Self Destroying Mosfet with PIC

[Josuetas]

Just that, i am using a IRF530 with a circuit as boost converter. Seems to work fine But
from time to time i receive complains and the MOSFET is destroyed.

How do you interface MOSFET to PIC to avoid destruction?

I am posting my circuit, in case there is any flaw on it.

Thanks FOr all your POSTs


<img src="http://www.picbasic.co.uk/forum/attachment.php?attachmentid=2114;stc=1&amp;d=11938 56902" width="400" height="350" />

[BrianT]

A couple of thoughts.

a. Try a small series resistor in the gate drive. I have read this somehow helps with the large input capacitance (1300 pF). I guess it reduces gate current.

b. If the gate drive is held on too long, the inductor will saturate and the drain current will be enormous. Max allowable is 16 amps continuous or 64 amps pulsed.

c. If the PIC driving the gate has other tasks to do, can you be certain that you are not leaving the gate high long enough to saturate the inductor?

HTH

Brian

[Josuetas]

i will try with the resistor at gate.

I am using HPWM that is supposed to be on all the time the device is on... it is like the first part of my code.

Appart from that i never use HPWM off or something like that, but it is true that a cide BUG could be guilt for this behaviour..... As i said i am usin HPWM switching at 4000 with dutty 20.

Would there be any other idea? Would a NPN transistor like tip31 work better?

thanks

DJC

[mackrackit]

I will echo what Brian said with the added thought about amps.

You will need a good heat sink and if the amps are near tolerance run two or more MOSFETS in parallel.

Another high current/voltage solution would be an IGBT.

[Acetronics2]

Hi,

Would certainly be interesting to scope the Mosfet Drain voltage ... I would not be surprised it goes much over the 100 v rating of the '530.

1mH is a BIG inductor...

You did not tell about your diode ... very fast recovery, I hope !

a series resistor in the gate will lengthen the switching time, and may be hide the problem.
but also increase losses ...

some simulation could help to choose the best components match ...


May be also too close OFF/ON "togglings" of your device ( mains failure i.e. ) can generate some Mosfets destroyed ... my computer inverter supply did it to me once my finger slipped on the ON switch ...

Alain

[Ioannis]

I will double what Alain stated about MOSFET protection from the coil. Use a diode anti-parallel to the D-S of the MOSFET. If you have a scope, check the spikes on th D pin, but anyway, put that fast diode there.

Ioannis

[PHPHOLLY]

Hi Josuetas
When I use MOSFET'S in 5 Volt Design I tend to use Logic Level types (i.e. IRL530) to ensure clean switching. As previously mentioned a series gate resistor of about 100R would probably help.

Pete

[Josuetas]

Hi, thanks to all for your concern.

First of all my drawing is wrong, the coil is a small 101k choke(100uH) i found it, tried it and it worked.

<img src="http://www.picbasic.co.uk/forum/attachment.php?attachmentid=2119&stc=1&d=119401832 1" />
The condensor it's also wrong (why didn't i check this?) i am using 1000uF 50V, and a 40v zener diode to clamp output voltage at condensor is not in the drawing.


This application is not a high voltage one, i am just getting like 80mA 40V output.

The Diode do is a fast switching diode, its even one of those big 6A ones.

I will double what Alain stated about MOSFET protection from the coil. Use a diode anti-parallel to the D-S of the MOSFET. If you have a scope, check the spikes on th D pin, but anyway, put that fast diode there.

Ioannis

Question: The diode proposed is from Drain to 12V voltage source? (Snubber diode). Wouldn't this avoid the condensor from charging? i though the idea of the boost converter was to use the current when the switcg is off to charge the condensor, wouldn't this kind of diode ruin that?.
Or
If the diode is from Drain to Source, could you draw it for me? i've never seen this before and.... i just don't get it... can you explain it?

I already added the resistor to gate.

Thanks for all your suggestions, they are all been taken in count.

[ra68gi]

Hi Jos,
The MOSFET is a voltage driven device. Make sure that the device is fully enhanced by giving a gate voltage greater than 10volts, which will greatly increase the charge in the inversion layer( or conduction layer ) and reduce Rds(drain to source resistance). Since its a voltage controlled device people generally over look at the drive requirement.
Mind you that a weak drive or any input signal noise voltage greater than 20volts can puncture your gate oxide layer permanently destroying your device.
To keep the input impedance low try to connect 470E resistor from gate to ground. Make sure your drive can both source and sink 100ma +. Use a totem pole or npn-pnp combination as the drive. Use a free wheeling diode across the mosfet if it does not have a bultin reverse diode. you can also play with the Dc resistance value of the inductor such that under saturation the current does not exceed the device rating .
Regards
Raghunathan.

[GeoJoe]

I've built many boost supply circuits. Here are the main areas of failure I have found:
P1. Inadequate gate voltage for non logic level FET.
S1. Use a fet driver. This might seem like overkill but it will solve other probles too.
P2. Inadeqate current available on pic pin driving gate. This could be caused by your PIC power supply regulator. It might only happen when at some time you are overloading the PIC power supply by turning on some LED's or somthing that uses a lot of power from this source.
S2. Again a fet drive chip will solve this problem. With a FET driver chip you can use your 12V power source which is likely to be a lower imedance source to drive the gate.
P3. As before mentioned you may be generating voltages from Drain to Source that exceed ratings.
S3. High speed snubber diodes can help in low current cases like yours, however I've found RC snubbers in parrallel with diode snubbers to be a superior solution. You can google rc snubbers for many solutions.
P4. Sudden load removal can be devastating to a boost circuit.
S4. You can add some type of fail safe feed back. The easiest way to do this is a comparator circuit that pulls the HPWM line low when the boost output voltage exceeds a set limit.
Notes:
a low value resistor (10 ohms or so) between a FET driver chip and the gate is needed for large gate capacitance FETs.
An SOIC for boost circuits can be used for this circuit too.
Microchip has some easy to use fet drivers and even a handy spread sheet calculator.