I am a great fan of reverse polarity tvs diodes, because I once bought a broken device for 80 Euro instead of 400 Euro because the previous owner plugged in the wrong power supply and the tvs diode failed short. It was an easy repair for me thought.
These educational videos are amazing and refreshing! By the way, this is just what I needed as it's a part of my project and Master's thesis as well. Keep it up Zach!
this is great. was thinking that connectors will not allow reverse polarity but then most electronics have batteries that are easily reversed. so reverse protection is a must if you sell your works. thank you
In the case of the P-channel - you mention the leaking current is what makes sure we are in the linear region. What does the calculation actually look like. It is just the leaking current from the mosfet datasheet multiplied with the Resistor (R) to see what the voltage would be - is it that simple? Any other pitfalls to be aware of with this design?
Yeah I would agree, I have used the n-channel on the power rail with charge pump driving, I'll have to do a video on this topic as it's pretty interesting IMO.
A P-channel MOSFET is usually my favourite but if the input voltage is already on the higher side of what I'd need, a series diode does the trick and is normally a little cheaper for very low cost stuff.
"Canal" or "kanal" is the name for "channel" such as P-channel Mosfet in most western European languages other than English. Including Germanic, Romance and Slavic languages. They don't generally have the word channel, but use canal/kanal instead. Both channel and canal are from Latin and means the same, but in English they are used a bit differently.
When we using MOSFET for reverse polarity protection, we need to use it for low ampere application. Because when MOSFET is used for DC current, we violating the SOA. For DC current, mosfet will generate lot heat than mosfet used in switching applications. Also from datasheet, 10 to 15% of Id(drain current) will be used to DC application. Its from my experience. Correct me if i am wrong. Also any experts design, pls provide details of selecting mosfet for DC current for 50A DC.
Zack sir for polarity identification supposed in the Gerber silkscreen layer is not given,we don't have assembly drawing,we don't have a schematic .can we find the polarity any rule by using track or by decoupling capacitors or near components are placed near to ic.any method to find the polarity of ic without data like assembly drawing, schematic diagram, silkscreen not given in the Gerber .
In the PMOS method, the MOSFET drain terminal is connected to the supply while the source terminal is connected to the circuit. How does this allow current flow yet for MOSFETs, current is from source to drain and not vice-versa?
as long as the gate is lower voltage (by Vgsth) than the source the P-MOSFET will pass current in either direction. There is nothing in the FET to stop current flowing in either direction unlike the bipolar transistor which has a directional (PN) junction to pass through.
I like the n-channel mosfet version since n-channels are the most common mosfets. Some reccomend a mosfet driver if it ever be used on the high(?) side
If you want 0 voltage drop and can sacrifice a little current put a diode (+resistor) in series with the coil of a rele to switch the rele on if polarity is right. If you use an LED you also have an "on" light.
I'm hesitant on the NPN option because I feel breaking the ground connection and forcing it through that pinch point would lead to more problems than interrupting the positive line. Anyone have experience?
Hey Zach, I have done research before about the P channel MOSFET type but couldn’t get any information why there would be a resistor. I know diode clamps the voltage for MOSFET but resistor is not explained. Is it for limiting the leakage current? If so, is the leakage current that high to be limited by a resistor? Also, I have seen values such as 1K, 10K, or 100K for the resistor. How do you calculate these values?
You need the resistor to get current flowing through the diode. You calculate resistor value based on the current you need. You would want the value to be as high as possible to have minimal loss.
@@geirlindahl3297 where else would it flow if resistor is not connected? And also as I said, the leakage current wouldn’t be too high already. So how would the calculation be made for a specific current value since we are not dealing with high values? Would it make the resistor values arbitrary chosen values?
@@ugurkeles7429 Lets say the turn on current for the zener, Iz, is 5mA. You should choose a resistor giving a minimum of that current. Assuming a voltage drop across the zener Vf=5V and input voltage 10V. That gives R=(10V-5V)/5m=1k
Switching GND can lead to issues when the load has external GND or signal connections. I’m a bit surprised this wasn’t mentioned. So a high-side switch/pFET might be the option with less downsides.
I personally use the P-Channel MOSFET almost all the time. What I don't like with the N-MOSFET option is that it introduces a voltage difference between the circuit GND and the real GND. I am not sure if that in reality makes a difference, but IMHO this is a headache that I want to avoid.
Some devices has specified input voltages as -48V for solving this problem. They basically use custom psu:s for them in field. N-Channel is usually used when you want minimal power losses on that component.
