Help with Mosfet data sheet.

[chapmjw]

Can one of you EE's please help an ME understand the details of a
mosfet data sheet. The values I've calculated are not even close to
real world results. My understanding of the data sheet is limited to
what I have read on this and other news groups.

Here's what I've done thus far.
To test my calculated values to a real world example I did the
following.

24V dc precision power supply
1 mosfet STP1920-NB (TO-220 case)
1 Heating Element reads 13.xx ohms

With this connected and powered up, I get a current reading of 1.540
amps.

From the data sheet I get the following info.

RDS(on) .15 ohm
Tja 62.5 C/W

From search results online I used the following to predict the rise in

temp on the case.
Ambient Temp 27.8 C
Power dissipation (1.54 * 1.54) * .15 = .35574 W
Rise in temp = Tja * PowerDissipation
62.5 * .35574 = 22.23 C
Case temp = ambient + RiseInTemp
27.8 + 22.23 = 50.01 C
In about 10 seconds time the actual value I'm reading on the case is
almost triple that at 148 C .

Data sheet for the mosfet is here:
http://rocky.digikey.com/WebLib/ST Micro/Web Data/
STB19NB20-1,%20STP19NB20_FP.pdf
Any help is greatly appreciated
Jim

 

[Fred Bartoli]

chapmjw a écrit :

Can one of you EE's please help an ME understand the details of a
mosfet data sheet. The values I've calculated are not even close to
real world results. My understanding of the data sheet is limited to
what I have read on this and other news groups.

Here's what I've done thus far.
To test my calculated values to a real world example I did the
following.

24V dc precision power supply
1 mosfet STP1920-NB (TO-220 case)
1 Heating Element reads 13.xx ohms

With this connected and powered up, I get a current reading of 1.540
amps.
RDS(on) .15 ohm
Tja 62.5 C/W

temp on the case.
Ambient Temp 27.8 C
Power dissipation (1.54 * 1.54) * .15 = .35574 W
Rise in temp = Tja * PowerDissipation
62.5 * .35574 = 22.23 C
Case temp = ambient + RiseInTemp
27.8 + 22.23 = 50.01 C
In about 10 seconds time the actual value I'm reading on the case is
almost triple that at 148 C .

Data sheet for the mosfet is here:
http://rocky.digikey.com/WebLib/ST Micro/Web Data/
STB19NB20-1,%20STP19NB20_FP.pdf
Any help is greatly appreciated

Several points:
1) 24V/13.xx = 1.84A and you measure 1.54A This seems to indicate your
mosfet isn't saturated. How much gate drive do you have? You can check
the source-drain voltage which should be under half a volt.
2) Apart this your calc is almost OK, except the you forgot to correct
RDSon for temperature (see curve GC74110 page 5).

 

[linnix]

Can one of you EE's please help an ME understand the details of a
mosfet data sheet. The values I've calculated are not even close to
real world results. My understanding of the data sheet is limited to
what I have read on this and other news groups.

That's why you always need to test the theory.

Here's what I've done thus far.
To test my calculated values to a real world example I did the
following.

24V dc precision power supply
1 mosfet STP1920-NB (TO-220 case)
1 Heating Element reads 13.xx ohms

With this connected and powered up, I get a current reading of 1.540
amps.>From the data sheet I get the following info.

If your supply is stable at 24V, the resistance across the MOSFET is 2
ohm, not .15 ohm. R(mosfet) = V / I - R(ele)

RDS(on) .15 ohm
Tja 62.5 C/W


temp on the case.
Ambient Temp 27.8 C
Power dissipation (1.54 * 1.54) * .15 = .35574 W
Rise in temp = Tja * PowerDissipation
62.5 * .35574 = 22.23 C

With 2 ohms, the junction temperature would be close to 300 C.

Case temp = ambient + RiseInTemp
27.8 + 22.23 = 50.01 C
In about 10 seconds time the actual value I'm reading on the case is
almost triple that at 148 C .

Sound right, with the temperature gradient.

Data sheet for the mosfet is here:http://rocky.digikey.com/WebLib/ST Micro/Web Data/
STB19NB20-1,%20STP19NB20_FP.pdf
Any help is greatly appreciated
Jim

You can't just drive this with a constantly on MOSFET.

 

[Ban]

Can one of you EE's please help an ME understand the details of a
mosfet data sheet. The values I've calculated are not even close to
real world results. My understanding of the data sheet is limited to
what I have read on this and other news groups.

Here's what I've done thus far.
To test my calculated values to a real world example I did the
following.

24V dc precision power supply
1 mosfet STP1920-NB (TO-220 case)
1 Heating Element reads 13.xx ohms

With this connected and powered up, I get a current reading of 1.540
amps.
RDS(on) .15 ohm
Tja 62.5 C/W

temp on the case.
Ambient Temp 27.8 C
Power dissipation (1.54 * 1.54) * .15 = .35574 W
Rise in temp = Tja * PowerDissipation
62.5 * .35574 = 22.23 C
Case temp = ambient + RiseInTemp
27.8 + 22.23 = 50.01 C
In about 10 seconds time the actual value I'm reading on the case is
almost triple that at 148 C .

Data sheet for the mosfet is here:
http://rocky.digikey.com/WebLib/ST Micro/Web Data/
STB19NB20-1,%20STP19NB20_FP.pdf
Any help is greatly appreciated
Jim

The real world is more like this:
24V supply
-0.15V meter loss
-(14R*1.54A) Heater has pos tempco
= 2.29V across FET *1.54A = 3.5W
Why does your fet have almost 1.5R ?
Your gate voltage is too low, or you have made up a wrong circuit. Your fet
must have the Source on gnd, and the load is between +24V and Drain.

 

[Joerg]

That's why you always need to test the theory.




If your supply is stable at 24V, the resistance across the MOSFET is 2
ohm, not .15 ohm. R(mosfet) = V / I - R(ele)

I would agree. Even if 13.xx would be 13.99 ohms that should result in
about 1.7A, provided the 24V are really precise. Either the FET
gate-source voltage isn't high enough or maybe gate and source are
reversed and Jim is accidentally running it via its body diode, which is
slowly frying the thing up.

With 2 ohms, the junction temperature would be close to 300 C.

Without a heat sink it's probably just a few seconds until smoke signals
will emerge and some bits and pieces fly off ;-)

Sound right, with the temperature gradient.




You can't just drive this with a constantly on MOSFET.

Why not? A FET with an Rdson of 0.15ohms looks alright here. If it's fed
the proper gate voltage, that is.