Strange transistor failure
Dr. Dark Current, Fri Oct 13 2006, 04:09PMWell, I was playing with the "Single transistor flyback driver" and was optimising it for efficiency. I was using the original flyback primary and a HOT (1,5kV, 8A). I put two 680V MOV's across C-E of the HOT, and uf4007 diode from emitter to base (anode at emitter) for extra transistor protection, and the HOT has built-in C-E damper diode. Everything worked great, I could get around 30kV with just 10-15watts in, but after a while the HOT died (this happened when the HV was not loaded). Now I can't explain this. The HOT wasn't even warm (no overheating) and was protected by MOV's (no overvoltage).
Any ideas why it died???? I really like the efficiency of this circuit, so does anyone have an idea of how could I protect the HOT from blowing up?
thanks, J.M.
EDIT: possible reason for HOT death - can this be the B-E diode? Maybe there is not enough negative base bias -> reduced maximum C-E voltage -> death from overvoltage??
Re: Strange transistor failure
HV Enthusiast, Fri Oct 13 2006, 07:11PM
Check to see what your voltage is across the base-emitter junction at all modes of operation.
HV Enthusiast, Fri Oct 13 2006, 07:11PM
Check to see what your voltage is across the base-emitter junction at all modes of operation.
Re: Strange transistor failure
Sulaiman, Fri Oct 13 2006, 08:03PM
I don't KNOW what caused the failure, but here are a few ideas;
1) When you operate with no load,
allmost all of the power from the supply goes to heat in the MOV(s).
MOVs WILL die under such circumstances.
Check ('scope) that your MOVs are still clamping C-E voltage spikes.
Personally I never use MOVs for this reson - I use TVS
(I bought 100 1.5KE350CP cheap off eBay)
For my Ignition Coil invertor I'm using 10 x 1.5KE350CP in parallel epoxied to a heatsink.
You could also consider Diode to capacitor//resistor as a clamp - pretty good
The thing to ALWAYS consider is
If the power isn't going into a load, where does it go?
2 Usually HOTs fail from overvoltage or overheating
but depending on the circuit its possible to fuse (open) either the emitter or base bonding wire.
Check for open circuit base or emitter.
(usually C-E is short-circuit)
If you want to continue using MOVs try connecting from Collector to Base
this way the MOV isn't stressed so much
the MOV causes the transistor to clamp its own flyback spike.
Even in this mode of operation a TVS is much more reliable.
Consider a Xenon flash tube from a disposable camera as a clamp
the small ones flashover at about the correct voltage AND give a visual indication
Keep us updated and good luck
Sulaiman, Fri Oct 13 2006, 08:03PM
I don't KNOW what caused the failure, but here are a few ideas;
1) When you operate with no load,
allmost all of the power from the supply goes to heat in the MOV(s).
MOVs WILL die under such circumstances.
Check ('scope) that your MOVs are still clamping C-E voltage spikes.
Personally I never use MOVs for this reson - I use TVS
(I bought 100 1.5KE350CP cheap off eBay)
For my Ignition Coil invertor I'm using 10 x 1.5KE350CP in parallel epoxied to a heatsink.
You could also consider Diode to capacitor//resistor as a clamp - pretty good
The thing to ALWAYS consider is
If the power isn't going into a load, where does it go?
2 Usually HOTs fail from overvoltage or overheating
but depending on the circuit its possible to fuse (open) either the emitter or base bonding wire.
Check for open circuit base or emitter.
(usually C-E is short-circuit)
If you want to continue using MOVs try connecting from Collector to Base
this way the MOV isn't stressed so much
the MOV causes the transistor to clamp its own flyback spike.
Even in this mode of operation a TVS is much more reliable.
Consider a Xenon flash tube from a disposable camera as a clamp
the small ones flashover at about the correct voltage AND give a visual indication
Keep us updated and good luck
Re: Strange transistor failure
Dr. Dark Current, Fri Oct 13 2006, 08:31PM
thanks Sulaiman, the MOVs were not getting even warm, I jus put them there for protecting from occasional spikes that can occur. The HOT died in a way that B-E reads near zero and C-E has about 100 ohms. I will try to remove the base diode and see what happens, but I have only one HOT left so when it dies I must wait until monday and buy some more (not too cheap). I will also buy some lower voltage MOV's and see if they get warm.
(btw, my soldering iron has developed a bad power cord, and obviously I can't fix this with my iron
so experimenting is suspended until I borrow a soldering iron to fix my own)
Dr. Dark Current, Fri Oct 13 2006, 08:31PM
thanks Sulaiman, the MOVs were not getting even warm, I jus put them there for protecting from occasional spikes that can occur. The HOT died in a way that B-E reads near zero and C-E has about 100 ohms. I will try to remove the base diode and see what happens, but I have only one HOT left so when it dies I must wait until monday and buy some more (not too cheap). I will also buy some lower voltage MOV's and see if they get warm.
(btw, my soldering iron has developed a bad power cord, and obviously I can't fix this with my iron
so experimenting is suspended until I borrow a soldering iron to fix my own)Re: Strange transistor failure
Steve Conner, Fri Oct 13 2006, 08:33PM
MOVs have quite a "soft knee". The voltage across them can be pushed quite far above the clamping voltage printed on them, if you supply enough current. The TVS that Sulaiman recommends are a good deal more accurate.
As far as I know, MOVs never fail open circuit, unless they are hit with enough energy that they blow completely apart.
I believe it's better to use a suitably sized snubber capacitor across the HOT, as they do in TVs anyway. This limits the maximum possible voltage according to conservation of energy (1/2*L*I^2 = 1/2*C*V^2)
Steve Conner, Fri Oct 13 2006, 08:33PM
MOVs have quite a "soft knee". The voltage across them can be pushed quite far above the clamping voltage printed on them, if you supply enough current. The TVS that Sulaiman recommends are a good deal more accurate.
As far as I know, MOVs never fail open circuit, unless they are hit with enough energy that they blow completely apart.
I believe it's better to use a suitably sized snubber capacitor across the HOT, as they do in TVs anyway. This limits the maximum possible voltage according to conservation of energy (1/2*L*I^2 = 1/2*C*V^2)
Re: Strange transistor failure
Dr. Dark Current, Sun Oct 15 2006, 03:56PM
Yeah, I dug out my old unkillable flyback driver with two irfp250's in parallel, and voila, instant 50kV! Why I didn't thought of that before?! Oh well, I'll scrap the STFD. (now I must just watch to not kill the flyback
)
J.M.
Dr. Dark Current, Sun Oct 15 2006, 03:56PM
Yeah, I dug out my old unkillable flyback driver with two irfp250's in parallel, and voila, instant 50kV! Why I didn't thought of that before?! Oh well, I'll scrap the STFD. (now I must just watch to not kill the flyback
)J.M.
Print this page