Choosing a MOSFET

Thread Starter

carto

Joined Feb 3, 2014
45
Hi, I am hoping someone can help me choose a MOSFET for my application. I am using an ATMEGA328 MCU and an N-channel logic-level MOSFET to switch on and off a 12V power supply to a device that uses 25-300mA (the current varies over time). I need to keep power consumption of the system as low as possible. The MCU sleeps most of the time, switching on the 12V power for two minutes every hour (the timing is done with an IC2 clock chip alarm) for five months. The power to the MCU (and therefore the MOSFET's gate voltage) will start at 4.8V and may go as low as about 2.7V by the end.

Currently I am using an IRL540 MOSFET. It works, but I don't know if it's the best for when the gate voltage decreases. I'd like to find the best MOSFET for my application. I keep thinking that I'll get the hang of the datasheet, but so far I am not getting very far. I am stuck at very basic questions, like "Does Gate-Source Voltage mean the voltage coming from the MCU, or is it that voltage minus the source voltage?" This question by itself probably shows my level of expertise. I tried searching for "MOSFET for dummies" but that didn't help much.

I am thinking that before I invest more time with the datasheets, perhaps someone could look a this specific application and help me get started on the right path? I need to use a through-hole package if I can.

Thank you!
 

wayneh

Joined Sep 9, 2010
17,496
Currently I am using an IRL540 MOSFET.
Good news, bad news. The good news is that's a perfectly good choice. Bad news, it may be hard to do better. I'll look a little and come back if I see anything. It's good that your switched current is not high. You can tolerate some Rds-on without huge concern about heat dissipation. That means the gate voltage can be lower and still be OK.
Does Gate-Source Voltage mean the voltage coming from the MCU, or is it that voltage minus the source voltage?
It's the delta between the gate and source of the MOSFET that controls the conductivity of the MOSFET. Typically the MCU and the source pin would share a common ground and that means the MCU output voltage is the gate voltage.

Do you have 12V always available?
 

Thread Starter

carto

Joined Feb 3, 2014
45
Thanks for the reply. If the IRL450 will work at 2.7V it's fine with me to use that one. I just couldn't figure out what is the minimum MCU voltage that will work. The MCU and the -12V source share a common ground, so it sounds like this means that the gate voltage is simply the MCU voltage. The 12V is always available.
 

Thread Starter

carto

Joined Feb 3, 2014
45
I could power the MCU from the 12V battery, but I have been thinking that this would require a voltage converter, which would use power all the time. The design I have now uses AAA batteries to power the MCU. These last for >5 months, and that works fine for my purpose. I need the system to run unattended for 5 months, and everything needs to be small and light.
 

wayneh

Joined Sep 9, 2010
17,496
I could power the MCU from the 12V battery, but I have been thinking that this would require a voltage converter, which would use power all the time. The design I have now uses AAA batteries to power the MCU. These last for >5 months, and that works fine for my purpose. I need the system to run unattended for 5 months, and everything needs to be small and light.
OK, so the challenge is how to switch a 12V, 300mA load using a control signal as low as 2.5V. Have you considered a BJT instead of a MOSFET? The base current would draw a little current from the MCU. Another approach would be to use a BJT to control the voltage on the MOSFET gate, using the 12V supply. You’d need almost zero base current to change the voltage on gate.
 

Thread Starter

carto

Joined Feb 3, 2014
45
I can look at these. I was thinking that the challenge for me was to figure out how to read the datasheets so I could choose the right MOSFET. I thought that the voltage-controlled MOSFET would yield the lowest power use. But it sounds like a MOSFET can't be used down to 2.5V, or at least not with a 12V or 300mA load?
 

wayneh

Joined Sep 9, 2010
17,496
I can look at these. I was thinking that the challenge for me was to figure out how to read the datasheets so I could choose the right MOSFET. I thought that the voltage-controlled MOSFET would yield the lowest power use. But it sounds like a MOSFET can't be used down to 2.5V, or at least not with a 12V or 300mA load?
Oh I think it might work, and careful selection of the MOSFET might get you to the low voltage you need. I’m just thinking that adding a transistor ensues robustness.
 

Thread Starter

carto

Joined Feb 3, 2014
45
I'm all for robustness if it won't use more power! Reliability is extremely important for this project. On the other hand, I already have a pcb made, ready to plug in a MOSFET.
 

wayneh

Joined Sep 9, 2010
17,496
I'm all for robustness if it won't use more power! Reliability is extremely important for this project. On the other hand, I already have a pcb made, ready to plug in a MOSFET.
It would not be too hard to experiment and collect data. Power your MCU and circuit with various voltages and see what happens. If the MOSFET is the weak link, then looking for an alternative will make sense. But it may work fine until the MCU fails from low voltage.
 

Thread Starter

carto

Joined Feb 3, 2014
45
I can experiment, but I was thinking that's what the datasheet is for--to avoid that. I just don't know how to read it.
I am not concerned with the MCU--I know that it can operate down to 2.7V because it says that in plain English on the datasheet.
 

wayneh

Joined Sep 9, 2010
17,496
I can experiment, but I was thinking that's what the datasheet is for--to avoid that. I just don't know how to read it.
I am not concerned with the MCU--I know that it can operate down to 2.7V because it says that in plain English on the datasheet.
The datasheet I found has a few things that are relevant.
Screen Shot 2021-04-09 at 1.47.02 PM.png

The table shows that the Rds will be only 63mΩ at 15A (way more than your situation) and Vgs of 4.0V. It also shows the increase in Rds when the Vgs dropped from 5 to 4 was only 10mΩ. It bodes well to still be reasonable at 3V.

This chart reveals even more. Look at the line for Vgs = 2.5V. For your D-S of 12V, the chart shows the MOSFET could pass over 7A. So I think the resistance will be minimal at 300mA.

Screen Shot 2021-04-09 at 1.53.57 PM.png
 

Thread Starter

carto

Joined Feb 3, 2014
45
It helps a lot to know which are the most important parts in the tables and charts. I see now how to read these.

Possibly to find a MOSFET built for lower gate voltage I'd need to switch to surface-mount. But the IRL540 seems OK for this application. I'll test it with weak batteries in the fridge. Should work though.

Thanks very much for the help!
 

wayneh

Joined Sep 9, 2010
17,496
It helps a lot to know which are the most important parts in the tables and charts. I see now how to read these.

Possibly to find a MOSFET built for lower gate voltage I'd need to switch to surface-mount. But the IRL540 seems OK for this application. I'll test it with weak batteries in the fridge. Should work though.

Thanks very much for the help!
C'mon back with data once you have some. It's a fairly common question.
 

Audioguru again

Joined Oct 21, 2019
6,671
The datasheet for the IRL540 shows that its max threshold Vgs voltage is 2.0V when it conducts only 0.25mA.
The on-resistance is low when the Vgs is 4V and 5V.
But the graphs show only "typical" ones that you cannot buy because many are worse than typical.
 
Top