K3067 busted - recommended replacement?

[trodas]

( back photo - http://s232.photobucket.com/albums/e...067busted2.jpg )

So, recently I turned one of my folding machines - a "webserver" one, using Abit KD7-G mobo and powered by Enermax 431W PSU ( http://www.pcstats.com/articleview.cfm?articleID=409 ) and mid loud explosion happen. Machine did not power on anymore.

The PSU is recapped with quality caps, but I'm affraid that with the Panny TS-ED caps in the input, their ripple is maybe too high and ESR too low, so maybe that is what caused the mosfet pictures above to end how it end. After cleaning I find out that also the 2M ohm resistor to the base of this mosfet is damaged (part of it's leg to the mosfet is gone) and 10A 250V fuse is busted badly - almost like the wire exploded and wallpaper itself on the glass of the fuse

Nevermind. It looks as Fairchild component, very similar (likely) to this Toshiba mosfet: http://www.alldatasheet.com/datashee...A/2SD3067.html

Basicaly it is 600V mosfet, 2A current (1mS pulse 5A, 100uS pulse 8A) with RDS(ON) 4.5 ohm and plastic core. Can't find anything on http://www.fairchildsemi.com

So, anyone got idea about available higher grade (mainly lower RDS(ON) and higher current, at least 4A ... ) replacement, guys?

[Gianni]

Re: K3067 busted - recommended replacement?

Hi trodas

read this datasheet:

https://cdn.badcaps-static.com/pdfs/...80e2fd6a5a.pdf

https://cdn.badcaps-static.com/pdfs/...8d779406a5.pdf

https://cdn.badcaps-static.com/pdfs/...e379eb3d0f.pdf

I think one of them can replace the Toshiba.

If the Power Mos is dead, check carefully the control circuit and the drive because usually not only the MOSFET died.

I'm in a hurry now, hope to be back again tonight.

Ciao
Gianni

[trodas]

Re: K3067 busted - recommended replacement?

Hi, thanks a bunch! All I managed to quickly find was a NTE2387 at Mouser, but they asked $13 for it, so... a bit crazy, and it was not TO-220FP - eg. not the isolated one and I cannot have like 250V on the heatsink, that is no-go

K3067 - 600W, 4.5 Ohm RDS(ON), 2A (TO-220FP)
STP3NK60ZFP - 600V, 3.5 Ohm RDS(ON), 2.4A (TO-220FP)
NTE2387 - 800V, 2.7 Ohm RDS(ON), 4A
STP4NK60ZFP - 600V, 2.0 Ohm RDS(ON), 4.0A (TO-220FP)
STP5NK60ZFP - 650V, 1.6 Ohm RDS(ON), 5.0A (TO-220FP)

The best choice is clearly the STP5NK60ZFP, but no stock in Digi or Mouser, but just a slightly worser STP4NK60ZFP is available: http://search.digikey.com/scripts/Dk...=497-5980-5-ND

So I think I will go with it.

You are right about the "mosfet did not failing out of the blue", but I was unable yet to find anything besides the gate resistor (2 Megaohms!) damaged (one of it's leg near the mosfet was missing, burned/blasted away) and the fuse blown. And the mosfet cracked.

I beginning to slightly suspect the mosfet failed. The resistor is up to date still 2M (okay, like 1.96 my DMM say) and nothing else is visibly damaged, so, we see when I obtain thru friend or somehow the mosfet.

I probably give it another inspection, just in case I missed something.

I suspect the mosfet failed on the Power On, because I replaced the input caps in the PSU with hi-grade Panny TS-ED caps and they have like twice the ripple the original crap caps had... so maybe there come the source of failure?

[kc8adu]

Re: K3067 busted - recommended replacement?

no
less ripple=better.
mosfet could ave been wounded by a spike then finally got around to dieing.
or pwm chip could have locked it on.bang!

[Gianni]

Re: K3067 busted - recommended replacement?

and they have like twice the ripple the original crap caps had... so maybe there come the source of failure?

If you mean that you have measured a ripple that is 2 times greater than before changing the caps it is possible but difficult to say for sure.
If the ripple was lower than kc8adu is right.

At minimum input voltage, the duty cycle at full load is at maximum value so it is possible that having a low voltage under the min. value, due to the high ripple, made the PSU working out of spec and the controller has failed: maybe the Power Mosfet Ton was too long and the transformer has saturated so the peak current has destroyed the Power Mosfet.
You have the maximum duty cycle also during start-up because the PSU must charge the output caps and reach the regulated output voltage.

Anyway the power on is a critical phase for any SMPS and the failure can happen if the circuit is not well engineered.
I don't know what kind of IC controller you have on the PSU, but I suggest you to find the datasheet to proceed in some test.
With PWM controller like UC3842/43/44/45 or SG3524/25 etc... you can check if the driving circuit works properly: you can feed the controller with a regulated voltage, it must be between 12 - 16Vdc, and the with a scope you can check the waveform like the sawtooth, driver output etc...
Be sure to feed the PWM IC with a voltage higher than the minimum required for start up, for UC384X is between 11V and 16V IIRC and depend on the last number; the PWM controller has a minimum operating voltage also, if you decrease the voltage usually around 8 - 10V they stop until the voltage rises to the upper threshold.

At this point, if the driving circuit is working and you can see the square waveform between Gate and Source on the Power Mosfet, you can feed the PSU primary side with another power supply, better if it is possible to regulate the voltage starting from 0V; you can slowly increase the input voltage and check if the output voltages of the PSU appear. Probably at no load is it possible that you can see the right voltage value with an input voltage of 50-70Vdc.
On the other hand, if nothing appear and the current on the second PS increase quickly this means that there can be a short circuit on the primary or secondary side. To find where the problem is, you can disconnect the secondary side of the transformer e check again.
All the power supply you use to feed the controller and primary side of PSU must have a current limiter so you can avoid some problem: burning tracks, components, cables...

Pay attention also where you connect the power supply to feed the primary side ot PSU: if it has an active PFC, you must connect the power supply to the output cap of PFC; if the PSU has a standard bridge rectifier, you can connect the power supply to the input power cable.

Mosfet gate resistor: you wrote it is a 2M but it is strange, usually the resistor around the Mosfet are in the range of 4.7R up to 22K. You should find a resistor connected in series with the gate, usually low value (4.7R up to 100R), and it is possible that there is also a resistor between gate and source (1k up to 22K).
When you replace a Power Mosfet with another type, it is not important only the RDSon, Max Voltage a Max Power but also the other characteristic like Input capacitance, Gate charge etc...
In fact, even if 2 power mosfet have the same ratings, RDSon....., you can have more spikes, more power dissipation due to different behavior: one can turn on very fast respect the other one generating higher DV/DT that can cause spikes and increase snubber/clamp power dissipation. It can also overload the driving circuit if its input impedance is low.
In your case I don't know if you can see any difference using ST instead of Toshiba/Fairchild, but if I were you I would try it.

I have never repaired an ATX PSU (only recapped ) so I don't know how they are generally designed, I have experience with flyback and step-down converter, so I can't give you more details.

I don't know if I have wrote clearly and correctly, my English is not so good, anyway I hope it help you.

Gianni

[Wizard]

Re: K3067 busted - recommended replacement?

I had a SMD 8 pin mosfet (N and P gate mosfets on one IC) burn out on a notebook due to bad controller IC. Luckily have correct IC from junked notebook and a compatible 8 pin mosfet IC got it back in business.

Cheers, Wizard