Spring has officially sprung here in Colorado with a high today of a balmy 65°F. Never mind that Wednesday is supposed to be 34° and snowy - we're just going to go with the whole "Spring is here!" thing. Today is the first Monday of the month, which means we're back with another installment of "According to Pete." Check it out:
In this month's edition, Pete explores the wide world of MOSFETs. Give it a look and, as always, feel free to leave any comments, questions, or suggestions below. Hope you enjoy the video and have a great Monday!
To protect MOSFETs in circuits where the load can possibly fail in a short circuit mode, I recommend fuses which will open up to protect the PCB and associated wiring from being damaged due to the failed load.
Since standard fuses are a pain to replace, a very neat device to use is a PTC resettable fuse that heats up, then goes open when the load draws too much current. When the load no longer draws any current (like when the power is removed from the circuit), the PTC fuse cools down and allows current to once again flow to the load. Very nice.
When heat sinks are difficult to use, remember that the heat that the MOSFET will dissipate is proportional to the resistance of the channel. The lower this is, the less the MOSFET will dissipate. You can switch huge loads without heatsinks if you select the right MOSFET with a low channel resistance.
One can even parallel multiple MOSFETs to further reduce the channel resistance when using them to switch even higher current loads.
It is difficult to cover as big a topic as you did within a few minutes.
Great job!
Good advice! Many MOSFET driver chips also include current-sensing features that you can use to shut down the MOSFET when a fault occurs.
if you just want to protect the fet from shorts and the likes, it is muhc simpler to use something like ST omnifets, they look and behave like regular fets but have current limiting and over-temperature shutdown
Provided of course, that you turn them on and off fast enough. Unfortunately, this can lead into more problems, with gate current limits, and EMI radiated if Vgs rises faster than VS rises into the source lead inductance.
Turning high current fets on and off slowly is a primary cause of running hotter than you'd expect.
If you provide for a gate resistor, you can usually tune that value once you have hardware, to a value that achieves a good compromise.
Thanks Pete!
werd. Ninja edited that zener to a diode without saying early on.
So, this video has been up for years...and is still very helpful. Thanks Pete for bringing some of these concepts together for me.
Hey Pete, you should have snapped your fingers at 9:06 seconds, would've made a good visual affect!
Awesome video Pete. Thanks.
Ive been waiting for this topic to come up for awhile now. Thanks for finally covering it. I always use MOSFETS for all my switching needs because of the very straight forward operation of the PICs I use. Just thanks again. Nice job Pete
pls is the atmega 324p an 8bit micro contoller. am a beginner
Yep
There are logic-level MOSFETS. I believe that would take care of the biasing problem Pete talked about.
I love According to Pete! He clarifies so much. He's going to be the next Forrest M. Mims III. :D
More like Larry the Cable Guy of electronics, but thanks for the vote of confidence.
More Americans know Larry the Cable Guy than Forrest M. Mims III, if that's any consolation...
Not a big deal, but Vcc and Vdd aren't quite the same.
Indeed. When dealing with NPN BJTs, the correct label for V+ is 'vcc', while 'vdd' is the correct label when dealing with N-channel FETs.
In case anyone is curious: The V- label for BJTs is 'vee', and the V- label for FETs is 'vss'. The V+ and V- labels are swapped when dealing with PNP BJTs or P-channel FETs (EG: Vcc would refer to V- in a PNP circuit).
'Vcc' is also (technically improperly) used outside of dealings with transistors as an alias for V+.
At least, that's my understanding of it. You can find a wiki article on the topic here.
Doesn't matter how you label V+ and GND, as long as it's consistent and people know the difference.
It does when you're dealing with large power electronics systems, GND is Ground, and whatever you connect your circuit to is probably 0V or one of several 0V types, they very often aren't the same thing.
Meh. Bit of a letdown at the end: "FEEDBACK AT SPARKFUN DOT COM!" with the whole jazz hand thingy turned into a little _feedback@sparkfun.com_ :-P. On the other hand, very nice video.
Pete,
I enjoy these videos a lot and they offer some good information but you mentioned something about the body diode I would like to re-clarify. The body terminal in the majority of discrete MOSFETs is tied to the source terminal because that is the most common usage. The body terminal must be connected to something otherwise you have a BJT structure with the NPN but the base terminal would be floating causing random behavior. However, tying the body diode to the source terminal creates a parasitic diode from the body P-doped structure to the N-doped drain terminal. The body diode doesn't have any usefulness for ESD. The whole ESD issue for MOSFETs is the electric field created between the Gate and Body terminals. The oxide layer is so thin that you get enormous electric fields (electric field is Volts/meter) across the oxide layer. Some discrete MOSFETs do include ESD protection but it is a added TVS-type diode from Gate to Source.
Thanks for the video, keep them coming!
Well thanks for setting me straight. I thought I was going to catch some flack for this one.
Another thing to keep in mind with MOSFETs in high current drive circuit designs when driving inductive loads is that the protection diode built into the device can see currents due to back EMF that exceed the design specs of the protection diode, which may cause the MOSFET to fail.
This is because for most MOSFETs, the protection diode is designed to protect against ESD damage, not to absorb the large amount of energy that results from the collapse of the magnetic field after shutting off the drive to the inductive load.
When this is a concern, consider adding an external protection diode from drain to source which is rated to comfortably handle the currents that will occur when the MOSFET shuts off.
The reverse diode in a standard mosfet is there as a side effect of how mosfet are build, it is not there to protect against esd. And for most parts it has the same current rating as the fet in forward direction, but mind the power limits and the slow speed.
With inductive loads like relay the diode will function as a zener that a normal diode, clamping the voltage at the break down voltage. if the part is avalanche rated that is fine as long as you stay with in the energy limits
Adding a standard silicon diode in parallel with a fet will often have no effect since it will be in parallel with the internal diode which is hotter and thus have a lower on voltage
Which makes a Schottky diode a good choice since it will have a lower on voltage.
But it also has a much lower Reverse breakdown voltage in almost all cases, hence no good if your circuit exceeds that normally. Having said that, most hobbyist stuff probably wouldn't, but likely best in most cases to just use a suitably rated fast recovery diode, fast soft, or Zener, depending on the app.
It has been a long time, 20+ years, since I have worked with MOSFETs. I have a new project where they are perfect for the switching I need to do, so this was a great refresher.
I also have another project where I need to do some current limiting, 300ma-500ma at 3V-4V, along with some PWM. Dimming 'high power' LEDs. It would be nice to get a video about how to use a MOSFET in that situation.