Thank you so much for explaining exactly why 95% of the h-bridge circuit diagrams on the interweb don’t work. I’d almost thought it was an impossible circuit.
@@samueladitya303 sure but if you do any tinkering at some point you are going to blow something up. Have a look at strange parts video series where he works on making an atm into a slot machine. He really screws the pooch again and again and that is really not cheap
Thank you for this video ,the explanation is so simple and understandable for a bad english speaker and a electronics noob like me. Thanks you very much, Sir.
Hmmm i suppose you could perhaps drop a zener diode between the gates? That way the P channel gate becomes offset toward the positive rail and the N channel gate goes to the negative rail, the downside is that your mosfets are being driven less meaning you now have to balance shoot-through with RDSon for your new drive voltages
I have have used many half bridge drivers, the chips are only a few bucks and the powerful gate drive (1 to 4 amps I have used) make high speed pwm very efficient as the switching losses are low.
Also if your going to drive a lot of current you might want the high side to be n-channel as well. RDS(on) is much higher in p-fets. There must be a lot of h-bridge drivers out there, so there is no point in trying to do it with discrete components yourself. Great vid
Maybe that's why the original board used relays? The switching delay of the relays (which can be lengthened via capacitors) occurs in step with the switching time of the MOSFETs, thus avoiding that problem. You should be able to achieve your goal with a simple centre off, single pole dual throw toggle switch operating the coils of two relays, maybe salvaged from the failed board. Each relay is wired so that it supplies current to the motor in opposite directions. No need for any electronics at all. But first check to see whether the relay contacts have failed, maybe even welded together, thus causing the loss of voltage to the PIC while the MOSFETs failed. Just guessing, of course, but still worth investigating. Thanks Adam.
I think, with a second transistor of pnp-type, instead of the pullup, and the pullup (470 ohm to 1k for slow switching) between the collectors of the transistors, the problem can be avoided with little effort. At this time, I have no p-Channel MOSFET handy to test, but I think, you should try that configuration on your breadboard. The transistors base must be coupled in a way that there is a voltage interval in which both transistors conduct so that the output of the MOSFETs is floating. That can be achieved by base/emitter resistors of about 470 ohms and two base/base resistors of 4k7 in series. Between these you can apply your 12V rail to rail logic signal. Using optocouplers to isolate the arduino from the powerstage makes no sense as long as both can use a common ground. Again, thank you for cool videos!
Actually it may have worked fine. Once you have a dead short across your mosfets it would pull your control transistor to 0V on both legs thus turning off the mosfets. So your dead short would only last for the fall time of the mosfet which is what a few hundred nano-seconds?
This is the "shoot-through effect". It's necessary to add a dead-time between both gates conducting time. You can do it of course with some parts, but also with dedicated PWM gate drivers.
The worst case scenario with the circuit using a single bipolar NPN transistor and a single pull-up resistor to control an N-channel and a P-Channel MOSFET, is with a relatively high value for the pull-up resistor and the NPN transistor switches off. The NPN collector and therefore the MOSFET gates will see a slow voltage rise due to the gate capacitance charging via the resistor...
I have a long wire sticking out of my wall above my desk...where many TO-220 Mosfet hang there final resting place as a constant reminder of the dumb stuff I have done and learned...make sure you show us how you fixed it..my brother has a caravan with the same problem....
Does it also work for P-P motors and brushless motors? I had done a test with Graetz's schematic but at six rectifier diodes but didn't see any operation. need to try this simple schematic.
Nevertheless though, you've just invented an outstanding circuit that will draw current when the voltage is in between two points! Put a lamp in series with the Mosfets and it'll glow when the voltage is in the middle! An interesting "power good" indicator without using op-amps etc. :)
@@AdamWelchUK It's an interesting idea, and of course you would only need one part of the "H" to make such a circuit. It could probably be "tuned" by changing the resistance value and/or adding some Zenner diodes somewhere in the line between the Mosfets.
Interesting video. Will look forward to your final solution. I'm starting to dabble in solar, battery technology, and also have a caravan that I use to vacation in.
You drive them by four pins of arduino. Giving about 50 microseconds gap between switching off and on. First switch of the mosfet that is on wait for about 50 microseconds then switch on mosfet of other leg. Just few days ago I did it. No need of any transistors even. It works fine. To increase the current you can have parellel n and p channel mosfets and drive by same arduino pins.
