1. Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
    Dismiss Notice

Solid State Tesla Coil Burning Mosfets

Discussion in 'General Electronics Chat' started by Nissan20det, May 24, 2017.

  1. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    Sry I missed this post. Does this still apply that I fried all 4? Seems like it, you tell me...Okay, for sure, Ill lower my 50% duty on my reset pin to 25% arduino signal, and Ill test that on my low power supply. I will make a video to show how as I add turns on primary it just dies.. maybe is just too weak of a power supply?
     
  2. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    Okay this is interesting. I took another look at the CT after replacing my mosfets. I decided to look at the signal out of the inverter that the CT is making. I had it running off my antenna, sent my CT output through two inverters and scoped that second inverters output, open(to scope only), as well as my antennas output from the inverter. My CT signal was almost at 98% duty on time (Pic below). For some reason it wants me to have my DC blocking cap prior to my D10-D11 diodes. (Lonescience has the cap after). Once I placed the cap prior to the diodes I got a square wave that was identical antenna inverter output, it just was about 1Us faster. So this showed me I am getting somewhere.(Pic bellow "CT_InvOutOpen_Antenna_InvOut.jpeg." CT inverter output is blue, antenna is yellow.)

    Now I tried flipping phase anyway possible. Reverse secondary ground direction through toroid, one inverter not two, then swapped pin 1 with pin 2 on the CT. Still nothing, what the hell. I then tried the circuit bellow and it worked.
    schemeit-project.png
    Although this runs like this, with CT straight to gate driver, this is not something to do on any higher power supply. (pic below CT_Out). This shows a sine wave with the top and bottom cut off(clamped by diodes) so it is slightly a square wave but the rise and fall are slow."curved". This signal is not being cleaned up by the Schmitt making it dangerous to use at high currents. So any thoughts on this? Why I cant get this signal out of either type of inverter to work when connected to the driver? This doesnt make since to me, so I'll try moving it on my breadboard, maybe a portion of the board is screwed. The inverters will clean it, as well as output the CT sig but only connected to the scope. If I connect the CT inverter output to the gate driver it just goes flat. If I place my scope on scan I can see it try when I first power up the driver, like a couple pulses then flat.

    A side note here, You say my scope sucks but from what I see the 74HC14 is faster but the CD40106 outputs a 30% cleaner signal from what I see...Is this chip fried? I have another to try..pictures are labeled below.
     

    Attached Files:

    Last edited: Jun 5, 2017
  3. DerStrom8

    DerStrom8 Super Moderator Most Helpful Member

    Joined:
    Aug 9, 2010
    Messages:
    5,512
    Likes:
    507
    Location:
    Vermont (GMT-5:00)
    Your scope does "suck" and you're still not seeing the entire picture because of its low bandwidth. It's incapable of capturing the high-speed events that are very important to see when playing with Tesla coils.

    Did you remember to change the logic power supply to 5 volts (keep the driver supply at 12 volts) when you switched to the 74HC14? If not, you definitely fried your 74HC14.
     
  4. dave

    Dave New Member

    Joined:
    Jan 12, 1997
    Messages:
    -
    Likes:
    0


     
  5. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO

    Yes I absolutely did not forget to power it with 5v, as well D10 went to 5v not 12v. I will try a second chip tho.

    dude you have gatta get over the scope, Yes it does not have that high of bandwidth, But put this next to a benchtop 1Mhz scope you would be impressed. So for what it can do, it does not suck. I think it just sucks in your mind because its not the proper scope for this project, which your right. I got this for Vehicle testing and it has never failed me, not once. Maybe if you tried one your opinion may change. I will most likely get a higher bandwidth scope soon. I am loving this hobby and my later want to make a DRSSTC.

    Any thoughts on the CT working like that?
     
