I've not used MOSFETs in circuits before. They look like ideal for what I want to do - PWM control some RGB LEDs.

The LED strips are rated 12V 1.2 Amp (they come pre-soldered, sets of 3 LEDs in a chain with a resistor, on flexible strip).

I'm using a PIC for the software, which I've got a handle on. I've been using just a basic 20mA LED on each pin with a resistor to test, so that's all working. The PIC runs at 5V and can sink 30mA per pin max.

Now I think I need to link MOSFETs to the PIC so it can drive the 1.2A 12V stuff.

I've found a MOSFET that I *think* will meet spec and is low cost - STP16NF06, N-channel MOSFET Transistor 16A 60V, 3-Pin TO-220
http://uk.rs-online.com/web/p/products/486-2206/

Not really understanding them though, I don't know if my thinking is right.

The PIC outputs 5V, which is my "VGS". If I'm reading the graph right, then if I were using a 15V "VDS", I could switch about 9Amp with a "VGS" of 5V. Since I'm only using 12V it would be less, but well over the 1.2A I need, and it looks like it would be about 3A at 4.5V, so even if the PIC output is a little lower I should be fine.

Spec sheet says max 45W power dissipation... well the whole circuit wont use that much (I calc about 15W total) so again I should be fine.

• The 12V supply goes to the LED + wire.
• The output pin of the PIC goes to the MOSFET G pin, and there should be a 1kohm pull-down resistor to ground.
• The MOSFET D goes to LED - wire.
• The MOSFET S pin is tied to ground.

Have I got this right, or am I about to make things go "pop!" ?

 

To turn this particular NFET on all the way, it requires a Vgs of 10V. Your PIC port only puts out 5V at most. This NFET is barely begining to turn on with only 5V on the gate.

To directly control the gate of a NFET from a 3.3V or 5V logic output from a ucontroller, you should buy a "logic-level" NFET, which is spec'ed to fully turn on with a lower Vgs. Typically, such a FET will have a Vth of ~1V

 

Thank you. What would be a typical Vgs to look for? The searches I've found don't seem to allow for looking for "logic level" or "Vgs(th)" (which I assume is the same as Vth).

 

Also, just so I'm clear... are you saying that there would be a problem if the gate is barely turning on like that, or is it just a case of it would work but another solution would be better (though possibly a lot more expensive).

 

There is a problem; Rdson, the resistance across the MOSFET when it's turned "on". With a high current and too much Rdson, the MOSFET will have to dissipate more heat ( I^2•R) than can be tolerated by the package it is in. A low gate voltage causes Rdson to be more than it is at a higher gate voltage.

 

The search feature of the supplier you're looking at, while lots of searchable parameters are given, doesn't seem to have a filter for logic level vs. standard MOSFET. Digikey does, so I looked there until I found something your vendor carries.

This should work. Look at the first page of the datasheet.

 

The below are some logic level MOSFETs and the V/I enough to suit what you need --

BUK9507-30B : Nch_30V,75A,157W,7mΩ
IRLU8721 : Nch,30V,45A,11.8mΩ
PSMN017-30PL : Nch,30V,32A,17mΩ
IRL540 : Nch,100V,28A,Vgs4V,77mΩ

 

I've not used MOSFETs in circuits before. They look like ideal for what I want to do - PWM control some RGB LEDs.

The LED strips are rated 12V 1.2 Amp (they come pre-soldered, sets of 3 LEDs in a chain with a resistor, on flexible strip).

I'm using a PIC for the software, which I've got a handle on. I've been using just a basic 20mA LED on each pin with a resistor to test, so that's all working. The PIC runs at 5V and can sink 30mA per pin max.

Now I think I need to link MOSFETs to the PIC so it can drive the 1.2A 12V stuff.

I've found a MOSFET that I *think* will meet spec and is low cost - STP16NF06, N-channel MOSFET Transistor 16A 60V, 3-Pin TO-220
http://uk.rs-online.com/web/p/products/486-2206/

Not really understanding them though, I don't know if my thinking is right.

The PIC outputs 5V, which is my "VGS". If I'm reading the graph right, then if I were using a 15V "VDS", I could switch about 9Amp with a "VGS" of 5V. Since I'm only using 12V it would be less, but well over the 1.2A I need, and it looks like it would be about 3A at 4.5V, so even if the PIC output is a little lower I should be fine.

Spec sheet says max 45W power dissipation... well the whole circuit wont use that much (I calc about 15W total) so again I should be fine.

• The 12V supply goes to the LED + wire.
• The output pin of the PIC goes to the MOSFET G pin, and there should be a 1kohm pull-down resistor to ground.
• The MOSFET D goes to LED - wire.
• The MOSFET S pin is tied to ground.

Have I got this right, or am I about to make things go "pop!" ?

Just to confuse you even more - there are 2 types of MOSFET (apart from polarity); enhancement and depletion - the depletion variety work exactly like tubes, that is they conduct heavily with no gate/grid bias - you have to bias it negative to turn the current off.

With an N-channel enhancement MOSFET, no current flows when gate bias is zero, you have to give it a positive bias to make the MOSFET conduct - this is actually a PITA for biasing in linear circuits, most people use them in switching applications, from the tiny logic level devices to the somewhat larger PWM motor control devices and choppers in SMPSU circuits.