This is an excellent explanation. I realy like the way you give context (NMOS Low-side and PMOS High-side) before explaining the problem (RDS is too high with PMOS High-side, but NMOS High-side circuit dont work) to finish with the solution (Cap. + diode bootstrap). Thank you very much.
I whacked the subscription bell soon after you said you use automotive spec in many designs. The automotive industry has ramped up its tech over the decades, and now it allows hobbyist a lesser expensive route to quality and robust electronics where previously there was only expensive aircraft silicon. With the advent of electric vehicles, that has only widened the selection and ubiquity of very interesting electronics. I’m happy the cost bar has lowered!
Very good video, I like your videos a lot. I recently learned that there are some optocoupler that are driver mosfet. Like the TLP591 - DC Input, Photovoltaic Output. They are not cheap, but the voltage in Gate is independent of the PSU of the load, which makes it immune to voltage drops and other problems.
It is very powerful circuit and cheap. I test it, work like charm!. Thanks you very much. I used boost converter but it is costly and design is not very good for small package. I used it for bicopter, I will upload soon.
Sweet! A use for all those cheap low Rs n-channel mosfets . I used a open collector output comparator to drive the bootstrap. Worked great well up into ultra sonic frequency.
Thanks. I like this idea. I'd love to see more hints like this. While tearing apart old consumer electronics P-Channel MOSFETs and Power PNP BJTs seem to be extremely rare. In fact the only place I know I will find a P-Channel MOSFET (without ordering anything) is on lithium battery protection PCBs, typically in a SOT-8 package. Even the cheapest Chinese adjustable power supply kits, sold for around $5 use a power NPN BJT. I'd like to see you explain the different configurations to avoid the use of power P-Ch/PNP devices. It's handy information to have in those situations where I want to build something now and can salvage a part instead of ordering something and waiting a week or more ;) BTW, if we are switching any power MOSFET with a stable Gate voltage, how easy/accurate is it to use the voltage drop of the Rds resistance in place of a shunt to measure the current? Would I need to measure the Rds resistance value of the actual part used? What's my potential measurement resolution? -Jake
Upcycle Electronics - As I have experienced MOSFET RDs on data, the problem is not the voltage, it's the extremely low current demand. It's almost like a charge dependant gate. Even static electric effects switch these things. So, in the absence of current, I don't know how to measure voltage. Even on a Tektronix 556, I have trouble sorting it out.
many many thanks, now i understand whats going on with bootstrapping, theres alot of indepth videos and very dry explanations out there, however yours was very interesting and in the short time perfectly understandable in comparison to the hours in class of trying to understand the subject matter, keep up the good work. Top, thanks once again.
@6:00 I would add a bleeding resistor with value depending on the loading and wiring conditions. On other side we may use a BJT transistor to connect the bleeding resistor to the cap to discharge it once the PWM is tunrned OFF. I was thinking about where else can we insert the bleeding resister! I need to build this circuit and try it for further testing.
7:53 you can increase holding on time by using a 1n400x diode instead of 1n4148. 1n400x has lower leakage. Learned that from Bob Pease. Also use non-electrolytic cap.
I must admit, it´s the first time I understand a bit of how bootstraping works, still I have 2-3 questions, and hope that any commenters/viewers are gonna explain a bit to me: 6:50 1: "the transistor will be on, so, its collector will be at ground, thus grounding any voltage to the gate of the Mosfet..." Does the gate need 24V to open (cause 24-12=12V are needed after all?), so, 12V going through the IN4148-diode, the 10K- and the 1K-resistor (while at the same time anyway connected to ground), surely isn´t enough to trigger the gate ??? The question may sound as being self-answering, but such stuff isn´t clear to me at all (that´s what you get, when learning electronics at YT ;-) 7:25 2: "so, the voltage of the load is gonna increase. and as it´s increasing, the voltage at the diode/cap-midpoint is gonna increase at the same rate..." I know caps unload their charge in milliseconds, so, more than a single cap-charge/cycle is demanded for it to work? Or is the gate-current-demand so low, that a single cap-charge per cycle is sufficient??? I mean, how long can the 24v be sustained, since the cap unloads quickly (in few ms), if the cap isn´t supposedly charged several times in a single cycle, thus keeping its 12v-charge-level constant??? Also 7:25 3: What is meant by "the voltage at the diode/cap-midpoint is gonna increase at the same rate (as by the load)", is the voltage at the diode/cap-midpoint the sum of the "battery"-voltage rushing-in, and being added to the 12v of the cap? In other words: diode/cap-midpoint has 12v from the cap plus 0v from the battery at the beggining, and as soon as battery-current rushes-in, you get 12v + "X"v, "X" growing at the same rate it grows at the load (as is apparent/logical), till you reach 24v??? Or does some other wizardry happen, that I didn´t get yet??? Thanks á priori for all answers !!!
This is great! Thanks for sharing! What would be the approach if I'm looking to leverage the pwm of an arduino and a high side n-channel mosfet to control 12V LEDs? Should I build a mosfet driver?
Thanks for the info... As I understand it correct, I can use 4 LL N-channel fets for an h-Bridge on 24V. As log as I feed the top ones 29V to switch using Optocouplers or so.... thank you!