    Last edited: Jun 6, 2017
  6. DerStrom8

    DerStrom8 Super Moderator Most Helpful Member

    Joined:
    Aug 9, 2010
    Messages:
    5,512
    Likes:
    507
    Location:
    Vermont (GMT-5:00)
    I don't need to try one because you're already showing me images from it which prove my point (though you have no point of reference of what a good scope should show in this case, so you aren't seeing how awful it really is). Sure, it may do the trick for basic projects, like automotive ones. I'm not complaining about that. I'm complaining about you insisting to use it for this project where it isn't even beginning to do what we really need it to do. I certainly encourage you to move forward with a higher BW scope. A simple Rigol DS1052e is one of the cheapest on the market and yet will do most of what is needed for Tesla coil projects and a wide variety of other uses. I wouldn't get anything cheaper though, because it is on the brink already.

    I'm wondering if the junction capacitance of the clamping diodes is coming into play (I'd have to do some more digging into datasheets). If the lower one was acting as a capacitor it would shunt the output from the CT to ground, thus giving you a flat line into your logic. I have never seen this happen, and I wouldn't expect the junction capacitance to be a problem until you get into extremely high frequencies, but I can't say for sure. Then again, if this was the problem then I wouldn't think placing the DC-blocking capacitor before them rather than after would make much of a difference.... I'll take a look at the datasheets when I get home from work if I think of it.
     
  7. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    Okay sweet thanks, I am still on the 1n,4148's. I tried a bunch of those 5819's and they wouldn't work. Looks like the other 74hc14 is working better, still won't work with the CT and driver. The ct output looks great through the inverter by itself. Very confusing that I get a inverter output but not once that output is connected to the driver..

    Okay cool, I will look in to that model for my next one.
     
  8. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    Hey so this is a thought I had about what you said earlier:
    You said P = I^2 * R for Mosfets, So Bellow I wasnt doing that.
    So is it really P = 13a^2 * 100v ? cuz thats 16900W/4 = 4225 watts a piece? I'm lost....LOL sry.... But shouldnt I be close to popping these at 15amp? Sure they say 50amp but 250v @ 50amp aint anywhere near 400W which is there rating.. Just on my mind since all four blew, thats never happend, and it seem to run great.
     
    Last edited: Jun 6, 2017
  9. DerStrom8

    DerStrom8 Super Moderator Most Helpful Member

    Joined:
    Aug 9, 2010
    Messages:
    5,512
    Likes:
    507
    Location:
    Vermont (GMT-5:00)
    P = I^2 * R, not P = I^2 * V.
     
  10. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    ahhhh okay thks!!!!

    So I finally solved my CT issue Ugh!!!!!!!Yay!!!!!!Okay maybe you can explain whats going on here, I have 3 schematics that I'll explain. Inverters are now 74hc14's, on 5v rail, with D2 - 5v. D1-D2 are 1n4148


    #1
    Okay so I started to notice that as I changed the DC blocking caps value, I could get it to slightly spark up, sometimes it would hold steady. If I came close to the top load with a ground probe it would spark and go out at a certain distance.
    schemeit-project (5).png


    #2
    Now this is when it actually worked, Great actually!! I have a momentary button on my driver supply at the moment. as fast as i click it works every time. As well I can come as close as I want to the top load with a ground probe. So this setup acts exactly like my antenna, slightly better actually, If I quickly touch the ground probe to the topload it would loose signal on antenna, Now its solid!!!
    schemeit-project (2).png


    #3
    Okay, so in this one I had to reverse the direction of the secondary "ground" through the toroid to 180 my phase, This alowed me to add another inverter. Now this one does work although it will hesitate to start up sometime. I cant click my driver circuit on and off fast, after turning off I have to wait to turn back on. It also sometimes wont start right away.
    schemeit-project (3).png

    Now is this okay to leave this going through one Schmitt? Two cleans a bit better right? Can you see whats going on here? If so and two is better then one, do you know how to make the two schmitts act like one?
     
  11. DerStrom8

    DerStrom8 Super Moderator Most Helpful Member

    Joined:
    Aug 9, 2010
    Messages:
    5,512
    Likes:
    507
    Location:
    Vermont (GMT-5:00)
    The number of Schmitt triggers determines the phase of the output. If you only have one it inverts the signal from the CT (shifts it roughly 180 degrees out of phase). Adding a second trigger puts the output back in phase with the CT. If it works with one Schmitt trigger inverter, but you want to add a second one, then swap your CT leads and add the second inverter. This will put everything (roughly) back in phase. I say roughly because there are delays caused by C1, R1, C2, and your Schmitt trigger inverters that will also shift the signal slightly.

