Nice! One thing you didn't mention (presumably to avoid overcomplicating?) is the issue when using for example a 3.3v GPIO from a CPU to switch a high-side PFET (with pullup as you showed) on a 5v or 12v power rail - this is something people want to do quite often of course. -- Hints for newbies reading this; there's two issues, one more obvious than the other: 1) Let's say you don't have the pull-up resistor shown, and just drive the P-FET gate from your GPIO; typically your GPIO will only go to 3.3v which isn't high enough (relative to the e.g. 12v power rail) to turn the P-FET off. The fix for this appears to be to add the pull-up resistor to the power rail (which will mean the FET is normally off) and then use your GPIO in "open drain" a.k.a "open collector" mode (or just switch it to an input), end result being the GPIO only pulls down (to turn the load on) and doesn't try to pull it up to turn off (letting the resistor do that).. .. however the subtle issue then comes into play ... 2) You have to watch out for "input protection diodes" which almost all chips have on their GPIOs, which are diodes inside the chip that protect it from over-voltage, they go from the GPIO pin to the chip's VCC (e.g. 3.3v); if the input exceeds VCC by (say) 0.6v then the diode starts conducting, clamping the GPIO pin to stop it going any higher and damaging things. This however screws up your "open drain" pin; if the resistor on your PFET is pulling its gate to 12v, and you connect that gate to your GPIO, then very probably the input protection diode on the GPIO is going to start conducting, and it'll pull down the gate, again making the FET impossible to turn off regardless of what you do with the GPIO pin. ...The fix for this is to use a second (N-channel) fet/ or (NPN) transistor, similar to what was shown for the "latching on" in the video (but without the resistor to the load output to make it latch) - and switch _that_ with your GPIO. Hope that wasn't too verbose, but if you try a high-side PFET on a significantly higher voltage rail than your GPIO, you may well have the "it won't turn off!!" problem for one of the two reasons above.
An opto coupler does not work, but this neat isolated DC DC converter is easy to use and cheap. Search for "b0505s-1w" You connect your 5 V control signal to the input and the gate / source from the Mosfet to the output. Works great! :)
Lord Rutherford said once that "An alleged scientific discovery has no merit unless it can be explained to a barmaid." In this case, I am the barmaid craving that elusive understanding. Due to your clear, concise, and well-said explanation, I now feel confident enough in my understanding of these parts to breadboard a few experiments - and learn even more. I'd been slamming my head against the proverbial 'brick wall' in my attempts to understand proper MOSFET operation. Many existing guides/tutorials provide a set of 'newbie' formulae which tend to assume *some* pre-existing understanding of MOSFETs, while others provide examples using complimentary NFET and PFET packages, without explanation for which-formula-applies-to-which-channel FET. Again, Thank You for being the one clear tutor in a sea of obscurity.
LOL, I missed the crickets during edit until just now when I came to the youtube page and started looking around for crickets here in the room. Yes, the wires are small amount as I used "real" wires. (Note that this was not built on a solderless breadboard :) I have a 4 lead volt meter and could have nulled out the leads and wires were it critical.
Your better than any of my professors. I remember when we get our first led. That red light cost the shop 290.00 . only samples. Now they are everywhere.
Wonderful idea and great explanation. Never thought of using FET's in this way, but, I'm sold. God I wish there was a magic cheap alternative to the Relay for moderate current flow with once off triggering for on and off. Efficient and cheap switching is the holy grail of modern automated systems and it just seems to me like we're still in the stone age somehow.
subbed mate... i learn from you very quick . really appreciate you dropping in tech terms and explaining them broading my knowledge . In this one i liked your reference to the holes being bigger and therefore more resistive than the negative ions to movement. thanks for your effort.
Hi, Thanks for this video. It is working well when I follow your instructions. I tried controlling the gate alone but it did not work. I'm puzzled why I have to connect the Base to the Drain via 4.7k. Then using a 100r to the Base, driving 0v is OFF and using +5v to turn the Fet back ON.
Thanks for the tutorial, Appreciated. But one question you say the resistor pulls it back high when the gate is switched off. But what size resistor should we use? Thanks
Great video! Can you explain a bit about current through a MOSFET? Trying to make a MOSFET h-bridge, but can't get it to draw the same current as my BJT version...
Hi Bill, would you be able to do a video on circuits relating to various methods of making a constant current LED driver? This video reminded me of some of the logic behind the topology of the constant current LED driver circuits.
What is the comparison to the power consumed when on and off vs a low side? As in which is better for a mostly-on circuit and which is better for a mostly-off circuit? (from a power conservation stand-point?)