I have a question , on the first example with the N channel you had a resistor going to ground (pulling down the gate ). can I put the resistor to vcc instead of ground ? So it would be pulling high unless the signal goes low , and the load would be allways turn on unless the signal turn it off
To drive a 1700V mosfet, I assume the same circuit will work if the components meets the voltage ratings, capacitor, BJT, diode and resistors to say 1500V
There are also p-channel MOSFETs with considerably lower Rdson like the IRF4905 with 20 mOhm or even the IXTP140P05T with just 9 mOhm. The p-channel MOSFETs for higher voltages have mainly higher values like the 500 V IXTP10P50P with 1 Ohm.
As time goes on manufacturers are making P-channel MOSFETs with lower RDS on resistance. P-channel MOSFETs have holes as majority carriers as N-channel have electrons so P-channel MOSFETs are slower in switching and have higher input capacitance (Ciss). The main purpose of the video was to show an example of bootstrapping as it is using in many gate driver ICs.
@@0033mer usually the main reason for the use of n-channel high side switches is, that they have comparable characteristics like the n-channel low side switch like gate threshold voltage and transconductance. And they are usually less expensive. The use of gate driver ICs is mostly their switching current capability and shoot-through protection. But the industry offers also n-MOS and p-MOS pairs with nearly comparable parameters like irfz24 / irf9z24 or irf540 / irf9540. But it is not recommented to operate them with a 12 V rail to rail switching signal, especially if they have low channel resistances. Even 100 to 200 mOhm devices can be destroyed by the energy of a 1 µs shoot through coming from a low ESR electrolytic capacitor.
If I were to use an n channel depletion type mosfet instead of enhancement, can I turn this circuit from 'normally closed' to 'normally open'? Also, can I control that npn transistor with a raspberry without killing it? What were to happen If I used an electrolytic cap instead of the cap used in this video? Thanks! Really handy video.
6:50 i'm confused, since the transistor is ON, wouldn't the mosfet be off because the node in between the transistor and 1Kr resistor is 12V (fed from VCC and the diode), in short words: not enough voltage to drive the mosfet on? Also, 7:14: i believe the capacitor is now configured as a voltage source, working as an "isolated power supply", adding up it's voltage to VCC, putting 24V (well, i'm not counting the voltage drop across 10Kr and 1Kr resistors, so it should be less) to the mosfet GATE, creating a voltage on the gate above VCC, making it conduct. You said here that 12V will turn the mosfet on, but that's not enough to turn it on, at least it was what you were saying before - i'm confused here, because when the capacitor is charging, it would reach VCC at some point, which would activate he mosfet. See what i meat here and please correct me if i'm wrong: imgur.com/a/CGDLOIV (ignore the left side where i say there will be only 12V, in fact that is used to drain the remaining voltage on the gate when the transistor is ON)
When the transistor is ON, mosfet gate drive is shunted to ground turning the mosfet OFF as stated. When the capacitor charges up to 12v it will start to turn on the load. As the load voltage increases so does the mosfet drive to 24 volts. This is the bootstrap operation. We have 24 volts from gate to GND, and 12 volts from gate to source turning on the mosfet with 12 volts across the load.
The capacitor doesnt "equal" 12 volts until just the point you turn off the grounding gate transistor. Before that, any "12v" on the cap is bled to ground before it can hit the gate. Thats why the gate resistor - lower resistance to ground thru the gate transistor so the "12v" charge goes to ground. When you turn off the transistor now that 12v will be applied to the gate. The other side of the cap was at "ground" but now the load turns on and the voltage drop across the load will "push up" the voltage on the cap. Meaning now the voltage at the gate will "float" up with load voltage, maintaining the same voltage difference between gate and drain that was there when the mosfet just turned on
Brilliant. Explained exactly as you need the information. P channels may be easier but I want to avoid them because of the high SD r. PS. I like the IRF1407. Their "F" but run OK off arduino. Also, I assume its going to work OK for brushless motor? Thanks!
great explanation. But i have a question why is MOSFET still working when is 24V on the gate, i looked up on datasheet and saw that max. voltage of Vgs is +-20V, why's that ?
The bootstrap circuit generates 24 volts from the gate to Ground. With 12 volts across the load the gate to source voltage (Vgs) will be 12 volts. Check video at 7:20.
can i add a resistor from the gate of an n-mosfet to 12V to keep the load ON until a normally opened reed switch connected from gate to ground is approached by a magnet (the magnet turns off the load)?
Great Video but It can be Replaced Also by using A Darlington Pair like TIP122 6A Or any other as they can operate in High Frquency also if I m not wrong
It looks like working but i cant measure any voltage at the gate of mosfet I took my oscilloscope and measure the gate to source voltage and it says 0V but the load gets 2A current and mosfet doesn't heat up much.I'm using it with 100kHz Very confused 😅
QUESTION: the datasheet for IRF1405 says that Vgs is ± 20V. So now with bootstrapping Vgs becomes +24V exceeding the maximum allowed. Why is this not damaging your MOSFET? What if my Vcc is 48V. Then Vgs would be 96V. Would this not destroy my MOSFET? If not then I need to understand what is the real meaning behind Vgs. Thank you.
The bootstrap circuit generates 24 volts from the gate to GROUND. With 12 volts across the load the gate to source voltage (Vgs) will be 12 volts. Check video at 7:20.
Thanks so much, great video! Can you recommend a good p-channel mosfet for high side switching, that can be driven by logic, and control 3.3-5 v? Want to interface to arduino.
I am new. Doesn’t the RDS on .8ohm issue go away after a few ms so we could use a P channel mosfet in continuous mode and avoid the capacitor draining/switch off?