    What you've done with C2 is created a low-pass filter with a critical frequency fc = 1/(2*pi*R*C). The critical frequency is the frequency at which the output amplitude is decreased by -3dB. Beyond that point, higher frequencies will cause the output amplitude to be attenuated drastically. Simply put, it will try to "block" frequencies beyond the fc. Assuming R1 is still 1k, with a 220pF capacitor your fc = 1/(2*pi*[1k]*[220pF]) = 723.4kHz. With this filter you may be blocking the harmonics that Guangyan mentioned in his notes on his schematic. However, you are still creating a phase shift, as mentioned earlier (the time constant of your RC is 220ns!). There is a good chance that you have added so much delay that it can no longer trigger at the correct time in order to sustain a high voltage arc. If your MOSFET switching waveform is out of phase with the actual current resonating in the secondary circuit, the waveforms will cancel each other out and you will lose your output. This is just a guess, but it would make sense. I suggest getting rid of as much delay as you can. Remove C1 and C2 and stick with one Schmitt trigger inverter (this should be sufficient to clean up the signal) and just make sure your CT leads are connected in the correct phase. Try this and let me know what happens.
     
  12. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    I have always been paying VERY big attention to my 180* shifts with toroid/connections/inverter count.

    I think you didnt understand what I meant.. I am not flipping 180*... I was on one schmitt, then ran the secondary ground through the toroid in the opposite direction i had it. This flips 180. Then I added another schmitt, this adds another 180*. Now Im back "360*" to where I started. It runs very crappy with two schmitts(on same Phase cycle). I think it changes when adding one more schmitt due to the propagation delay the second adds.

    I really understand you know a ton more then I do about this stuff but the one thing that has held us back this entire thread is you assuming I don't know something therefore I have to re-explain it. I understand my grammar might suck but you would have seen the part about the secondary ground if you where not assuming everything. I keep feeling like an idiot because how you talk to me but three posts later I finally get you to understand me. I dont see electronics like you I'm sry, few will. I'm nnot trying to start a fight here but this all started from post one, how you thought this was a DR that would have save a page of going back and fourth. I mean theirs more then one may to skin a cat, Just because you can tell I know less doesnt make me feel good when you skim over my post and say Im the one missing stuff.

    It will not work with no caps and through one schmitt, it will on antenna.... you saw my Previous phase shift on CT I'll test with my #2 setup and see If its less then the 1us the last shift was. I have tied a lot of combinations here and out of 115 this is the first one(#2)...

    What I was asking in that post is if it working with one schmitt is okay? Also my resistance has been changed to 56ohms. So 1/(2*pi*R*C) = 12918.42hz. does this make my RC constant better? Maybe this is moving me 180*out and if I invert that I'm back in to phase.?

    But over all #2 is the best I have came up with, what you suggested has been tried and will not work.


    Thank you so much again for the help
     
    Last edited: Jun 6, 2017
  13. DerStrom8

    DerStrom8 Super Moderator Most Helpful Member

    Joined:
    Aug 9, 2010
    Messages:
    5,512
    Likes:
    507
    Location:
    Vermont (GMT-5:00)
    No, I read this thoroughly and understood it. You have to understand that you are not the only one reading this thread. Sometimes I need to explain certain concepts for the benefit of other current and future readers. The point of my post was mainly to look at the delays caused by the different parts of your circuits (i.e. the propagation delay of the Schmitt trigger inverters and of the RC filters). I think you need to go back and reread my posts as you clearly are not understanding the intent behind them.

    I am assuming very little. I take what you write, I read it (often two or three times), and then I respond, sometimes adding background information for other readers, and to verify what you have already done. What has been holding us back this entire thread is you refusing to provide all of the necessary information. You STILL haven't posted any detailed photos of your setup, and all of the schematics you have posted are either incomplete or have been changed in your actual setup. All that on top of not using the right tools for the job makes it very difficult to help you.

