Interesting disappearing problem with my Sstc / Tesla Coils / Forums

Forums

4hv.org :: Forums :: Tesla Coils

« Previous topic | Next topic »

Interesting disappearing problem with my Sstc

1 2 next

Move Thread

LAN_403

brandon3055

Wed Jun 06 2012, 03:19AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

You may or may not have seen my audio modulated run on YouTube well when I made that video about a month ago I had a problem ocur with my bridge in order to run it for the full lenth of the song without overheating I had to run it at half power which worked grate but when i switched it back to 240v two of my mosfets blew and when i changed them and ran it again (for about 3 seconds) two of the mosfets and fast diodes were to hot to touch so I asumed that I had damaged one or more of my puny 3A super fast diodes so i decided to leave it until I could get sum descendant ultra fast's but the wird thigh is when I did a demonstration with it yesterday everything stayed completely cool even for a 30second run dose anyone have any idea why this might be (it guess its possible it just wanted to look good in front of an audience)

Sigurthr

Wed Jun 06 2012, 08:46AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

Have you considered trying another fet besides IRFP460? I have very limited experience with power fets but the experience I do have is mostly with IRFP460s, and they always seemed to fail for me in a manner consistant with what you describe - somewhat random. In identical setups like with yours where you simply swap out the fet (btw, I really like your method using a wire connector!) without changing anything some ran cool and lastes for nearly 20 hours while others blew instantly in 3sec.

I wasn't pushing the Vds limit with my circuit so I swapped to IRFP260s and all the problems vanished. Perhaps simply moving to a lower RdsON fet like I did will help as well. Obviously don't use IRFP260 in your circuit as it's Vds is only 200V, but an analog of the 460 with lower RdsON instead.

The only other suggestion would be to use MUR860 ultrafast diodes and heatsink them as well.

Kizmo

Wed Jun 06 2012, 11:55AM

Registered Member #599 Joined: Thu Mar 22 2007, 07:40PM
Location: Northern Finland, Rovaniemi
Posts: 624

Is that E style core for gdt? Also your gate wiring is far from low inductance. And power leads are running along with gate drive wires. That core type + wiring (=lots of leakage inductance and not very EMI proof) can be cause of your problem in form of bad gate drive waveforms. Have you scoped your gates to see how things look?

Full bridge of properly driven IRFP460s can do this:

Neet Studio

Wed Jun 06 2012, 01:48PM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

From a signal integrity point of view, you could at least twist the gate driving signal and its return wire to the MOSFET. They should be treated as high speed and high current signals. You really should minimize the loop area between them to minimize the inductance and magnetic noise pickup.

Edit:
I have also noticed that 2 of your MOSFET do not have insulators on them. Their metal tabs (drains) on the back could be shorting together via the heatsink? Given that you have diodes and what not in series with the drain, that's probably not by design.

The foil pattern of MOSFET at the left lower corner is not symmetrical with the one on the upper right. Is that narrow strip of copper supposed to be there?

brandon3055

Thu Jun 07 2012, 01:31AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

Thank Sigurthr I went for the 460's because they were the best I could afford and I've heard they are pretty good but if I get sum extra cash some time I might try something different because I've lost count of how many I've killed

Hi Kizmo the gdt core is a solid e core from the input filter of a tv I didn't expect it to work myself I just thought I would see what would happen (it termed out to be the best gbt I've ever built) but after hearing what you and Neet studio said I will definitely do something about the wires to the mosfets also I don't have a scope but I was able to take the driver and bridge to my college and use thairs (without HV input) I will post the pic below please tell me what you think I thought it looked ok

Neet studio I do have insulaters under all my mosfets that pic was taken wen I ran out of to-3p and had to try to-220 insulaters which didn't work out to well (BANG)

I have fixed that problem and I can't see what your talking about with the foil pattern

Edit I used this diver

Neet Studio

Thu Jun 07 2012, 01:46AM

Registered Member #4037 Joined: Fri Jul 29 2011, 03:13PM
Location:
Posts: 86

Oh wait. That yellow stuff there is insulating tape not the FR4 core? I could have sworn I have some FR4 board with yellow just like that.

Bad jpeg artifacts.

Sigurthr

Thu Jun 07 2012, 02:00AM

Registered Member #4463 Joined: Wed Apr 18 2012, 08:08AM
Location: MI's Upper Peninsula
Posts: 597

My novice eyes notice some mild inductive overshoot on the peak of each positive half cycle, but no heavy ringing. Rise and fall slopes are a little shallow but again nothing my novice eyes would say is a dead ringer for an issue. I can't remember what the gate capacitance was for the 460, but I don't think it was a lot. Both signals appear to be 180deg out of phase, as is proper, but one trace is showing a larger Vp-p than the other. It shouldn't be an issue as long as Vpk (not counting the overshoot) on both traces is >10V.

Did you get any measurements when you took it in to be scoped? Make sure Vpk doesn't exceed Vgs in the datasheet (Vpk = Vp-p / 2), include the overshoot as part of the Vpk measurement here.

Btw, I completely understand about using 460s. They're great fets as Kizmo awesomely pointed out, I just have had nothing but bad luck with them. 1/3 of an Ohm ON resistance is just too lossy for my preferences though. I've cooked so many of these fets that the 260s felt like a breath of fresh air!

brandon3055

Thu Jun 07 2012, 02:44AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

This is what I used for the gdt core and i used that wire because i have seen countless others use it and i think the voltage is over 10v also if I remember correctly I only needed 12 turns on the core that's good isn't it? I will make the modifications to the gdt later today and get bake to you

BTW this is the driver board

Edit what are your thoughts on the diodes I know I need to replace the sf37's with something better but i am also using 1n5822 for my schottky's (the same as richy's) do you think they might need to be bigger as well?

Edit: I was just about to make those modifications when I noticed this

I must not have reconnected them when i worked on it a month ago maby there is a problem with them that didn't show up wen i tested them the other two are still connected would I be better off without them?

brandon3055

Thu Jun 07 2012, 05:08AM

Registered Member #4548 Joined: Mon Apr 23 2012, 03:52AM
Location: tasmania
Posts: 271

What do you think I'm not shore whether or not the foil tape on the gdt is a good idea or not I want to know what you think before I test it and i still want to know about the zener diodes should I replace them or try it without them

BTW the blue cord from the driver is just two strands of the same wire as the gdt twisted together wrapped in foil tape and put back inside its original cat-5 cover and its grounded at the driver end

Steve Conner

Thu Jun 07 2012, 08:19AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

The rise and fall times of the gate drive waveforms look awfully slow.

What transistors did you use in Richie's circuit? Did you find the ZTX parts that he recommends?

Also, is your GDT wound multifilar? (Simple explanation: twist all the wires into a bundle, then wind the bundle on the core)

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob