# Mosfet provokes high voltage drop and become hot

I'm using a STP16NF06L mosfet to drive a pump (12V / 3A). The problem is: there is only 5V at the pump (so 7V voltage drop at the mosfet), and because of that the mosfet becomes instantly hot. I'm driving the mosfet with the raspberry pi through an optocoupler.

simulate this circuit – Schematic created using CircuitLab

I have measured 5V between mosfet gate and ground, I think it should be fully on.

• Your source needs about $4V$ (worst case) between itself and the gate. Your pump probably wants $12V$. Have you considered the idea of putting the IRF530 in the (-) leg of your pump circuit, instead? – jonk Aug 26 '16 at 0:42
• The MOSFET doesn't care where the Gate is relative to Ground - it only knows where the Gate is relative to its Source. – Peter Bennett Aug 26 '16 at 0:44

You are trying to use an N-channel MOSFET for a high-side switch, and you are not turning it on, because you are not driving the gate correctly.

For the MOSFET to turn on, the gate should be a certain voltage (VGS)about MOSFET's source. If the source itself is at some positive voltage, then the VGS has to be on top of that voltage. Imagine that you want to have +12V at the positive terminal of the pump, and you want to drive the gate with VGS=5V. Then the gate voltage with respect to ground would have to be +17V. You don't have a high enough supply rail in your circuit for that.

How to explain the power dissipation which was observed? When you try to turn the MOSFET on, the voltage at the positive terminal of the pump was somewhere between 0V and +5V. There is still enough voltage to turn on the MOSFET partially. As a result, there is still some current flowing though it, and there is a voltage drop across it. It can't turn on fully, and it can't turn off fully.

Options to fix this:

• Use the same N-channel MOSFET as a low-side switch.
• Use a P-channel MOSFET, if you need a high-side switch for some reason.
• Is it the same with bipolar npn transistor? – James Magnus Aug 26 '16 at 1:08
• @JamesMagnus: Yes. NPN's and PNP's are duals of each other the same way NMOS and PMOS are duals of each other. If you use an NPN in the same way, you will have one threshold voltage drop difference between the base voltage and the emitter (~.8V or so for power BJT). Instead use a PNP, in which case you just have one saturation voltage drop (~0.2V) instead. – jbord39 Aug 26 '16 at 1:20

Your problem is that you're using an N-channel MOSFET as a high-side switch. This is not recommended. Provided your pump doesn't need to be connected to the +12V ground, the correct configuration for your STP16NF06L as a low-side switch would be:

simulate this circuit – Schematic created using CircuitLab

• And just the mosfet place in the circuit explains the incredible power dissipation ? Oh, I see: in my circuit the gate voltage is between gate and pump + and not between gate and ground. – James Magnus Aug 26 '16 at 1:01
• Yes. Because it's not being driven correctly. In your configuration, it'll turn into a resistor since the gate is at a fixed voltage (~5V), but the source is trying to rise towards 12V. – user2943160 Aug 26 '16 at 1:05
• Just as a point, an NMOS makes a great switch for more negative voltages (cutting off the current return path, usually at ground or most negative). And a PMOS makes a great switch for more positive voltages (cutting off the current supply path, usually power supply). They are duals of each other, so both make poorer switches in the opposite situations (pmos on low side, or nmos on high side both suffer from one threshold voltage drop). – jbord39 Aug 26 '16 at 1:17
• The part where "so both make poorer switches in the opposite situations" is not entirely correct. It is much easier to drive them when nmos cuts the current return and pmos cuts the supply current, but they can be used in opposite configurations (in which case gate to source voltages are slightly harder to set). However, between 2 mosfets (one pmos and one nmos) manufactured under the same process (same gate length to width ratio) nmos has about 1/3 drain-source resistance, ho the high-end high-efficiency power inverters use nmos-only h-bridges with proper drivers. – nurchi Aug 26 '16 at 1:28