    That's another thing, without detailed images from a high bandwidth scope, it's difficult to track down where your problem is. I know I've been saying this over and over again from the start, but that's only because it is more important than you seem to realize, and you just keep blowing it off. "Oh, this scope has worked fine for me in the past." "Oh, I plan to get a better scope sometime." None of this is helpful NOW, and expecting detailed help without providing decent waveforms, schematics, photographs, etc. is unreasonable.

    I answered this in my last post.

    Changing the resistance to 56 ohms definitely helps the delay (your RC time constant is now 12ns) but now you have a filter that is practically useless (and FYI the fc with an R of 56 ohms and a C of 220pF is 12.918MHz, not 12.918kHz). Your filter wouldn't begin having an effect until the ~75th harmonic, which is absolutely useless. Theoretically you'd be better off taking it out altogether, which is what I mentioned earlier. We would need to determine the reason why adding a filter seems to be helping, and solve the problem further up the chain closer to its source. I think it's sheer chance that it is actually helping. The actual problem is most likely much deeper and you're just sticking a band-aid on it.
     
    • Thanks Thanks x 1
  14. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    you misread his and my schematic, stating I'm Making a DR.

    Then you assumed I was only flipping it only 180*, not 360*. this was two posts ago...

    I did state I reversed my direction for secondary ground through the toroid. and added a inverter, this 360*'s my phase.

    Maybe your not seeing posts as well but those examples are you not reading and assuming, Clearly.. Trust me I have missed alot of things in this thread.

    This is absolutely false. Maybe on the first page but I would say your DR assumption was the BIG delay here.. I understand you need you need info and pictures, I dont recall one time you have asked for something I have not given. When I made my first post, I explained in detail that other then the bridge, which I made a schematic... I changed a few parts in his schematic. I explained exactly what parts......Another misrepresentation on your part is I did state in the start, I replaced his arduino with a multivib at 50%. you assumed I connected my output signal other then I stated. I said this Is the same but with a multiivib generating my signal.... But I dont recall one thing your still waiting on. Did I miss something again?

    I 100% agree with this, I am absolutely patching a mistake, not fixing it. So lets make this easier, I was 100 of on my khz calculation sorry, but RC constant how are you calculating that I thought R*C but thats off? What should My filter frequency be? Is this complimentary to my RC delay? Are there values here I should be shooting for?
     
  15. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    I do agree with this and Im sorry Im not going to go buy one right now. I also understand this isnt helpful right now, but that wont change for a couple months.
     
  16. DerStrom8

    DerStrom8 Super Moderator Most Helpful Member

    Joined:
    Aug 9, 2010
    Messages:
    5,512
    Likes:
    507
    Location:
    Vermont (GMT-5:00)
    And as soon as I realized my mistake I acknowledged it and moved forward with the SSTC discussion.

    No I did not. I was simply explaining the concept, once again for other readers, of Schmitt trigger inverters shifting the phase, and the need to ensure the CT and output of the logic were in the correct phase with respect to one another. This was a general observation and explanation.

    I know, and that's exactly how I read it. See my above comment.

    I did not make any assumptions here. I know you used a multivibrator, and that was also shown in your schematic (which I appreciate, but wish it was kept accurate and up-to-date). This was never part of any sort of confusion. What post made you think I thought you were connecting the output of the multivibrator differently?

    PHOTOGRAPHS AND DECENT WAVEFORMS!!!

    100 off? Try 1000x. BIG DIFFERENCE!

    RC time constant is R*C. You said you have a 56 ohm resistor and a 220pF capacitor. 56*0.00000000022 = 12.32ns.
    I'm still not convinced you should have an RC filter where you have one in your circuit. I think it only just happens to work because it is shifting the waveform to make the timing match up correctly. This is a hack and should not be a permanent solution.
    If you use any filter, it should be sized between the fres of your secondary circuit and the first harmonic.
     