RDS on resistance is the On-state resistance of the Mosfet when it is in saturation as indicated in the data sheet, with the proper Vgs applied. The RDS on resistance value does not go away after switching. It actually increases with temperature over time. For a lower RDS on for a P channel check other part numbers for a lower spec.
@@0033mer thanks for this excellent video and your reply. I'm learning quite a bit from your other videos. I am really interested in high side switching for a project I'm working on. I'm interested in this F5305S based module with 817 opto isolation RDS on .06 ohms vs .8ohms. www.irf.com/product-info/datasheets/data/irf5305s.pdf . The module is shown in this video. ruclips.net/video/eAANkWDvusU/видео.html . I'm driving a large 300A bi-stable relay with a second pulse . I'm guessing the P channel mosfet have improved since your video. Also I think I need to understand how to measure the issue with RDS on, I'm guessing you may get voltage drop and heat with the higher RDS on?
Yes .. its all about getting the lowest RDS on so power dissipation is low with a lower running temperature. MOSFET technology is always changing so specs are always getting better for P channel MOSFETS. RDS on is the voltage drop across the Drain/Source divided by the Drain current when the MOSFET is fully turned on. Check data sheet for best spec and ensure turn on time is fast so you don't dwell too long in the linear mode region.
You don't really need the 1kΩ resistor. And reducing the 10kΩ resistor, in addition, should allow it to switch faster (=less partially on time). If short transient times are important, a push-pull driver could be used - but that requires 2 more transistors. Anyway, it's a good basic ciruit and you explained it well.
You are correct. The orginal circuit was built as you described but was modified as shown to be driven with CMOS logic which has limited sink current capabilities. The resistors limit the current from the charge on the gate to source capacitance and from the bootstrap capacitor. The circuit is for demonstration purposes only and if high speed is need a high side gate driver IC should be considered. Thanks for your feedback.
excellent job , this has clearedup an issue I was struggling to accept , have you gone on from this vide to discuss charge pumps which I also need an additional awareness of
What is the lower and higher limit of the frequency that this works? I used with an arduino pwm .. but the voltage over the gate doesn't rise to the 24V. Which means that the frequency is not suitable for your circuit. I don't know how to make it works.
Be aware, that the circuit will not work with any PWM or several Khz signal. This works only for very low frequencies from Arduino. To get the same effect, you need to use a boost converter to give double as much voltages to the gate which let the MOSFET to switch ON totally.
Sometimes the only option is to source feed the load as one side is connected to ground. An example would be a car horn, fuel pump, headlight ... etc. which is bolted to the frame of the car to get its ground connection, and you have to feed it switched 12 volts for activation.
@@0033mer I see, but no advantage. So rather than insulate the device from the return and low-side switch it, just do it this way. That makes sense. I guess I always knew that, but this video just got me to thinking. Thanks.
1. The 10k resistor to the transistor collector is also a leakage path for the bootstrap capacitor, albeit very tiny amount. 2. This circuit wouldn't work for very high speed switching right (~30+ KHz) ? The p channel bjt and the added capacitance being a bottleneck
This is a Demo circuit to demonstrate how bootstrapping works and not designed for high speed applications. The 10k resistor and the gate/source capacitance restricts the speed of the circuit. The 10k resistor is used to limit the current when the voltage on the gate/source capacitance is discharged so it can be driven with a CMOS logic gate without exceeding the maximum sink current. The 10k resistor can be omitted when the circuit is driven with an NPN transistor.
All semiconductor junctions have some leakage so the diode, mosfet and transistor would all contribute to the total leakage of the finite charge on the capacitor. The blocking diode is common to all bootstrap circuits so it was mentioned as the leaking source. The other component can differ in different designs. ( the transistor could be replaced with a relay contact) The purpose of the video was to demonstrate how a simple bootstrap circuit functions.
@@0033mer can you please explain this. When I drive a mosfet(NPN) from the output of a 555timer it turns on for 2 seconds when there is a led connected at the drain of the mosfet, but when I connect a small motor in parallel with the led the time reduces to 800ms.
Hello! is it important the frequency that you got in the pwm signal? Because I calculated the RC time constant is 22ms. Does this work well with a high frequency?
@@0033mer and what about if I wanna run it at high frequencies. Should I just change the capacitor value to sth smaller or there's another stuff I should thing about?
Think of the same circuit but use a bipolar PNP transistor as a high side switch. The emitter would be connected to VCC. Check out the schematic in this video: ruclips.net/video/-KuEyM1otls/видео.html
Hello, and thanks for the excellent tutorial. Quick question: I've seen you say multiple times in the comments that this setup is not intended for high-frequency applications, and that one should use an integrated circuit for such a case. What is the advantage of using an IC over switching out components for lower-valued ones that could support a higher frequency?
All MOSFETs have Gate/Source capacitance so switching them on/off very fast is not an easy task. There are ICs that will do this properly and keep the MOSFET out of its ohmic region during switching. This keeps the power dissipation down to a minimum and avoids MOSFET failure. To understand gate/source capacitance better check out these videos: ruclips.net/video/zTHzaNdByVA/видео.html ruclips.net/video/mjIubJeTRyY/видео.html
Yes .. you are correct. Technically Ciss = Cgs + Cgd and is the "input" capacitance as a whole for the Mosfet. Ciss is the capacitance between the gate and source terminals with the drain shorted to the source.(Cgd in parallel with Cgs) During switching the Cgd actually amplifies the Cgs so the Ciss is a spec used to compare different Mosfets input capacitances.