That configuration could work fine if your supply voltage is low enough that at no point can both FETs be in saturation or ohmic region (fancy words for linear region and "on"). If the threshold voltage is about 4V, you could get away with 6V to 7V supply and the shoot-through would be minimal. But of course it is not desirable to have such a narrow supply voltage for operation. 5V might work with the right FETs, but most normal power FETs need a bit more than that to achieve the Rds(on) banner spec.
I connect the the l298n to a stepper motor nema17 my DC power supply shows no current being drawn when I send a signal to move the stepper one revolution and then stops I notice it's still continuously drawing max amps and it's in CC mode an till I reset the Arduino then it goes back to normal again it's like the pins 8,9,10,11 are causing it to still run some how also after about 30mins the rc522 RFID scanner stops working any ideas what this could be?
On e-bikes or bldc controllers they seems like used lots of cascaded transistors stabilisimg resistors optocouplers for fixing the issue. Altough i didnt track all details but that circuit is used on all bldc motors controllerstoo ie electric weichles in general. Well they use 3 phase h bridge ...
Nice explanation... Does the motor driver have any proportional control or is it just left or right motor on or off? You obviously need to introduce a dead zone and delay, which you could do using the arduino and another 2 ports, but you could also add some logic to get that 'for free' - maybe a clock with a resettable counter or shift register that you can tap with some gates to get trigger signals at the right times.
2 more ports and code to provide a delay that is slightly greater than the slowest MOSFET's discharge time. Sounds like something Adam could do over one cup of tea ;-)
@Undefined Lastname Both have advantages - I like the ability to offload tasks to hardware that just works, but it's also good to be able to have software control of all parameters and build an optimized model.
Its not easy to suffer massive blow outs when you dealing with mains rectified H bridge circuits and get your timing even a tiny bit wrong and you have a massive blow out of high voltage expensive Mosfets.
Can someone please explain what the issue is here with the transistor driving the H-bridge? I understand that if you drive a transistor with a variable voltage input then you will get the scenario that he is talking about here where only a portion of the voltage is dropping across the transistor thus varying the voltage at the gate of the mosfets. But if you are driving the transistors fully on and fully off then you don't have to worry about the states in-between. Driving the transistor fully on will put the transistor in the saturation region and dropping around 0.2V from collector to emitter, effectively 0V or ground, and the remainder of the voltage will drop across the resistor. Turning the transistor fully off will have no current flowing through the transistor and the gates of the mosfets will be at the supply voltage.
the difference is you will get a few micro seconds of shoot through or high loss via dead time what adam is the almost dead short do to no dead time hope this helps
Couldn’t you just use 2 x 3 pole relays. 1 pole for power and the other 2 for forward. And the second relay for reverse. So now relay will work if the other one is on.
I used the H-bridge shown in the following picture: prosje.be/Projects/H-brugVI.png (from the webpage prosje.be/Projects/BelastingSchakelen.html). It is in dutch, but I try to translate it. As you want to use the left P-mosfet in conjunction with the right N-Mosfet, you can switch the P-mosfet on by pulling its matching N mofset low and keeping it conducting trough a capacitor. The N-mosfet is the only active switching component and therefore got a mosfet drivers to improve the switching characteristics. The only drawback is that you may never switch input A and B on at the same time. Hopefully it helps you further ;)
ps. keep switching fast as slow as possible ! ie to fast ringing on gate to slow leaner regain destruction more in depth and great calculators that help kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/ great stuff good luck Dan
You should probably breadboard your original setup and examine the current flow with the transistors fully on and off, instead of using a variable power supply on the gates of the mosfet because this demonstration is not an accurate depiction of your original circuit. I'm just saying.
Take a look at the circuit in th-cam.com/video/dbwZgP9gIbA/w-d-xo.html. When the inputs (D1 D2) are both low all 4 mosfets are off. When one input goes high. Only the desired mosfets are turned on. All you have to do is make sure you turn off the one input, delay a smig, then turn the other input. I think.
MrBrymstond, go Ohm! I'm saying this with some Reluctance: The Power of your words is meeting with Resistance here, because it makes you sound like a Hooray Henry from a Coulomb a Short Circuit away from blowing a Fuse. It makes me want to Choke, even though Anode what you meant! Under the Current climate, do you think it would be wise to make a Volt for the door. A? C? To Siemen penning jokes along that kind of Line is rather down to Earth of you, but writing as an Unbiased Neutral, the Power Factor here is less than ideal. While this Sparks some more funny comments from you, I'll go to Ground Faraday, and build a PCB with my Dielectric Soldering Ion, and Flux, while you Collector thought or two. No more Base comments from this Emitter! DC Watt I mean? To Condenser word or two of that , thanks for the laugh!