No mention of using n-channel MOSFETs in the positive side with a gate controller? These are common in automotive as you can easily add isolation with a 2nf FET and n-channel devices are cheaper and better than p.
the best way is the P-Channel mosfet because it will disconnect the live voltage side and not ground, so the rest of circuitry beyond it is not live, however in the N channel method, the rest of circuit is actually live it's just that there is no ground path. I wouldn't really consider the other methods at all unless you have size constraints or very very low power and don't mind the voltage drop.
Why include the Zener diode and resistor instead of just connecting the PFET gate directly to ground? That would give max Vsg and min RDSon. Just make sure the PFET's VSG rating is high enough to handle the full supply voltage across it. Also important to note that the transistors in your circuit are (intentionally) connected backwards relative to "normal" circuit use.
A FET with a high VGS rating may not be good at e.g. 3.x V. So it’s a tradeoff. They usually stop at 20 V VGS. Also MOSFET Gates are quite sensitive towards ESD events and power inputs are often big enough for you to potentially touch. So the z-diode is for me also ESD protection. I usually like to add an MLCC for additional/faster ESD mitigation.
Also I don’t think there is not such a case as the circuit you mentioned is never used. I have seen in a video of Phil’s Lab channel, only MOSFET is used for reverse polarity protection. I assume, of course, that was mainly because a simple and introductory hardware design video.
I didn't get that either. Can you think of a reason why they implemented it with the diode? Is it just to clamp the voltage across the gate if vgs is too high? Is it also considered an esd protection maybe?
my lab tests show that doing everything wrong but putting a silkscreen logo of zachs big arms protects my circuit from all evil effects
I'm really glad Zack took my suggestion seriously! 😊
I am a great fan of reverse polarity tvs diodes, because I once bought a broken device for 80 Euro instead of 400 Euro because the previous owner plugged in the wrong power supply and the tvs diode failed short.
It was an easy repair for me thought.
These educational videos are amazing and refreshing! By the way, this is just what I needed as it's a part of my project and Master's thesis as well. Keep it up Zach!
I was just researching this topic to implement on my circuit for the first time, properly.
How did you know??
The Zach works in mysterious ways
They know. It has also happened to me a couple of times.
That's a great video! In my opinion using MOSFETs is the best option in most embedded hardware designs.
Great video, as always!
this is great. was thinking that connectors will not allow reverse polarity but then most electronics have batteries that are easily reversed. so reverse protection is a must if you sell your works. thank you
In the case of the P-channel - you mention the leaking current is what makes sure we are in the linear region. What does the calculation actually look like. It is just the leaking current from the mosfet datasheet multiplied with the Resistor (R) to see what the voltage would be - is it that simple? Any other pitfalls to be aware of with this design?
In general I would avoid the N-channel as this puts loss in the common ground. This could lead to a bouncing reference at high current use.
Yeah I would agree, I have used the n-channel on the power rail with charge pump driving, I'll have to do a video on this topic as it's pretty interesting IMO.
Thank You so much for the information! ❤
A P-channel MOSFET is usually my favourite but if the input voltage is already on the higher side of what I'd need, a series diode does the trick and is normally a little cheaper for very low cost stuff.
"Canal" or "kanal" is the name for "channel" such as P-channel Mosfet in most western European languages other than English. Including Germanic, Romance and Slavic languages. They don't generally have the word channel, but use canal/kanal instead. Both channel and canal are from Latin and means the same, but in English they are used a bit differently.
When we using MOSFET for reverse polarity protection, we need to use it for low ampere application. Because when MOSFET is used for DC current, we violating the SOA. For DC current, mosfet will generate lot heat than mosfet used in switching applications. Also from datasheet, 10 to 15% of Id(drain current) will be used to DC application. Its from my experience. Correct me if i am wrong. Also any experts design, pls provide details of selecting mosfet for DC current for 50A DC.
What about using a full bridge rectifier for a dc input so that it can also flip to polarity no matter what ?
If the application requires lower current and can stand the voltage drop of 2 diodes, then this is also a solution.
Zack sir for polarity identification supposed in the Gerber silkscreen layer is not given,we don't have assembly drawing,we don't have a schematic .can we find the polarity any rule by using track or by decoupling capacitors or near components are placed near to ic.any method to find the polarity of ic without data like assembly drawing, schematic diagram, silkscreen not given in the Gerber .
In the PMOS method, the MOSFET drain terminal is connected to the supply while the source terminal is connected to the circuit. How does this allow current flow yet for MOSFETs, current is from source to drain and not vice-versa?
as long as the gate is lower voltage (by Vgsth) than the source the P-MOSFET will pass current in either direction. There is nothing in the FET to stop current flowing in either direction
unlike the bipolar transistor which has a directional (PN) junction to pass through.
@@Tokody The sorce has been tied to the bulk. Doesn't that make the device unidirectional?
