A little help with mosfet specs...

Thread Starter

geekoftheweek

Joined Oct 6, 2013
1,201
I'm having a problem finding my answer... What exactly does the power dissipation refer to in fet type specifications? Does it refer to the amount of power that can pass through source and drain, or is it somehow related to the gate? I hate to ask such a simple question... after a couple hours of searching I thought I would ask you all. Thanks
 

BobTPH

Joined Jun 5, 2013
8,813
Power dissipation in a MOSFET is the Vds * Id

Where Vds is the voltage across the drain to source and Id is the drain current.

When fully on, the Vds will be low, it will be Id * Rdson.

When fully off the current will be 0 so power is 0.

You can typically switch a load much higher than the power dissapation within the MOSFET. Just because the load is 100W does not mean the MOSFET is dissipating 100W.

Bob
 

Thread Starter

geekoftheweek

Joined Oct 6, 2013
1,201
Thanks Bob... I'm sure I just skimmed over that somewhere while trying to find an answer. I'm looking to use some p channel to switch on and off a few circuits of 5 volts around a half amp or give or take and wanted to make sure I'm going to order the right parts. I can't use n type being the circuits involve some pics and I've noticed that just doesn't work with them. It's a matter of over a dollar per piece if I don't understand what I'm doing.
 

Papabravo

Joined Feb 24, 2006
21,159
In switching applications the actual power consumed by the device is small. What you need to watch for is slow switching waveforms. As the gate makes its transition through the linear range between conduction and pinchoff there can be large peak currents that will cause a temperature rise. A high Vgs(th) and a large gate capacitance are usually impediments to fast switching waveforms.
 

Thread Starter

geekoftheweek

Joined Oct 6, 2013
1,201
I looked around a bit and thought maybe this would work... http://www.onsemi.com/pub_link/Collateral/NTS2101P-D.PDF

I tried to locate something with a lower gate capacitance as Papabravo mentioned, and with at least twice the current capacity I was looking to handle (just to eliminate problems with heat and such I don't want to deal with).

A little background on my project. I am looking to power three different circuits at the moment. They will all be powered from a converted cigarette lighter usb charger that puts out 5.1 volts (measured) and capable of delivering 2 amps according to the markings. The main circuit will involve a pic18f2525 that will turn on and off the other two circuits based on inputs. One of the other circuits will have a pic18f2550 and a 16x2 lcd display run by a pic16f886. I've estimated current consumption on that circuit to be around 500 mA (that's probably on the high side, but I'm trying to leave plenty of wiggle room). The other circuit will power various sensors and such that estimated total power consuption will be less that 500 mA (I think as it stands right now it's more like 150 - 200 mA). I plan to use pull up resistors in the area of 100k on the gates and use the pic to pull the gate low and turn the mosfets on or off depending on the needs at the time.
 

Papabravo

Joined Feb 24, 2006
21,159
A three processor project run from a cigarette lighter usb charger seems pretty ambitious for such a setup. Will this be a one off breadboard or are you going into production with a PC board(s)? The reason I ask is because layout will have an effect on the MOSFET behavior and characteristics. A schematic would convey far more useful information than a short paragraph.

With respect to the datasheet, look at the box on page 1 labeled "THERMAL RESISTANCE RATINGS" and at the parameter called R
θJA in °C/Watt. See the number 430? What does this number mean? Glad you asked, it means if the part is dissipating 1 Watt the difference between the ambient temperature and the Junction (Channel) will be 430°C!!. Of course the magic smoke will be let out long before that. See (Note 1) it tells you that even for this situation the part needs to be sitting on 1 square inch of copper.

How much power will you dissipate at 1.4A through 65 mΩ? About 127 mWatts which corresponds to a 55 °C temperature rise. That's hot enoungh to burn your lips. Keep it out of the linear range, make fast transitions, and put it on a PC board.
 
Last edited:

Thread Starter

geekoftheweek

Joined Oct 6, 2013
1,201
Thanks papa... This stuff is making my mind numb!! I'm planning on etching a board for this, but am starting to get a bit spooked by all the little things I never knew I had to consider. Since it's just a one time thing I may end up going with the overkill option that costs a bit more, but will be way more than adequate for the job. Learning this stuff is fun, but the magic smoke turns my hair gray.
 

Alec_t

Joined Sep 17, 2013
14,280
Can you use low-side switching for your circuits? That could be done with N-MOSFETs, giving you a wider choice of low Rds(on) types and hence lower power dissipation.
 

ian field

Joined Oct 27, 2012
6,536
I'm having a problem finding my answer... What exactly does the power dissipation refer to in fet type specifications? Does it refer to the amount of power that can pass through source and drain, or is it somehow related to the gate? I hate to ask such a simple question... after a couple hours of searching I thought I would ask you all. Thanks
Power dissipated in the device is the product of V & I.

The contribution made by the gate, is the consideration of whether it controls the drain current in a linear fashion (potentially maximum dissipation) - or if the gate is driven for the channel to be either conducting hard, or cut off, there you have 2 states; fully conducting but with very little volt drop (low VxA) - or cut off so all the voltage is across the device, but at hardly any current.
 

Thread Starter

geekoftheweek

Joined Oct 6, 2013
1,201
Alec... Unfortunately low side switching won't work for what I want to do with the circuit. I have however rethought my design that's working now that uses a relay switched by a n-mosfet and decided maybe I should stick with it for now while I research and learn a bit more about mosfets in general. My main goal was to eliminate the mechanical action (and possible point of trouble over time... as in it finally quits working) of the relay.

Ian... I had in mind for the mosfet to be either full on or full off.

Thanks for the help so far. I've found a few other threads in the forums over the last few days that have expanded on what you all have said so far and added a few things that haven't been said. I guess I just needed a starting point to start putting the puzzle together. It's starting to make a bit more sense.

I'm making this a somewhat modular project with the power supply and switching circuits on one board with separate boards for the pic's, outside i/o, and whatever else comes along. I've been working on this well over four months now and every time I think I'm ready to call it done I think of something else to add to the mix. I'm to a point where I want to get the things that won't change finalized and make a board to handle them instead of punched boards with wires connecting everything.
 

ian field

Joined Oct 27, 2012
6,536
Alec... Unfortunately low side switching won't work for what I want to do with the circuit. I have however rethought my design that's working now that uses a relay switched by a n-mosfet and decided maybe I should stick with it for now while I research and learn a bit more about mosfets in general. My main goal was to eliminate the mechanical action (and possible point of trouble over time... as in it finally quits working) of the relay.

Ian... I had in mind for the mosfet to be either full on or full off.

Thanks for the help so far. I've found a few other threads in the forums over the last few days that have expanded on what you all have said so far and added a few things that haven't been said. I guess I just needed a starting point to start putting the puzzle together. It's starting to make a bit more sense.

I'm making this a somewhat modular project with the power supply and switching circuits on one board with separate boards for the pic's, outside i/o, and whatever else comes along. I've been working on this well over four months now and every time I think I'm ready to call it done I think of something else to add to the mix. I'm to a point where I want to get the things that won't change finalized and make a board to handle them instead of punched boards with wires connecting everything.
There are various ways of obtaining gate drive voltage for a high side switch.

In the old days when a PSU was fed by a mains transformer, a charge pump voltage doubler could be fed from one of the AC terminals on the bridge rectifier.

In a typical half bridge MOSFET output stage, you'd run the doubler from the output terminal, but this only works if its continually switching.
 
Top