• This is my first time playing around with MOSFETs. I wanted to know how much continuous current could be handled by MOSFETs when they are controlled by Arduino IO. Also, I would like to know if heatsinks are necessary. The datasheets gave some indication, but I could not find the graphs I was looking for, so I made my own.

Materials

The following MOSFETs were compared:

• - FQP30N06L
• - IRLZ44N
• - IRF3205

The first two MOSFETs are "Logic Level" according to the datasheet and should be suitable for use with Arduino's.

Method

For every MOSFET the drop voltage (Vds) was measured at different currents (Id).
The resistance Rds(on) and the dissipated power Pd were then calculated.
The MOSFETs were without a heatsink and the current was increased until the dissipated power reached about 1 watt.
During this experiment it was determined at which currents these MOSFET were deemed "too hot to handle" by using a finger probe.
All this was done for a gate voltage (Vg) of 3.3 V and 4.5 V. The choice for 4.5 V (rather than 5V), was made because the working voltage of Arduino's is just above 4.6 V when powered from USB. The MOSFETs will perform better at 5V.

Results

In the graph below it can be seen that all three MOSFETs performed OK at a gate voltage of 4.5V. The IRLZ44N had the lowest resistance and the least power dissipated. The IRF3205 performed worst, but it still did a good job, especially considering it is not advertised as a logic level MOSFET.


New graphs in the post below....


The IRF3205 was deemed too hot to handle at a current of 4.1 A. (You can keep your finger on 3.6 A, but not on 4.1 A) The IRLZ44N and the FQP30N06L could go 0.5 A higher.

At a gate voltage of 3.3V, again the IRLZ44N performed best. Remarkably, the "logic level" FQP30N06L now performed worse than the IRF3205. A second FQP30N06L was tested but gave the same results.

New graphs in the post below....

The MOSFETs were too hot to handle at IRLZ44N: 4.1 A, IRF3205: 3.6 A and FQP30N06L: 2.6 A.

Conclusions

All three MOSFETs can be used without a heatsink for the currents tested.
The FQP30N06L performs the worst at 3.3V. It gets rather hot and the maximum continuous current is limited. It does much better at 5V.
The IRF3205 is not advertised as a logic level MOSFET, but still does a pretty good job.
The IRLZ44N performs best at both voltages.

• 
  

 

I found one of my test leads to be defective. I have used this lead to measure the voltage drop in the graphs. I decided to do the measurement again. The new graphs have different values, but the conclusions are the same. I used my time during the measurement to clean up the graphs a bit. Here they are:

 

How many of each MOSFET type did you test?

 

How many of each MOSFET type did you test?

For the best and the worst performing MOSFET (FQP30N06L and IRLZ44N) I did a few extra readings with a second MOSFET on the curve to see if the values were the same. Especially for the FQP30NL06L, because I expected it to be defective. The values were pretty much the same. They all came from the same batch though. Maybe I will buy some more of the same type, just to see if they are the same.

 

For the best and the worst performing MOSFET (FQP30N06L and IRLZ44N) I did a few extra readings with a second MOSFET on the curve to see if the values were the same. Especially for the FQP30NL06L, because I expected it to be defective. The values were pretty much the same. They all came from the same batch though. Maybe I will buy some more of the same type, just to see if they are the same.

I decided you were right to ask that question and I have been trying to remember how rigorous have been. During the measurement I did some quick checks that made me comfortable to trust the result, but I did not write down the results. That is why I did 2 of the measurement again, both with two different MOSFETs. You can check the results below and draw your own conclusions. Please note that the original measurements have a better resolution (every 0.5A).

Also I ordered some more IRLZ44N. I hope they are from a different batch. Also on the way are some IRL2203N. The datasheet suggests that these might stay cool enough to touch at 5.1 A (which is the maximum of my power supply), even without a heatsink.

 

I applaud your diligence in testing.

You have two types of dissipation to deal with, DC due to Rdson and AC due to MOSFET C and miller
effect. If you are just switching static loads, on and off, at low rep rates, AC can be ignored.

Heatsink design -

https://toshiba.semicon-storage.com/info/docget.jsp?did=13417

https://www.infineon.com/dgdl/Infin...N.pdf?fileId=5546d4626102d35a01610888511573d1

http://www.ti.com/lit/an/slva462/slva462.pdf

https://electronics.stackexchange.c...n-connect-cooling-pad-on-pcb-backside-by-vias

https://www.digikey.com/Site/Global/Layouts/DownloadPdf.ashx?pdfUrl=F51974C9A6D544F1A7D8F119514B67FF



Regards, Dana.