Never understood the concept mosfet bootstrapping. This is something I need to know. A while back I experimented with a three half-H bridge to try to drive a BLDC motor. At least I know how it's possible to build a high side switch with an N-channel mosfet.
If you can build one of these bootstrap circuits, you can essentially build up a 24v rail and then feed into the gates of your h-bridge with smaller transistors.
@@OtherDalfite _"...you can essentially build up a 24v rail and then feed into the gates..."_ That is an excellent idea. Just today I started researching charge pumps. That's another circuit I want to build. Thanks for he tip.
@@OtherDalfite It'll help. I'm just a hobbyist but I have to say I do expect to run into a few hurdles. A couple years ago I breadboarded a triple half-H bridge using transistors and never could get it to run smoothly. The hfe of each transistor never matched and add to that trying to get complimentary pnp's and npn's to cooperate with each other. If there is a charge pump IC I'll order mosfet driver IC's too and try to build the circuit that way. I know I'm talking a lot but now I'm thinking about ordering parts so I can test it out on my oscilloscope.
@@3deeguy hey, was doing some simulations using this idea as a springboard. Currently the best way I have to do this is get a voltage doubler (a bit less complex than bootstrapping) and make a dedicated rail circuit 10v or so above your Vds. Then use an opto-coupler to send the voltage to the gate of the mosfet you're wanting to control. Opto-coupler can be controlled with a 5v PWM from an Arduino. If anyone has a better way just let me know. Tried normal npn/pnp transistors but couldn't get a configuration that worked.
+2 component solution for high side n channel switching... photovoltatic optocouplers. logic level control through resistor and optocoupler's led. Wire the optocoupler's photodiode between mosfet source and gate. The gate voltage is then the source voltage plus what ever the photodiode is outputting.
This is an excellent explanation.
I realy like the way you give context (NMOS Low-side and PMOS High-side) before explaining the problem (RDS is too high with PMOS High-side, but NMOS High-side circuit dont work) to finish with the solution (Cap. + diode bootstrap).
Thank you very much.
really good and simple. thanks!!
This is how tutorials should be presented. Excellent.
Thank you. For the past hour I couldn't tell how the capacitor and voltage source acted in series until you motioned with your pencil.
Well organized, well presented, clear and concise. Thank you for an excellent tutorial!
You're very welcome! Glad it was helpful.
Finally someone that explained exactly how a "boot-strapping" circuit works 😍😍
Glad it was helpful!
I whacked the subscription bell soon after you said you use automotive spec in many designs. The automotive industry has ramped up its tech over the decades, and now it allows hobbyist a lesser expensive route to quality and robust electronics where previously there was only expensive aircraft silicon. With the advent of electric vehicles, that has only widened the selection and ubiquity of very interesting electronics. I’m happy the cost bar has lowered!
Thank you for this turorial. Bootstrap driving of mosfets is now much clearer thanks to you.
Glad it helped!
finally, you make it clean and clearer, appreciate it.
Glad it helped!
In addition to the bootstrap circuit you mentioned the charge pump circuit. Thank you for sharing your knowledge!
You are welcome!
I can't say how grateful I am for this video. Thank you very much!
Your Welcome!
ohhhhh... that's why my solenoids were being weird with my optocouplers.. thanks so much for this clear explanation.
thank you so much, I've been struggling to understand bootstrap and finally you have made me understand it very clearly
Glad it helped!
Very good video, I like your videos a lot.
I recently learned that there are some optocoupler that are driver mosfet. Like the TLP591 - DC Input, Photovoltaic Output. They are not cheap, but the voltage in Gate is independent of the PSU of the load, which makes it immune to voltage drops and other problems.
I like mosfets and i like bootstrapping, now I know how to do both at the same time!
Your videos are the best!! Your explanations are very clear and entertaining
Thanks for the feedback!
I FINALLY get it now. Thank you for this wonderful explanation.
Glad it was helpful!
It is very powerful circuit and cheap. I test it, work like charm!. Thanks you very much. I used boost converter but it is costly and design is not very good for small package. I used it for bicopter, I will upload soon.
Thanks, this subject was playing in my head today
Sweet! A use for all those cheap low Rs n-channel mosfets . I used a open collector output comparator to drive the bootstrap. Worked great well up into ultra sonic frequency.
Thanks. I like this idea. I'd love to see more hints like this.
While tearing apart old consumer electronics P-Channel MOSFETs and Power PNP BJTs seem to be extremely rare. In fact the only place I know I will find a P-Channel MOSFET (without ordering anything) is on lithium battery protection PCBs, typically in a SOT-8 package. Even the cheapest Chinese adjustable power supply kits, sold for around $5 use a power NPN BJT.
I'd like to see you explain the different configurations to avoid the use of power P-Ch/PNP devices. It's handy information to have in those situations where I want to build something now and can salvage a part instead of ordering something and waiting a week or more ;)
BTW, if we are switching any power MOSFET with a stable Gate voltage, how easy/accurate is it to use the voltage drop of the Rds resistance in place of a shunt to measure the current? Would I need to measure the Rds resistance value of the actual part used? What's my potential measurement resolution?
-Jake
Upcycle Electronics - As I have experienced MOSFET RDs on data, the problem is not the voltage, it's the extremely low current demand. It's almost like a charge dependant gate. Even static electric effects switch these things. So, in the absence of current, I don't know how to measure voltage. Even on a Tektronix 556, I have trouble sorting it out.
