Oh man, the rabbit hole just keeps getting deeper and deeper. This is far above what we covered 20 years ago in the electronics classes for avionics nav and com repair. I guess designing a circuit is far more complex than troubleshooting and repairing one with like components. Great video and I’ll have to watch this a few more times along with others to understand this a little better.
This is great. I'd been wondering what considerations should go into sizing a gate resistor for a while and hadn't found anything terribly useful. One thought I had, though... You never really commented much on how the current you derive relates to the circuit in question. You kinda did without really bringing much attention to it when you increased the voltage to 19 volts and swapped in the 100 ohm resistor, but didn't really go into much detail or explanation. I'm thinking if the turn-on time is mostly defined by charging up the Miller capacitance, then the voltage driving that charge current would be the gate-drain voltage, wouldn't it? It seemed like you were just taking the voltage on the drain and dividing by the desired current to get a resistance, but shouldn't we calculate Vd - Vgs(th) and divide that by our required charge current to get the appropriate resistor? Isn't that the voltage across the Miller capacitance? For higher voltages, this might not make much difference, but for lower voltage circuitry, this seems like it could make a big difference. Or am I missing something?
Hi brettski74, I’m not sure if I follow you. In the example given, the bulk of the charge was on the gate to source capacitance since it was larger than the gate to drain capacitance. However, one gets a delay charging the gate to drain capacitance, shown with the Millar Plateau, due to the rapidly decreasing voltage on the drain. I see the majority of the turn on time was on the gate to source capacitance, not the Millar Plateau. You wrote “I'm thinking if the turn-on time is mostly defined by charging up the Miller capacitance.” I don’t see that as the case. Am I mad in thinking the gate and gate to source capacitance took longer to charge than the gate to drain capacitance? I must point out that I took the gate to source charge only and intended to look at the gate to drain charge later. As with many people making videos, mistakes are made whilst doing stuff on the fly without a script. I left the video ‘as is’ and figured it was ok for the basics.
@@JohnB-2021 Let's start with - I don't know the correct answer here. What I was trying to understand from your video is that we can calculate the charge current that we want in order to achieve a given turn-on time. That's straightforward enough, but if we want to go from that current to a resistance to connect to the gate, we would need to know what voltage is driving that current. Then we can calculate the resistance using ohm's law. So the question at that point is what voltage is across our gate resistor to drive the desired current. You did mention at about 10:40 in the video that you had 19 volts on the drain and proceeded to divide 19 volts by 100 ohms to get 190mA of current. I'm not seeing how this relates to the resistor or the desired charging current, though. I initially figured you mentioned drain voltage because the Miller capacitance is across the gate-drain, but on reflection, that doesn't seem like it should matter. We're talking about ohm's law and the calculation you did there implies that there was 19 volts across the gate resistor, but was there? Looking at your oscilloscope trace, it looks like the gate voltage tops out a little below 15V and this is presumably the output voltage of the gate driver. I'm also assuming that since the plateau region is where the MOSFET first starts to conduct, this is the threshold voltage and looks to be around 5V. So wouldn't the voltage across the gate resistance be around 10V? I don't doubt your results, but I'm trying to understand why it works given the numbers and images presented.
@@brettski74 I see your point. I had the power supply connected to the driver, it was showing gate voltage peak to peak and drain voltage on the other lead. Changing the voltage on the drain would influence the turn on time. That same supply was connected to the driver. I see some points in the video are not as clear as I would like. Thank you for your comments, I do appreciate them.
You guys are both right. The gate charge determines the total time of charging the gate to your driver voltage (typically 15V) and the duration of the Miller plateau depends on the Vds voltage, since it is charging the Miller capacitance Vdg (often called Crss). The data sheet typically specs both the gate charge and the Cgs and Crss capacitances since both are important in determining your gate resistor and the switch on and off times of your MOSFET. If you are interested in more details on how to set your switch on/off times and the gate resistor(s) that go with that, I highly recommend the video by Prof. Sam Ben-Yaakov called "MOSFETs’ gate resistors" here on youtube. That video also addresses why you should switch on slow and switch off fast, and thus typically provide two different gate resistors for each direction.