I like the n-channel mosfet version since n-channels are the most common mosfets. Some reccomend a mosfet driver if it ever be used on the high(?) side
That's where charge pump driving is used, it can provide reverse polarity protection and surge protection so it's pretty useful.
very informative.
If you want 0 voltage drop and can sacrifice a little current put a diode (+resistor) in series with the coil of a rele to switch the rele on if polarity is right. If you use an LED you also have an "on" light.
I did a design with reverse polarity protection, just copy-pasted the mosfet approach the RP CM dev board uses
I'm hesitant on the NPN option because I feel breaking the ground connection and forcing it through that pinch point would lead to more problems than interrupting the positive line.
Anyone have experience?
drop a link for the intro music please. it's catchy.
Hey Zach, I have done research before about the P channel MOSFET type but couldn’t get any information why there would be a resistor. I know diode clamps the voltage for MOSFET but resistor is not explained. Is it for limiting the leakage current? If so, is the leakage current that high to be limited by a resistor? Also, I have seen values such as 1K, 10K, or 100K for the resistor. How do you calculate these values?
You need the resistor to get current flowing through the diode. You calculate resistor value based on the current you need. You would want the value to be as high as possible to have minimal loss.
@@geirlindahl3297 where else would it flow if resistor is not connected? And also as I said, the leakage current wouldn’t be too high already. So how would the calculation be made for a specific current value since we are not dealing with high values? Would it make the resistor values arbitrary chosen values?
@@ugurkeles7429 Lets say the turn on current for the zener, Iz, is 5mA. You should choose a resistor giving a minimum of that current. Assuming a voltage drop across the zener Vf=5V and input voltage 10V. That gives R=(10V-5V)/5m=1k
For reasanoble range of input woltage can connect a gate to ooposite rail directly. So, it's eliminate an additional zener and resistor. Am I right?
As long as you can guarantee that you'll never see more than the maximum V_GS at your supply input, yes.
Definitely preferring the N-Chanel over the P-Chanel because of lower R_DSon.
Switching GND can lead to issues when the load has external GND or signal connections. I’m a bit surprised this wasn’t mentioned.
So a high-side switch/pFET might be the option with less downsides.
I personally use the P-Channel MOSFET almost all the time.
What I don't like with the N-MOSFET option is that it introduces a voltage difference between the circuit GND and the real GND.
I am not sure if that in reality makes a difference, but IMHO this is a headache that I want to avoid.
Some devices has specified input voltages as -48V for solving this problem. They basically use custom psu:s for them in field. N-Channel is usually used when you want minimal power losses on that component.
@LimbaZero Gotcha, I can see how the NMOSFET can be more power efficient than PMOSFET in some applications. Good point.
No mention of using n-channel MOSFETs in the positive side with a gate controller? These are common in automotive as you can easily add isolation with a 2nf FET and n-channel devices are cheaper and better than p.
I did mention it in terms of charge pump gate driving of an n-MOSFET, but I'm leaving that topic for an upcoming video.
the best way is the P-Channel mosfet because it will disconnect the live voltage side and not ground, so the rest of circuitry beyond it is not live, however in the N channel method, the rest of circuit is actually live it's just that there is no ground path. I wouldn't really consider the other methods at all unless you have size constraints or very very low power and don't mind the voltage drop.
There are diodes with really low voltage drop... Check SBR (SUPER BARRIER RECTIFIERS)... 0.1V to 0.2V drop...
Why include the Zener diode and resistor instead of just connecting the PFET gate directly to ground? That would give max Vsg and min RDSon. Just make sure the PFET's VSG rating is high enough to handle the full supply voltage across it. Also important to note that the transistors in your circuit are (intentionally) connected backwards relative to "normal" circuit use.
That helps protect the FET gate from reverse voltage spikes. You can never be idiot-proof, but you can at least be idiot-resilient.
A FET with a high VGS rating may not be good at e.g. 3.x V. So it’s a tradeoff. They usually stop at 20 V VGS.
Also MOSFET Gates are quite sensitive towards ESD events and power inputs are often big enough for you to potentially touch. So the z-diode is for me also ESD protection. I usually like to add an MLCC for additional/faster ESD mitigation.
Also I don’t think there is not such a case as the circuit you mentioned is never used. I have seen in a video of Phil’s Lab channel, only MOSFET is used for reverse polarity protection. I assume, of course, that was mainly because a simple and introductory hardware design video.
I didn't get that either. Can you think of a reason why they implemented it with the diode? Is it just to clamp the voltage across the gate if vgs is too high? Is it also considered an esd protection maybe?
FW bridge on input.
Emeric Poulin is a French name. Canal is French for channel.
This video is a pretty good summary on circuit protection as well: th-cam.com/video/GkXqE4x5Y3A/w-d-xo.htmlsi=QYYv9Xu6nyCj91p1