Boonedock Journeyman Rds would be Drain-Source voltage divided by Drain-Source current. Are you talking Gate current ?
I love you. Ty for saving my presentation in university
Same here!
Great idea bud. Very simple and probably very functional even in long period cycles if one use good quality, low leakage parts.
3:11 how this can be done if both mos are n type in Half bridge construction ??
In which upper mosfet ground is floating .
Ooh you did mentioned that later in the video ...
Monitory H carrier of p type
Congratulations from Brazil.
many many thanks, now i understand whats going on with bootstrapping, theres alot of indepth videos and very dry explanations out there, however yours was very interesting and in the short time perfectly understandable in comparison to the hours in class of trying to understand the subject matter, keep up the good work. Top, thanks once again.
Thanks for the feedback!
@6:00 I would add a bleeding resistor with value depending on the loading and wiring conditions. On other side we may use a BJT transistor to connect the bleeding resistor to the cap to discharge it once the PWM is tunrned OFF. I was thinking about where else can we insert the bleeding resister! I need to build this circuit and try it for further testing.
The circuit was meant to simply demonstrate how bootstrapping works and is not optimized for speed or PWM applications
7:53 you can increase holding on time by using a 1n400x diode instead of 1n4148. 1n400x has lower leakage. Learned that from Bob Pease. Also use non-electrolytic cap.
Why non electrolytic?
this is excellent design! the explanation is better!!
My eyes was messing with me! I was scrolling down, and im like what the heck is Boost trapping! Ha-ha
I must admit, it´s the first time I understand a bit of how bootstraping works, still I have 2-3 questions, and hope that any commenters/viewers are gonna explain a bit to me:
6:50 1: "the transistor will be on, so, its collector will be at ground, thus grounding any voltage to the gate of the Mosfet..."
Does the gate need 24V to open (cause 24-12=12V are needed after all?), so, 12V going through the IN4148-diode, the 10K- and the 1K-resistor (while at the same time anyway connected to ground), surely isn´t enough to trigger the gate ???
The question may sound as being self-answering, but such stuff isn´t clear to me at all (that´s what you get, when learning electronics at YT ;-)
7:25 2: "so, the voltage of the load is gonna increase. and as it´s increasing, the voltage at the diode/cap-midpoint is gonna increase at the same rate..."
I know caps unload their charge in milliseconds, so, more than a single cap-charge/cycle is demanded for it to work?
Or is the gate-current-demand so low, that a single cap-charge per cycle is sufficient??? I mean, how long can the 24v be sustained, since the cap unloads quickly (in few ms), if the cap isn´t supposedly charged several times in a single cycle, thus keeping its 12v-charge-level constant???
Also 7:25 3: What is meant by "the voltage at the diode/cap-midpoint is gonna increase at the same rate (as by the load)", is the voltage at the diode/cap-midpoint the sum of the "battery"-voltage rushing-in, and being added to the 12v of the cap? In other words: diode/cap-midpoint has 12v from the cap plus 0v from the battery at the beggining, and as soon as battery-current rushes-in, you get 12v + "X"v, "X" growing at the same rate it grows at the load (as is apparent/logical), till you reach 24v??? Or does some other wizardry happen, that I didn´t get yet???
Thanks á priori for all answers !!!
Awesome video, this is what I was looking for :D
Very practical and Very useful circuit idea. Thanks.
Interesting circuit nice refresher for an old timer
This is great! Thanks for sharing! What would be the approach if I'm looking to leverage the pwm of an arduino and a high side n-channel mosfet to control 12V LEDs? Should I build a mosfet driver?
Thanks for the info...
As I understand it correct, I can use 4 LL N-channel fets for an h-Bridge on 24V. As log as I feed the top ones 29V to switch using Optocouplers or so....
thank you!
Can i use Gate drive Transformer instead of GDIc ? What will be the disadvantages ?
I wanna use it for driving Half bridge of Class D Amplifier?
I have a question , on the first example with the N channel you had a resistor going to ground (pulling down the gate ). can I put the resistor to vcc instead of ground ? So it would be pulling high unless the signal goes low , and the load would be allways turn on unless the signal turn it off
thank you for making this video. i see you have few views and thats a shame cause it helped me so much.
keep it up! I wan't MORE!
my first bootstrap configuration ever made was due to you and you're video :)
Thanks for the feedback.
Check out these videos:
ruclips.net/video/zTHzaNdByVA/видео.html
ruclips.net/video/mjIubJeTRyY/видео.html
You just earned yourself another subscriber :D Excellent explanation. Thank you :)
Thanks and welcome
Why does the collector ground at the beginning? And if it is ground, the node after 1n4148 a ground also, isn’t it? Then how the capacitor charges up?
So would this circuit work for PWM as long as it didn't hit 100% Duty cycle?
If I'm understanding it correctly, it should
Very very very very very informative 🔥🔥
Glad it was helpful!
To drive a 1700V mosfet, I assume the same circuit will work if the components meets the voltage ratings, capacitor, BJT, diode and resistors to say 1500V
Best explanation, appreciate
Glad it was helpful! Thanks for the visit.
There are also p-channel MOSFETs with considerably lower Rdson like the IRF4905 with 20 mOhm or even the IXTP140P05T with just 9 mOhm. The p-channel MOSFETs for higher voltages have mainly higher values like the 500 V IXTP10P50P with 1 Ohm.