Hi. Such a great video. Informative and easy to understand. Thank you. You mention a "toolbox" of 100-odd circuits that you have. Is this perhaps written down, maybe in a book or some such artefact? Sounds like a very valuable resource. Kind regards Schalk
I see the following bands: Yellow, Purple, Red and Gold which means, if my eyes aren't deceiving me, that it's a 4.7Kohm resistor. However beyond that I don't have any idea as to the decision process leading to this choice of value...
what's the problem with shuting down RPi? 1 gpio to initiate halt, 2nd to provide hearbeat signal and cut the power after processing absence of heartbeat be that in software via MC or hardware with something like comparator and timer.
Around 7:10, Vgs(th) should be -1.0 V, not + 1.0V (ref: ww1.microchip.com/downloads/en/DeviceDoc/VP3203%20B082613.pdf ). A positive Vgs(th) would imply a N Chanel MOSFET.
In High School Electronics and later Going through the USAF's Basic Electronics course Hole Theory never settle well in my head. You simply can not have a flow of something that is not. LOL I always got by with More or Less then flowing nots. Missing Electrons I could fathom. Vacancies in valance bonds I could see. Flowing Holes ?
The "bubble" (small circle) is a common way to mark low-active signals. So for low-active input on Gate (P-channel MOSFET), just imagine the bubble on Gate and then you know in which direction to draw the arrow (always away from the bubble in order to avoid popping it) ... and vice versa (if you see the arrow, you know it's N- or P-channel, depending on whether the arrow would pop the imaginary bubble).
Nice! One thing you didn't mention (presumably to avoid overcomplicating?) is the issue when using for example a 3.3v GPIO from a CPU to switch a high-side PFET (with pullup as you showed) on a 5v or 12v power rail - this is something people want to do quite often of course.
-- Hints for newbies reading this; there's two issues, one more obvious than the other:
1) Let's say you don't have the pull-up resistor shown, and just drive the P-FET gate from your GPIO; typically your GPIO will only go to 3.3v which isn't high enough (relative to the e.g. 12v power rail) to turn the P-FET off. The fix for this appears to be to add the pull-up resistor to the power rail (which will mean the FET is normally off) and then use your GPIO in "open drain" a.k.a "open collector" mode (or just switch it to an input), end result being the GPIO only pulls down (to turn the load on) and doesn't try to pull it up to turn off (letting the resistor do that)..
.. however the subtle issue then comes into play ...
2) You have to watch out for "input protection diodes" which almost all chips have on their GPIOs, which are diodes inside the chip that protect it from over-voltage, they go from the GPIO pin to the chip's VCC (e.g. 3.3v); if the input exceeds VCC by (say) 0.6v then the diode starts conducting, clamping the GPIO pin to stop it going any higher and damaging things. This however screws up your "open drain" pin; if the resistor on your PFET is pulling its gate to 12v, and you connect that gate to your GPIO, then very probably the input protection diode on the GPIO is going to start conducting, and it'll pull down the gate, again making the FET impossible to turn off regardless of what you do with the GPIO pin.
...The fix for this is to use a second (N-channel) fet/ or (NPN) transistor, similar to what was shown for the "latching on" in the video (but without the resistor to the load output to make it latch) - and switch _that_ with your GPIO.
Hope that wasn't too verbose, but if you try a high-side PFET on a significantly higher voltage rail than your GPIO, you may well have the "it won't turn off!!" problem for one of the two reasons above.
Wouldn't a optoisolator also work?
An opto coupler does not work, but this neat isolated DC DC converter is easy to use and cheap. Search for "b0505s-1w" You connect your 5 V control signal to the input and the gate / source from the Mosfet to the output. Works great! :)
6:57 I love how the crickets kick in when he shows his lackluster list of P Channel MOSFET
Lord Rutherford said once that "An alleged scientific discovery has no merit unless it can be explained to a barmaid." In this case, I am the barmaid craving that elusive understanding. Due to your clear, concise, and well-said explanation, I now feel confident enough in my understanding of these parts to breadboard a few experiments - and learn even more.
I'd been slamming my head against the proverbial 'brick wall' in my attempts to understand proper MOSFET operation. Many existing guides/tutorials provide a set of 'newbie' formulae which tend to assume *some* pre-existing understanding of MOSFETs, while others provide examples using complimentary NFET and PFET packages, without explanation for which-formula-applies-to-which-channel FET. Again, Thank You for being the one clear tutor in a sea of obscurity.
LOL, I missed the crickets during edit until just now when I came to the youtube page and started looking around for crickets here in the room. Yes, the wires are small amount as I used "real" wires. (Note that this was not built on a solderless breadboard :) I have a 4 lead volt meter and could have nulled out the leads and wires were it critical.
Your better than any of my professors. I remember when we get our first led. That red light cost the shop 290.00 . only samples. Now they are everywhere.
Glad to see I am not the only builder still using the trusty Fluke 77!
Wonderful idea and great explanation. Never thought of using FET's in this way, but, I'm sold.
God I wish there was a magic cheap alternative to the Relay for moderate current flow with once off triggering for on and off.