An easy fix (that doesn't require new PCBs) is to add a 5th mosfet to VCC to disconnect power when the mosfets are changing state (direction). Fixing it properly requires creating "dead time" where no mosfets should be switched on (in reality mosfet gates behave like capacitors). Overly technical on the topic. hackaday.io/project/3176-gator-quad/log/11741-pwm-control-and-dead-time-insertion ps: solder a EMI supression cap to the end of that motor if you plan to use it in projects.
if you don't need speed control of your motor, you can use simple relays. Even with an h bridge relay, you can add just one mosfet in serie to pwm and the adjust the motor speed. With dpdt relay, you just need one pin : 4.bp.blogspot.com/-z_QGf7IYPWM/T8cOTopVn8I/AAAAAAAAAN0/TqP4AR_qgHE/s1600/dpdt.JPG. I've use relays for my for my gate opener (2 motors) and it works fine since about five years. If you want more details, you can contact me, i would be glad to help.
Thank you so much for explaining exactly why 95% of the h-bridge circuit diagrams on the interweb don’t work. I’d almost thought it was an impossible circuit.
It takes a man to admit when he is wrong. You're a gentleman and a scholar. 👍
You’re too kind.
If you don't screw up you're not learning. And we learn with you
@@samueladitya303 sure but if you do any tinkering at some point you are going to blow something up. Have a look at strange parts video series where he works on making an atm into a slot machine. He really screws the pooch again and again and that is really not cheap
Thank you for this video ,the explanation is so simple and understandable for a bad english speaker and a electronics noob like me. Thanks you very much, Sir.
Lesson learned. That's all that matters. You'll know next time.
Don't be disheartened.
Thanks. You’re right.
Hmmm i suppose you could perhaps drop a zener diode between the gates? That way the P channel gate becomes offset toward the positive rail and the N channel gate goes to the negative rail, the downside is that your mosfets are being driven less meaning you now have to balance shoot-through with RDSon for your new drive voltages
I have have used many half bridge drivers, the chips are only a few bucks and the powerful gate drive (1 to 4 amps I have used) make high speed pwm very efficient as the switching losses are low.
Also if your going to drive a lot of current you might want the high side to be n-channel as well. RDS(on) is much higher in p-fets. There must be a lot of h-bridge drivers out there, so there is no point in trying to do it with discrete components yourself. Great vid
Maybe that's why the original board used relays? The switching delay of the relays (which can be lengthened via capacitors) occurs in step with the switching time of the MOSFETs, thus avoiding that problem.
You should be able to achieve your goal with a simple centre off, single pole dual throw toggle switch operating the coils of two relays, maybe salvaged from the failed board. Each relay is wired so that it supplies current to the motor in opposite directions. No need for any electronics at all.
But first check to see whether the relay contacts have failed, maybe even welded together, thus causing the loss of voltage to the PIC while the MOSFETs failed. Just guessing, of course, but still worth investigating.
Thanks Adam.
I think, with a second transistor of pnp-type, instead of the pullup, and the pullup (470 ohm to 1k for slow switching) between the collectors of the transistors, the problem can be avoided with little effort. At this time, I have no p-Channel MOSFET handy to test, but I think, you should try that configuration on your breadboard. The transistors base must be coupled in a way that there is a voltage interval in which both transistors conduct so that the output of the MOSFETs is floating. That can be achieved by base/emitter resistors of about 470 ohms and two base/base resistors of 4k7 in series. Between these you can apply your 12V rail to rail logic signal.
Using optocouplers to isolate the arduino from the powerstage makes no sense as long as both can use a common ground.
Again, thank you for cool videos!
Actually it may have worked fine. Once you have a dead short across your mosfets it would pull your control transistor to 0V on both legs thus turning off the mosfets. So your dead short would only last for the fall time of the mosfet which is what a few hundred nano-seconds?
This is the "shoot-through effect". It's necessary to add a dead-time between both gates conducting time. You can do it of course with some parts, but also with dedicated PWM gate drivers.