As time goes on manufacturers are making P-channel MOSFETs with lower RDS on resistance. P-channel MOSFETs have holes as majority carriers as N-channel have electrons so P-channel MOSFETs are slower in switching and have higher input capacitance (Ciss). The main purpose of the video was to show an example of bootstrapping as it is using in many gate driver ICs.
@@0033mer usually the main reason for the use of n-channel high side switches is, that they have comparable characteristics like the n-channel low side switch like gate threshold voltage and transconductance. And they are usually less expensive.
The use of gate driver ICs is mostly their switching current capability and shoot-through protection.
But the industry offers also n-MOS and p-MOS pairs with nearly comparable parameters like irfz24 / irf9z24 or irf540 / irf9540. But it is not recommented to operate them with a 12 V rail to rail switching signal, especially if they have low channel resistances.
Even 100 to 200 mOhm devices can be destroyed by the energy of a 1 µs shoot through coming from a low ESR electrolytic capacitor.
Please can 311V be the input to the entire circuit? Plus the bootstrap circuit
That was a great video. Thanks !!
You are welcome!
Excellent. Thank you very much.
You are welcome!
Thank you, best explanation
Glad it was helpful!
TNX 4 the upload
73 N8AUM
If I were to use an n channel depletion type mosfet instead of enhancement, can I turn this circuit from 'normally closed' to 'normally open'?
Also, can I control that npn transistor with a raspberry without killing it? What were to happen If I used an electrolytic cap instead of the cap used in this video?
Thanks! Really handy video.
6:50 i'm confused, since the transistor is ON, wouldn't the mosfet be off because the node in between the transistor and 1Kr resistor is 12V (fed from VCC and the diode), in short words: not enough voltage to drive the mosfet on?
Also, 7:14: i believe the capacitor is now configured as a voltage source, working as an "isolated power supply", adding up it's voltage to VCC, putting 24V (well, i'm not counting the voltage drop across 10Kr and 1Kr resistors, so it should be less) to the mosfet GATE, creating a voltage on the gate above VCC, making it conduct. You said here that 12V will turn the mosfet on, but that's not enough to turn it on, at least it was what you were saying before - i'm confused here, because when the capacitor is charging, it would reach VCC at some point, which would activate he mosfet.
See what i meat here and please correct me if i'm wrong: imgur.com/a/CGDLOIV
(ignore the left side where i say there will be only 12V, in fact that is used to drain the remaining voltage on the gate when the transistor is ON)
When the transistor is ON, mosfet gate drive is shunted to ground turning the mosfet OFF as stated.
When the capacitor charges up to 12v it will start to turn on the load. As the load voltage increases so does the mosfet drive to 24 volts.
This is the bootstrap operation. We have 24 volts from gate to GND, and 12 volts from gate to source turning on the mosfet with 12 volts across the load.
The capacitor doesnt "equal" 12 volts until just the point you turn off the grounding gate transistor. Before that, any "12v" on the cap is bled to ground before it can hit the gate. Thats why the gate resistor - lower resistance to ground thru the gate transistor so the "12v" charge goes to ground. When you turn off the transistor now that 12v will be applied to the gate. The other side of the cap was at "ground" but now the load turns on and the voltage drop across the load will "push up" the voltage on the cap. Meaning now the voltage at the gate will "float" up with load voltage, maintaining the same voltage difference between gate and drain that was there when the mosfet just turned on
Brilliant. Explained exactly as you need the information. P channels may be easier but I want to avoid them because of the high SD r. PS. I like the IRF1407. Their "F" but run OK off arduino. Also, I assume its going to work OK for brushless motor? Thanks!
Great video!! Thank you for uploaded it.
Glad you enjoyed it!
the diode and cap solution, is that equal to a charge pump circuit? cause u doubled the 12 Volts?
Thanks a lot for video! it opened a new window for me.
is there a way to control the high side of a circuit with 2 n channel mosfets controlled with a pin of an arduino
great explanation. But i have a question why is MOSFET still working when is 24V on the gate, i looked up on datasheet and saw that max. voltage of Vgs is +-20V, why's that ?
The bootstrap circuit generates 24 volts from the gate to Ground. With 12 volts across the load the gate to source voltage (Vgs) will be 12 volts. Check video at 7:20.
@@0033mer hahha oh that's right, my bad, thanks for fast response.
Thanks a lot man, great video, great explanation
You're welcome!
Good video. Thanks for this idea.
can i add a resistor from the gate of an n-mosfet to 12V to keep the load ON until a normally opened reed switch connected from gate to ground is approached by a magnet (the magnet turns off the load)?
Yes, that would work.
@@0033mer thank you, hope to watch new videos soon
@@0033mer i can confirm it's working, thank you again.
Very nice explanation , thank you !
You are welcome!
Great Video but It can be Replaced Also by using A Darlington Pair like TIP122 6A Or any other as they can operate in High Frquency also if I m not wrong
No need to bootstrap with a Darlington. It will work but Vsat can be as high as 2 volts so power dissipation is a concern.
It looks like working but i cant measure any voltage at the gate of mosfet
I took my oscilloscope and measure the gate to source voltage and it says 0V but the load gets 2A current and mosfet doesn't heat up much.I'm using it with 100kHz
Very confused 😅
Can this work with 400v across MOSFET in 3 phase
QUESTION: the datasheet for IRF1405 says that Vgs is ± 20V. So now with bootstrapping Vgs becomes +24V exceeding the maximum allowed. Why is this not damaging your MOSFET? What if my Vcc is 48V. Then Vgs would be 96V. Would this not destroy my MOSFET? If not then I need to understand what is the real meaning behind Vgs. Thank you.