  17. DerStrom8

    DerStrom8 Super Moderator Most Helpful Member

    Joined:
    Aug 9, 2010
    Messages:
    5,512
    Likes:
    507
    Location:
    Vermont (GMT-5:00)
    Then consider this thread on-hold for a couple months. We can't move forward without having some waveforms that actually show what's going on. I will not post in this thread again until you provide detailed photos of your setup and some good waveforms of the signals throughout the circuit (GDT primary, GDT secondaries, bridge, feedback CT, clamped & cleaned feedback, etc).
     
  18. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    that was a typo yes 1000X, I had my pico decimal place off


    I posted videos, after that, I don't recall you asking for photos.

    Ill tell you right now I will not buy a scope just so you can help me.....I feel its a good idea, with your insight, to purchase a bench top.

    I scoped the antenna and my #3/#2 setup on CT they is in phase to each other. (antenna/CT) . this pic is labeled "Antenna-40106_CT-74hc14"

    When first getting this to work I did not have C6. Now that C6 is in here it works very well with 2 inverters.

    Okay here are scope view of right now. These are .SVG files, the open in any browser. Each file can be zoomed way in or out as well scroll quite ways off the screen(right). then you can hover over any part of picture with your mouse and it will display the voltage and time below. I have ordered these from start to finish. I will state connection as well as how it was tested (Signal - Scope connections)

    1. CT signals output - Dig Gnd / CT pin 1 (this one is a sine but cuts off at top, in file)
    2. Inverter out. Cleaned CT signal - Dig Gnd / U2 pin 8
    3. Gate driver out "GDT Primary" - Gnd OutB / Probe OutA
    4. Mosfet Gate T4 - bridge Gnd / T4 Gate
    5. Mosfet Gate T3 - bridge Gnd / T3 Gate
    6. Mosfet Gate T2 - T2 Source / T2 Gate
    7. Mosfet Gate T1 - T1 Source / T1 Gate
    8. Bridge Output - PriOUT1 / PriOUT2
    All Pins and scope views correspond to the new updated schematic below.
     

    Attached Files:

    Last edited: Jun 6, 2017
  19. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    Im sorry these other file types wont upload.... Do you have an email that you can PM me to send these? They are quite a bit better them .BMP.

    Here they are in BMP

    Im thinking about sticking with my CT, Let me know if this is reasonable. In the picture comparing the CT to antenna, It looks like the Phase shift is equal to the antenna timing(this was over 1us off before). Then in the two .PNG files, If you look at the center wave, you can see the CT is outputting better waveform. Its a lot straighter up and down then my antenna. As well my filter should be at a reasonable frequency now that the 960pf has been added after the two Schmitts. Does this Add to my delay that was 12ns? also making this a bit more logical?
     

    Attached Files:

    Last edited: Jun 6, 2017
  20. shortbus=

    shortbus= Well-Known Member

    Joined:
    Mar 23, 2008
    Messages:
    1,230
    Likes:
    77
    Location:
    youngstown, oh
    As someone that has followed this on both AAC and here, and have seen the results from both threads, I can't understand why you are unhappy with what Derstrom has done for you. This is much farther along than it was at ACC, due to one person, Derstrom. This is not my area of expertise, but one of my many interests, I try and follow every thread on both sites on any Tesla coil build. Derstrom is the most knowledgeable person in all of them, listen to what he tells you and learn.
     
  21. Nissan20det

    Nissan20det Member

    Joined:
    May 24, 2017
    Messages:
    95
    Likes:
    1
    Location:
    Parker, CO
    Absolutely. I am EXTREMELY thankful for what he has said. I Never Said I was unhappy about his insight, at all. It just sucks sometimes, having an ignorant person talking to a smart one. Conversations will occasionally run in circles. As well getting replays in a way of speaking to other people at the same time, in the same sentience, confuses the crap out of me. So I apologize that I didnt understand what he was doing. This was the entire reason I thought I was being talked down to.

    Those photos are the best I can do as well as an updated schematic.
     

Share This Page