Thanks for the super informative video. Question: Could a potentiometer be placed in lieu of the gate resistor then watch the scope as the pot is manipulated? When the point of ringing is reached, measure the resistance of the potentiometer. Or would the characteristics of the pot invalidate the test? The calculations work well, I was just curious if what I described would work or not. Thanks the videos
Great video John. I would like to ask a question thou I am building an ev and designing a controller. I am currently using cm600ha-24h igbt bricks two in parallel. Have 20ohm resistors on each gate and 1k pull down resistors on each . I run controller it works great until I increase the load to 50 amps and then they destroy themselves
Wow, huge transistors. What is the load? What is the switching frequency? What is your supply voltage? 3 possibilities that spring to mind. You are exceeding the max power rating of 4100 watts each. Or if switching an inductive load, you are generating an emf voltage (spike) which may exceed the device voltage rating. Or, the gate voltage is being exceeded during the switching phase due to stray inductance etc, you could clamp the gates with a zener diode.
@@JohnB-2021 thankyou for your reply. I have a 11 inch dc sereis wound motor it is capable of handling 144vdc at 600amp max. I am only using 60vdc at 200amps with a peak of 800 amps using lithium prismatic 21v pack ×3 .I have capacitors for dc bus rated at 450vdc 500uf ×3. I also am using 2× 1200vd 1uf snubber capacitors and the spike is raising to 65vdv and the scope looks like 1khz frequency. So I am very confused I have destroyed 4 controller's I have built. So I am asking for any help as you no they are expensive.
@@craigivas4037 Are you blowing the controllers and transistors or just the controllers? 60 volts x 200 amps is 12 thousand watts. You need to measure the voltage across the transistors when the are switched on. It sounds like one of the possibilities above, exceeding the 4100 watts per device. Also, what scope do you have? Only a fast scope will show the true inductive spikes.
@@JohnB-2021 good morning John yes the scope is a Tektronix task 210. I have a 200 amp shunt and it reads up 40 amps all day long as soon as I push further like 50 amps it blows the igbt bricks.. the driver still works fine. I did notice that the 1k pulldown resistors were burnt out as well.
If the device that switches on the fet can’t switch it off, put a resistor from the gate to source. If you want to slow the switch on time or you want to limit the current from an ic, put a gate resistor between ic and fet. If you encounter gate loop ringing, increase the size of the gate resistor.
E quando a cituação é a seguinte : - NÃO SE PODE MUDAR A TENSÃO DE DRENO ! pra começar , e ela é a tensão de uma bateria de 12v e nada mais além disso. como se cálcular os resistores corretos para o MOSFET IRF640 ? Alguém saberia me informar isso com vídeo por favor ?
Oh man, the rabbit hole just keeps getting deeper and deeper. This is far above what we covered 20 years ago in the electronics classes for avionics nav and com repair. I guess designing a circuit is far more complex than troubleshooting and repairing one with like components. Great video and I’ll have to watch this a few more times along with others to understand this a little better.
Thank you for your interesting comments 😃
Indeed i was waiting for this marvelous knowledge n explanation .... Cheers
Thank you😁
@@JohnB-2021 my pleasure, plz make electronics n electrical videos
Really nice explanation
I wish that universities could also explain electronics with design exaples and shots from osciloscopes.
Thank you Vasquez
Excellent overview.
Thank you kind sir.
I have just uploaded 'gate loop ringing' Take 2
You may like to have a look.😁
This is great. I'd been wondering what considerations should go into sizing a gate resistor for a while and hadn't found anything terribly useful. One thought I had, though... You never really commented much on how the current you derive relates to the circuit in question. You kinda did without really bringing much attention to it when you increased the voltage to 19 volts and swapped in the 100 ohm resistor, but didn't really go into much detail or explanation. I'm thinking if the turn-on time is mostly defined by charging up the Miller capacitance, then the voltage driving that charge current would be the gate-drain voltage, wouldn't it? It seemed like you were just taking the voltage on the drain and dividing by the desired current to get a resistance, but shouldn't we calculate Vd - Vgs(th) and divide that by our required charge current to get the appropriate resistor? Isn't that the voltage across the Miller capacitance? For higher voltages, this might not make much difference, but for lower voltage circuitry, this seems like it could make a big difference. Or am I missing something?
Hi brettski74,
I’m not sure if I follow you.
In the example given, the bulk of the charge was on the gate to source capacitance since it was larger than the gate to drain capacitance.
However, one gets a delay charging the gate to drain capacitance, shown with the Millar Plateau, due to the rapidly decreasing voltage on the drain.
I see the majority of the turn on time was on the gate to source capacitance, not the Millar Plateau.
You wrote “I'm thinking if the turn-on time is mostly defined by charging up the Miller capacitance.”