The bootstrap circuit generates 24 volts from the gate to GROUND. With 12 volts across the load the gate to source voltage (Vgs) will be 12 volts. Check video at 7:20.
Thanks so much, great video! Can you recommend a good p-channel mosfet for high side switching, that can be driven by logic, and control 3.3-5 v? Want to interface to arduino.
There are a few out there but I have used the NDP6020P and they are fairly easy to obtain.
www.sparkfun.com/products/12901
I am new. Doesn’t the RDS on .8ohm issue go away after a few ms so we could use a P channel mosfet in continuous mode and avoid the capacitor draining/switch off?
RDS on resistance is the On-state resistance of the Mosfet when it is in saturation as indicated in the data sheet, with the proper Vgs applied. The RDS on resistance value does not go away after switching. It actually increases with temperature over time. For a lower RDS on for a P channel check other part numbers for a lower spec.
@@0033mer thanks for this excellent video and your reply. I'm learning quite a bit from your other videos. I am really interested in high side switching for a project I'm working on. I'm interested in this F5305S based module with 817 opto isolation RDS on .06 ohms vs .8ohms. www.irf.com/product-info/datasheets/data/irf5305s.pdf . The module is shown in this video. ruclips.net/video/eAANkWDvusU/видео.html . I'm driving a large 300A bi-stable relay with a second pulse . I'm guessing the P channel mosfet have improved since your video. Also I think I need to understand how to measure the issue with RDS on, I'm guessing you may get voltage drop and heat with the higher RDS on?
Yes .. its all about getting the lowest RDS on so power dissipation is low with a lower running temperature. MOSFET technology is always changing so specs are always getting better for P channel MOSFETS. RDS on is the voltage drop across the Drain/Source divided by the Drain current when the MOSFET is fully turned on. Check data sheet for best spec and ensure turn on time is fast so you don't dwell too long in the linear mode region.
You saved my energy
Thanks so much, great video! ,This mosfet IRF1405 and
Will this circuit you explain work with 24V
You don't really need the 1kΩ resistor. And reducing the 10kΩ resistor, in addition, should allow it to switch faster (=less partially on time).
If short transient times are important, a push-pull driver could be used - but that requires 2 more transistors.
Anyway, it's a good basic ciruit and you explained it well.
You are correct. The orginal circuit was built as you described but was modified as shown to be driven with CMOS logic which has limited sink current capabilities. The resistors limit the current from the charge on the gate to source capacitance and from the bootstrap capacitor. The circuit is for demonstration purposes only and if high speed is need a high side gate driver IC should be considered. Thanks for your feedback.
excellent job , this has clearedup an issue I was struggling to accept , have you gone on from this vide to discuss charge pumps which I also need an additional awareness of
Check out this video: ruclips.net/video/wkSrmPBDRys/видео.html
Awesome Explanation. 5 stars ★★★★★
Thanks for the feedback.
What is the lower and higher limit of the frequency that this works? I used with an arduino pwm .. but the voltage over the gate doesn't rise to the 24V. Which means that the frequency is not suitable for your circuit.
I don't know how to make it works.
It is a low frequency circuit designed to demonstrate Mosfet Bootstrapping. For fast switching use a Mosfet driver IC.
I can't get this to work! The waveform at my mosfet gate looks sawtooth instead of square, any idea what I'm doing wrong?
Double check your circuit wiring with the schematic.
Lower the 10K to 1k
...and make sure the tranistor is deep into saturation, so low drop across Collector - emitter
Nice video, thanks.
Be aware, that the circuit will not work with any PWM or several Khz signal. This works only for very low frequencies from Arduino. To get the same effect, you need to use a boost converter to give double as much voltages to the gate which let the MOSFET to switch ON totally.
The described circuit is for demonstration purposes. Check the description box for an application note.
can u do a similar video but using a dedicated gate driver chip
What exactly is the benefit of powering your load via the high-side vs. the low-side?
Sometimes the only option is to source feed the load as one side is connected to ground. An example would be a car horn, fuel pump, headlight ... etc. which is bolted to the frame of the car to get its ground connection, and you have to feed it switched 12 volts for activation.
@@0033mer I see, but no advantage. So rather than insulate the device from the return and low-side switch it, just do it this way. That makes sense. I guess I always knew that, but this video just got me to thinking. Thanks.
Thank you. Didn't know about this technique. Can you make a video about different types of mosfets, their main characteristics and circuits?
Check out these videos:
ruclips.net/video/zTHzaNdByVA/видео.html
ruclips.net/video/mjIubJeTRyY/видео.html
great explanation. thank you soo much
You're welcome!
When the transistor is on, won't the supply be shorted with with path of 12V - diode-10k-transistor-gnd ?
The 10k ohm resistor in series ensures the power supply is not shorted out.
@@0033mer yeah... Sorry...😅 A dumb question...😅 Didn't notice the guy there...😅
1. The 10k resistor to the transistor collector is also a leakage path for the bootstrap capacitor, albeit very tiny amount.