The worst case scenario with the circuit using a single bipolar NPN transistor and a single pull-up resistor to control an N-channel and a P-Channel MOSFET, is with a relatively high value for the pull-up resistor and the NPN transistor switches off. The NPN collector and therefore the MOSFET gates will see a slow voltage rise due to the gate capacitance charging via the resistor...
I have a long wire sticking out of my wall above my desk...where many TO-220 Mosfet hang there final resting place as a constant reminder of the dumb stuff I have done and learned...make sure you show us how you fixed it..my brother has a caravan with the same problem....
I have a RIP bag with more in than I would like to admit to🤔 .
I think a way around this is to drive them with Schmitt triggers instead of transistors, that way the "shoot through" zone is bypassed.
Does it also work for P-P motors and brushless motors? I had done a test with Graetz's schematic but at six rectifier diodes but didn't see any operation. need to try this simple schematic.
How do you correct this? The one I'm trying to move, is not moving FML
Nevertheless though, you've just invented an outstanding circuit that will draw current when the voltage is in between two points! Put a lamp in series with the Mosfets and it'll glow when the voltage is in the middle! An interesting "power good" indicator without using op-amps etc. :)
Every cloud eh? I great way to look at it. Thanks Xan.
@@AdamWelchUK It's an interesting idea, and of course you would only need one part of the "H" to make such a circuit. It could probably be "tuned" by changing the resistance value and/or adding some Zenner diodes somewhere in the line between the Mosfets.
Interesting video. Will look forward to your final solution. I'm starting to dabble in solar, battery technology, and also have a caravan that I use to vacation in.
You drive them by four pins of arduino. Giving about 50 microseconds gap between switching off and on. First switch of the mosfet that is on wait for about 50 microseconds then switch on mosfet of other leg. Just few days ago I did it. No need of any transistors even. It works fine. To increase the current you can have parellel n and p channel mosfets and drive by same arduino pins.
That configuration could work fine if your supply voltage is low enough that at no point can both FETs be in saturation or ohmic region (fancy words for linear region and "on"). If the threshold voltage is about 4V, you could get away with 6V to 7V supply and the shoot-through would be minimal. But of course it is not desirable to have such a narrow supply voltage for operation. 5V might work with the right FETs, but most normal power FETs need a bit more than that to achieve the Rds(on) banner spec.
I connect the the l298n to a stepper motor nema17 my DC power supply shows no current being drawn when I send a signal to move the stepper one revolution and then stops I notice it's still continuously drawing max amps and it's in CC mode an till I reset the Arduino then it goes back to normal again it's like the pins 8,9,10,11 are causing it to still run some how also after about 30mins the rc522 RFID scanner stops working any ideas what this could be?
On e-bikes or bldc controllers they seems like used lots of cascaded transistors stabilisimg resistors optocouplers for fixing the issue. Altough i didnt track all details but that circuit is used on all bldc motors controllerstoo ie electric weichles in general. Well they use 3 phase h bridge ...
why all the h ridges i have seen on solar inverters up to 3000 watts have 4 equal IGBTs?
How do I do it if all of my Mosfets are N channel?
Nice explanation... Does the motor driver have any proportional control or is it just left or right motor on or off?
You obviously need to introduce a dead zone and delay, which you could do using the arduino and another 2 ports, but you could also add some logic to get that 'for free' - maybe a clock with a resettable counter or shift register that you can tap with some gates to get trigger signals at the right times.
2 more ports and code to provide a delay that is slightly greater than the slowest MOSFET's discharge time. Sounds like something Adam could do over one cup of tea ;-)
@Undefined Lastname Both have advantages - I like the ability to offload tasks to hardware that just works, but it's also good to be able to have software control of all parameters and build an optimized model.
Very good clip
I was hoping you were going to show the "quite exciting" lol.
Its not easy to suffer massive blow outs when you dealing with mains rectified H bridge circuits and get your timing even a tiny bit wrong and you have a massive blow out of high voltage expensive Mosfets.
