prevent reverse current)


If the secondary and primary are wound one over the other  the leakage will be smaller.


Now here goes something,




The oriented grain laminations is not only about the thickness of the material. The way i cut and made my close loop core is completely wrong cause the I laminations and the L's i made were not in the right orientation for the bobbins. He told me that there is a fabric here where i can have oriented grain core made in whatever configuration i want. So basically if you don't use the grains with the right orientation, you are not using it well, at that point he told me that would be better to use non oriented core. He told me that the best thing was than to use a flyback core as is not oriented.. .


He also told me that the the thing is working and i want i can go there to make the measurements on the unit using oscilloscopes and the equipments.


VEry very very nice pearson.


He also explained to me the positions where the fuses must go, oh boy many things.


He told me that the best thing was to use a flyback transformer as is not oriented.. .


sebs
k ai ser
cheers


I think the vic would benefit from having all the coils wounded one over the other. Cause this would minimize the leakage, losses and the resonance would therefore induce or exchange the energy with the other coils in a better way.


I'm going to try modifying my core to benefit from the oriented grain... And i'm going to try get to the point tomorrow or this weekend if possible.


Can't wait to see.


If any of you don't know where to get OG cores and or have intention to have a core made in the shape and orientation you want we could make an arrangement.

Offline Login to see usernames

  • Global Moderator
  • Hero member
  • ****
  • Posts: 3607
Re: Stan Resonance finally cracked! by Sebosfato
« Reply #143 on: March 25, 2011, 05:11:17 am »
The leakage inductance is like a series inductor with the primary and the magnetizing inductance is in parallel.


The fuse should go after the filter capacitor (after the variac and bridge) but the diode i had should maintain to protect my house line so he told me that the cap in parallel with fuse is not needed.


Also, the transformer feeding the circuit, (electronic) should go before the fuse, cause this can never be off when the variac is on and the voltage should be applied so only after the circuit is pulsing..



Almost forgot, between drain and source there might be also a high speed diode if is not already internally build in into the transistor.


So 3 diodes. (one snubber, one parallel with the transistor and the other parallel with the transformer.



$ 1 2.6E-7 3.7524723159601 53 9.92 50
v 160 288 144 432 0 2 3000.0 20.0 0.0 0.0 0.5
r 544 192 592 224 0 3000.0
T 432 192 544 320 0 0.04 10.0 2.420068870404547 -0.1676025239313738 0.5
c 592 224 656 320 2 9.0E-10 2993.2624381511214
d 256 208 368 192 1 0.805904783
c 368 192 368 256 0 1.0000000000000001E-7 -0.010644594757621917
d 368 320 368 256 1 0.805904783
r 336 272 368 192 0 100.0
w 336 272 368 256 0
l 544 320 656 320 0 0.1 -0.1676025239313738
r 320 320 368 320 0 2.0
r 368 320 432 320 0 3.0
w 368 192 432 192 0
w 144 432 256 432 0
w 256 432 256 384 0
v 144 192 144 112 0 0 40.0 100.0 0.0 0.0 0.5
w 144 112 256 208 0
w 144 192 144 432 0
d 256 352 384 352 1 0.805904783
d 256 384 256 352 1 0.805904783
r 256 352 384 384 0 100.0
c 256 432 384 432 0 6.8E-8 -0.05356864303075781
w 384 352 384 384 0
w 384 384 384 432 0
t 208 368 256 368 0 1 0.7205879924013274 0.7738753198778348 100.0
r 160 288 208 368 0 100.0
d 320 320 256 352 1 0.805904783
c 80 112 80 448 0 1.0E-5 100.0
w 80 112 144 112 0
r 80 448 144 432 0 1.0E-4
o 1 64 1 35 160.0 9.765625E-5 0 -1
o 1 64 0 297 320.0 0.2 1 -1
o 3 64 0 42 20480.0 0.1 2 -1
o 7 64 1 291 299.3155353253689 9.765625E-55 3 -1
o 15 64 1 291 2560.0 9.765625E-5 4 -1
o 19 64 0 291 320.0 9.765625E-5 5 -1
o 20 8 1 35 640.0 9.765625E-5 6 -1
o 5 64 1 291 160.0 9.765625E-5 7 -1


« Last Edit: March 25, 2011, 16:45:23 pm by sebosfato »