P-MOS control and capactive load

[Koxx3]

Hi,

I am trying to drive an high capacitive load with a P-Mosfet.

I added a PTC to limit the current, but it doesn't seen to be working. The mosfet is often killed.
Can anybody help and provide a 'simple' solution to this issue please ?

Here is the schematics.

Thanks !!

 

[danadak]

You are operating the IRFR9120 right at its spec Vds breakdown so there is no margin
there for starters. Plus you have it in the active region a long time due to slow gate
charging so its dissipating a lot of power potentially.

And I am guessing you may be violating SOA :



How rapidly do you have to charge this 500 uF cap ?

Here is a sim driving part at 100 hz :



Regards, Dana.

 

[Koxx3]

thanks a lot. yes, I tend to understand that.
First, I've ordered some stronger MOSFET (HUASHUO HSU0139. datasheet.lcsc.com/lcsc/2209221830_HUASHUO-HSU0139_C701018.pdf ).
pulse drain is at 100A, and continuous at 30A...
but it might not be enought, no mosfet will handle ~450W (according to your simulation ?)

and as I understand, PTC might not be the best option. and NTC IRL might be a better option ?
something like that ?

B57237S0229M000 TDK | C125462 - LCSC Electronics

B57237S0229M000 TDK US$0.4958 - 2.2Ω 7A 3.1W ±20% Straight,P=7.5mm NTC Thermistors ROHS datasheet, price, inventory C125462

lcsc.com

 

[danadak]

Thats peak power not avg power which would be a lot lower
because of the duty cycle is so low. What freq do you drive
the circuit ? Your gate drive R's, 100K, 20k, can make the
mosfet switch faster and lower the avg power if they are
significantly reduced. But that means the peak drain
current will rise.

The 500 uF cap, a regular wet alum electrolytic ?

Can you describe what the circuit is for ?

Your new MOSFET is also a 100V part, how sure are you
that the 100V supply never has > 100V spikes on it ?


Regards, Dana.

 

[Koxx3]

Thats peak power not avg power which would be a lot lower
because of the duty cycle is so low. What freq do you drive
the circuit ? Your gate drive R's, 100K, 20k, can make the
mosfet switch faster and lower the avg power if they are
significantly reduced. But that means the peak drain
current will rise.

The 500 uF cap, a regular wet alum electrolytic ?

Can you describe what the circuit is for ?

Your new MOSFET is also a 100V part, how sure are you
that the 100V supply never has > 100V spikes on it ?


Regards, Dana.

Thats peak power not avg power which would be a lot lower
because of the duty cycle is so low. What freq do you drive
the circuit ?
Can you describe what the circuit is for ?
>> it's a humain activated power latch, so, VERY low frequency. power on/off every 5 minutes max

The 500 uF cap, a regular wet alum electrolytic ?
>> yes, but low ESR. It's not the actual load, but a "simulation" of the worse case senario. Load are DC/DC from 1 to 5A input for lights.

Your new MOSFET is also a 100V part, how sure are you
that the 100V supply never has > 100V spikes on it ?
>> power supply is lithium battery with a max voltage from 10S (36V) to 22S (92V). I don't expect spikes going up to 100V

 

[danadak]

I am not concerned about the power then. But low ESR cap thats
means the fet is looking into almost a dead short with 100V across
it, which must be violating SOA if I read the SOA chart right. What is
its ESR ? And for sure the Litiums are low esr as well.

You could run a quick turn on test with a milliohm shunt in the load
path, and using a DSO capture what peak current and Vds looks like
to see what you have.

You could try a compromise when you turn on fet more slowly to get
the cap ramped up slower, that however aggravates power loss in fet.

Or put a constant current limiter in drain of fet to force a ramp. A 3
terminal adjustable reg with a R will handle this -



You lose some head room in Cap V, ~3 V, depending on its leakage.


Regards, Dana.

 

[danadak]

Actually you cannot use the LM317,. my error, does not have enough
HV capability. Use a PNP version of one of these



or



or






Two diodes below could be a LV zener.


Regards, Dana.

 

[Koxx3]

Actually you cannot use the LM317,. my error, does not have enough
HV capability. Use a PNP version of one of these

View attachment 56399

or

View attachment 56400

or


View attachment 56401



Two diodes below could be a LV zener.


Regards, Dana.

PNP for 100V ? it is safe ?

 

[danadak]

Use a higher Vceo PNP than 100V.....give yourself some margin.

The BD140 was only a web example of the source, just pick a 200V part
with ratings you need.

Regards, Dana.