2. This circuit wouldn't work for very high speed switching right (~30+ KHz) ? The p channel bjt and the added capacitance being a bottleneck
This is a Demo circuit to demonstrate how bootstrapping works and not designed for high speed applications. The 10k resistor and the gate/source capacitance restricts the speed of the circuit. The 10k resistor is used to limit the current when the voltage on the gate/source capacitance is discharged so it can be driven with a CMOS logic gate without exceeding the maximum sink current. The 10k resistor can be omitted when the circuit is driven with an NPN transistor.
Tnks for tutorial,, but if the voltage more than 12volt, example from automotive battery ? Need change mosfet
Great video!!
Glad you enjoyed it
Excellent. Thank you!
If one were to increase the value of the cap will the duty cycle increase for the on state of the Fet?
Yes .. the leak-down time would be longer.
Does it invert the signal?
Excellent👍👏.
Many thanks
Nice video.
Don't you feel the leakage current via the 10K resistor would be more than the leakage current via the diode?
All semiconductor junctions have some leakage so the diode, mosfet and transistor would all contribute to the total leakage of the finite charge on the capacitor. The blocking diode is common to all bootstrap circuits
so it was mentioned as the leaking source. The other component can differ in different designs. ( the transistor could be replaced with a relay contact) The purpose of the video was to demonstrate how a simple bootstrap circuit functions.
@@0033mer can you please explain this.
When I drive a mosfet(NPN) from the output of a 555timer it turns on for 2 seconds when there is a led connected at the drain of the mosfet, but when I connect a small motor in parallel with the led the time reduces to 800ms.
The LED is clamping the voltage across the motor. Do not connect the LED in parallel with the motor.
cant pwm be used to keep the capacitor charged?
Hello! is it important the frequency that you got in the pwm signal? Because I calculated the RC time constant is 22ms. Does this work well with a high frequency?
This circuit is not meant to run at high frequencies. It is a low parts count circuit to demonstrate mosfet bootstrapping.
@@0033mer and what about if I wanna run it at high frequencies. Should I just change the capacitor value to sth smaller or there's another stuff I should thing about?
For high frequency operation you need to design your circuit around a HVIC.
www.onsemi.cn/pub/Collateral/AN-6076.pdfJP.pdf
So, can we run this circuit with 42 volts?
Its a bit confusing to me. Why is suddenly the source up and the drain down with die P channel fet?
Think of the same circuit but use a bipolar PNP transistor as a high side switch. The emitter would be connected to VCC. Check out the schematic in this video:
ruclips.net/video/-KuEyM1otls/видео.html
Hello, and thanks for the excellent tutorial. Quick question: I've seen you say multiple times in the comments that this setup is not intended for high-frequency applications, and that one should use an integrated circuit for such a case. What is the advantage of using an IC over switching out components for lower-valued ones that could support a higher frequency?
All MOSFETs have Gate/Source capacitance so switching them on/off very fast is not an easy task. There are ICs that will do this properly and keep the MOSFET out of its ohmic region during switching. This keeps the power dissipation down to a minimum and avoids MOSFET failure. To understand gate/source capacitance better check out these videos: ruclips.net/video/zTHzaNdByVA/видео.html
ruclips.net/video/mjIubJeTRyY/видео.html
@@0033mer Thanks for the answer and the resources!
thanks, usefull video 👍
Glad it was helpful!
Ciss is the input capacitance, gate to source and gate to drain (Ciss=Cgd+Cgs)
Yes .. you are correct. Technically Ciss = Cgs + Cgd and is the "input" capacitance as a whole for the Mosfet.
Ciss is the capacitance between the gate and source terminals with the drain shorted to the source.(Cgd in parallel with Cgs)
During switching the Cgd actually amplifies the Cgs so the Ciss is a spec used to compare different Mosfets input capacitances.
This circuit can be using a ignition coil???
Never understood the concept mosfet bootstrapping. This is something I need to know. A while back I experimented with a three half-H bridge to try to drive a BLDC motor.
At least I know how it's possible to build a high side switch with an N-channel mosfet.
If you can build one of these bootstrap circuits, you can essentially build up a 24v rail and then feed into the gates of your h-bridge with smaller transistors.
@@OtherDalfite _"...you can essentially build up a 24v rail and then feed into the gates..."_
That is an excellent idea. Just today I started researching charge pumps. That's another circuit I want to build. Thanks for he tip.
@@3deeguy I'm thinking of that very concept for something I'm working on at work. Hope it helps!
@@OtherDalfite It'll help.
I'm just a hobbyist but I have to say I do expect to run into a few hurdles. A couple years ago I breadboarded a triple half-H bridge using transistors and never could get it to run smoothly. The hfe of each transistor never matched and add to that trying to get complimentary pnp's and npn's to cooperate with each other.
If there is a charge pump IC I'll order mosfet driver IC's too and try to build the circuit that way.
I know I'm talking a lot but now I'm thinking about ordering parts so I can test it out on my oscilloscope.
@@3deeguy hey, was doing some simulations using this idea as a springboard. Currently the best way I have to do this is get a voltage doubler (a bit less complex than bootstrapping) and make a dedicated rail circuit 10v or so above your Vds. Then use an opto-coupler to send the voltage to the gate of the mosfet you're wanting to control. Opto-coupler can be controlled with a 5v PWM from an Arduino.
If anyone has a better way just let me know. Tried normal npn/pnp transistors but couldn't get a configuration that worked.
+2 component solution for high side n channel switching... photovoltatic optocouplers.
logic level control through resistor and optocoupler's led.
Wire the optocoupler's photodiode between mosfet source and gate. The gate voltage is then the source voltage plus what ever the photodiode is outputting.