Can someone please explain what the issue is here with the transistor driving the H-bridge? I understand that if you drive a transistor with a variable voltage input then you will get the scenario that he is talking about here where only a portion of the voltage is dropping across the transistor thus varying the voltage at the gate of the mosfets. But if you are driving the transistors fully on and fully off then you don't have to worry about the states in-between. Driving the transistor fully on will put the transistor in the saturation region and dropping around 0.2V from collector to emitter, effectively 0V or ground, and the remainder of the voltage will drop across the resistor. Turning the transistor fully off will have no current flowing through the transistor and the gates of the mosfets will be at the supply voltage.
the difference is you will get a few micro seconds of shoot through or high loss via dead time what adam is the almost dead short do to no dead time hope this helps
hi can you swap the positions of the mosfet's or will that not solve the problem?
Couldn’t you just use 2 x 3 pole relays. 1 pole for power and the other 2 for forward. And the second relay for reverse. So now relay will work if the other one is on.
I used the H-bridge shown in the following picture: prosje.be/Projects/H-brugVI.png
(from the webpage prosje.be/Projects/BelastingSchakelen.html). It is in dutch, but I try to translate it.
As you want to use the left P-mosfet in conjunction with the right N-Mosfet, you can switch the P-mosfet on by pulling its matching N mofset low and keeping it conducting trough a capacitor. The N-mosfet is the only active switching component and therefore got a mosfet drivers to improve the switching characteristics. The only drawback is that you may never switch input A and B on at the same time.
Hopefully it helps you further ;)
What is the purpose of capacitors C1 and C2 and diodes D5 and D6 in that circuit?
Useful video 👍
Search for L9110S on eBay. Dual chip boards for $1 or 15A 36V power versions for US$10
Peter Kay BTS7960B is another option. Have used it previously to drive wheelchair motors in my combat robots.
@@TornTech1 Another nice option with hi side driver as well and >40A but limited to
How can I simulate DRF1400 circuit if anyone have idea please share with me that will be highly appreciable thanks
ps. keep switching fast as slow as possible ! ie to fast ringing on gate to slow leaner regain destruction more in depth and great calculators that help kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/igbts/ great stuff good luck Dan
I want it to run at full speed without any pwm signal !!!
You should probably breadboard your original setup and examine the current flow with the transistors fully on and off, instead of using a variable power supply on the gates of the mosfet because this demonstration is not an accurate depiction of your original circuit. I'm just saying.
شكرا
Take a look at the circuit in th-cam.com/video/dbwZgP9gIbA/w-d-xo.html. When the inputs (D1 D2) are both low all 4 mosfets are off. When one input goes high. Only the desired mosfets are turned on. All you have to do is make sure you turn off the one input, delay a smig, then turn the other input. I think.
In SHORT, How dare you commit mosfetery?! Complete and utter Blastofme :-)
MrBrymstond, go Ohm! I'm saying this with some Reluctance: The Power of your words is meeting with Resistance here, because it makes you sound like a Hooray Henry from a Coulomb a Short Circuit away from blowing a Fuse. It makes me want to Choke, even though Anode what you meant! Under the Current climate, do you think it would be wise to make a Volt for the door. A? C? To Siemen penning jokes along that kind of Line is rather down to Earth of you, but writing as an Unbiased Neutral, the Power Factor here is less than ideal. While this Sparks some more funny comments from you, I'll go to Ground Faraday, and build a PCB with my Dielectric Soldering Ion, and Flux, while you Collector thought or two. No more Base comments from this Emitter! DC Watt I mean?
To Condenser word or two of that , thanks for the laugh!
@@RWBHere Ahahahahahaha
An easy fix (that doesn't require new PCBs) is to add a 5th mosfet to VCC to disconnect power when the mosfets are changing state (direction).
Fixing it properly requires creating "dead time" where no mosfets should be switched on (in reality mosfet gates behave like capacitors).
Overly technical on the topic.
hackaday.io/project/3176-gator-quad/log/11741-pwm-control-and-dead-time-insertion
ps: solder a EMI supression cap to the end of that motor if you plan to use it in projects.
You are love
if you don't need speed control of your motor, you can use simple relays. Even with an h bridge relay, you can add just one mosfet in serie to pwm and the adjust the motor speed. With dpdt relay, you just need one pin : 4.bp.blogspot.com/-z_QGf7IYPWM/T8cOTopVn8I/AAAAAAAAAN0/TqP4AR_qgHE/s1600/dpdt.JPG. I've use relays for my for my gate opener (2 motors) and it works fine since about five years. If you want more details, you can contact me, i would be glad to help.
I hate p channel mosfet to difficult to use
MOSFET p chanel
It’s called shoot-through. And it’s pronounced “et-cetera”. No K.