Efficient and cheap switching is the holy grail of modern automated systems and it just seems to me like we're still in the stone age somehow.
Oh man, these crickets (?) are annoying! Very interesting video! Thanks :)
how to connect switch to MOSFET
This one is so useful. Learned a lot and I'm using it now in my circuits. Would like to know more of your bag of tricks circuits.
Also found out about the MAX16054 (pushbutton on/off controller) and use that as the pushbutton on/off switch for the p-channel mosfet
subbed mate... i learn from you very quick . really appreciate you dropping in tech terms and explaining them broading my knowledge . In this one i liked your reference to the holes being bigger and therefore more resistive than the negative ions to movement.
thanks for your effort.
I see someone owns the Black Magic book, awesome! I recommended this and the other one from the same author for my universitys library.
A D-type flipflop is easy to wire for a toggle with one button instead of two..
I use the same circuit in my RPi for power, except with a N-channel FET and a PNP transistor.
High side fet switching....
Use the output of a photovoltaic optocoupler as it floats. Thus can be used to add a voltage offset to the source.
Hi, Thanks for this video. It is working well when I follow your instructions. I tried controlling the gate alone but it did not work. I'm puzzled why I have to connect the Base to the Drain via 4.7k. Then using a 100r to the Base, driving 0v is OFF and using +5v to turn the Fet back ON.
Unfortunately, the 2n3904 is getting too hot using a 350mA load. I therefore have to use those expensive SSR. And low current input.
Thanks for the tutorial, Appreciated. But one question you say the resistor pulls it back high when the gate is switched off. But what size resistor should we use? Thanks
Great video! Can you explain a bit about current through a MOSFET? Trying to make a MOSFET h-bridge, but can't get it to draw the same current as my BJT version...
Next step, add a zener reference and a buffer to turn it into a voltage regulator? Still would always prefer ground side switching.
Hi Bill, would you be able to do a video on circuits relating to various methods of making a constant current LED driver? This video reminded me of some of the logic behind the topology of the constant current LED driver circuits.
You should put links in the description. The links on the screen at the end of the video don't work for iOS devices.
Excellent! Thanks for sharing..
What is the comparison to the power consumed when on and off vs a low side? As in which is better for a mostly-on circuit and which is better for a mostly-off circuit? (from a power conservation stand-point?)
Hi. Such a great video. Informative and easy to understand. Thank you.
You mention a "toolbox" of 100-odd circuits that you have. Is this perhaps written down, maybe in a book or some such artefact? Sounds like a very valuable resource.
Kind regards
Schalk
How many ohms is the bias resistor? Can it be pretty much anything over the maximum drain source resistance? Or should it match it closely?
I see the following bands: Yellow, Purple, Red and Gold which means, if my eyes aren't deceiving me, that it's a 4.7Kohm resistor.
However beyond that I don't have any idea as to the decision process leading to this choice of value...
How much of that measured Rds is actually the wires you used to hook up the FET on the proto board?
+eformance Probably just a small fraction.
what's the problem with shuting down RPi?
1 gpio to initiate halt, 2nd to provide hearbeat signal and cut the power after processing absence of heartbeat be that in software via MC or hardware with something like comparator and timer.
Super awesome explanation thank you
How would I identify the IGBT that operates like this?
Great stuff! Thanks!
Thank you!! 👍🙏
Great video, as always! Bil Herd for president! :D
beautiful
Thank you.
3.3v io, 85vdc, switching highside @8a, headache lol.
Thank you very much.:)
very good thanks
that is very interesting ..thanks
Around 7:10, Vgs(th) should be -1.0 V, not + 1.0V (ref: ww1.microchip.com/downloads/en/DeviceDoc/VP3203%20B082613.pdf ). A positive Vgs(th) would imply a N Chanel MOSFET.
In High School Electronics and later Going through the USAF's Basic Electronics course Hole Theory never settle well in my head. You simply can not have a flow of something that is not. LOL I always got by with More or Less then flowing nots. Missing Electrons I could fathom. Vacancies in valance bonds I could see. Flowing Holes ?
4:10 ... i dont get it with this bubble, even after several replays.
The "bubble" (small circle) is a common way to mark low-active signals. So for low-active input on Gate (P-channel MOSFET), just imagine the bubble on Gate and then you know in which direction to draw the arrow (always away from the bubble in order to avoid popping it) ... and vice versa (if you see the arrow, you know it's N- or P-channel, depending on whether the arrow would pop the imaginary bubble).
Nice :) Thumbs up :D
Alex
hahaha.. i subcrabed youuu
Got a circuit you wanna build? Test it out here first. inspect androidcircuitsolver on google
bout 7min in my brain goes is that a cricket? why am i hearing cricket? im on 3rd floor and its 20 degrees f outside...pause video...oh.
🦗