I don’t see that as the case. Am I mad in thinking the gate and gate to source capacitance took longer to charge than the gate to drain capacitance?
I must point out that I took the gate to source charge only and intended to look at the gate to drain charge later. As with many people making videos, mistakes are made whilst doing stuff on the fly without a script. I left the video ‘as is’ and figured it was ok for the basics.
@@JohnB-2021 Let's start with - I don't know the correct answer here. What I was trying to understand from your video is that we can calculate the charge current that we want in order to achieve a given turn-on time. That's straightforward enough, but if we want to go from that current to a resistance to connect to the gate, we would need to know what voltage is driving that current. Then we can calculate the resistance using ohm's law. So the question at that point is what voltage is across our gate resistor to drive the desired current.
You did mention at about 10:40 in the video that you had 19 volts on the drain and proceeded to divide 19 volts by 100 ohms to get 190mA of current. I'm not seeing how this relates to the resistor or the desired charging current, though. I initially figured you mentioned drain voltage because the Miller capacitance is across the gate-drain, but on reflection, that doesn't seem like it should matter. We're talking about ohm's law and the calculation you did there implies that there was 19 volts across the gate resistor, but was there? Looking at your oscilloscope trace, it looks like the gate voltage tops out a little below 15V and this is presumably the output voltage of the gate driver. I'm also assuming that since the plateau region is where the MOSFET first starts to conduct, this is the threshold voltage and looks to be around 5V. So wouldn't the voltage across the gate resistance be around 10V? I don't doubt your results, but I'm trying to understand why it works given the numbers and images presented.
@@brettski74
I see your point.
I had the power supply connected to the driver, it was showing gate voltage peak to peak and drain voltage on the other lead.
Changing the voltage on the drain would influence the turn on time.
That same supply was connected to the driver.
I see some points in the video are not as clear as I would like.
Thank you for your comments, I do appreciate them.
You guys are both right.
The gate charge determines the total time of charging the gate to your driver voltage (typically 15V) and the duration of the Miller plateau depends on the Vds voltage, since it is charging the Miller capacitance Vdg (often called Crss). The data sheet typically specs both the gate charge and the Cgs and Crss capacitances since both are important in determining your gate resistor and the switch on and off times of your MOSFET.
If you are interested in more details on how to set your switch on/off times and the gate resistor(s) that go with that, I highly recommend the video by Prof. Sam Ben-Yaakov called "MOSFETs’ gate resistors" here on youtube. That video also addresses why you should switch on slow and switch off fast, and thus typically provide two different gate resistors for each direction.
Thanks for the super informative video. Question: Could a potentiometer be placed in lieu of the gate resistor then watch the scope as the pot is manipulated? When the point of ringing is reached, measure the resistance of the potentiometer. Or would the characteristics of the pot invalidate the test? The calculations work well, I was just curious if what I described would work or not. Thanks the videos
Yes you could use a pot.
You should not experience ringing unless you have high inductance somewhere. Thanks for your comments.😀
Great video John. I would like to ask a question thou I am building an ev and designing a controller. I am currently using cm600ha-24h igbt bricks two in parallel. Have 20ohm resistors on each gate and 1k pull down resistors on each . I run controller it works great until I increase the load to 50 amps and then they destroy themselves
Wow, huge transistors.
What is the load?
What is the switching frequency?
What is your supply voltage?
3 possibilities that spring to mind.
You are exceeding the max power rating of 4100 watts each.
Or if switching an inductive load, you are generating an emf voltage (spike) which may exceed the device voltage rating.
Or, the gate voltage is being exceeded during the switching phase due to stray inductance etc, you could clamp the gates with a zener diode.
@@JohnB-2021 thankyou for your reply. I have a 11 inch dc sereis wound motor it is capable of handling 144vdc at 600amp max. I am only using 60vdc at 200amps with a peak of 800 amps using lithium prismatic 21v pack ×3 .I have capacitors for dc bus rated at 450vdc 500uf ×3. I also am using 2× 1200vd 1uf snubber capacitors and the spike is raising to 65vdv and the scope looks like 1khz frequency. So I am very confused I have destroyed 4 controller's I have built. So I am asking for any help as you no they are expensive.
@@craigivas4037 Are you blowing the controllers and transistors or just the controllers?
60 volts x 200 amps is 12 thousand watts.
You need to measure the voltage across the transistors when the are switched on.
It sounds like one of the possibilities above, exceeding the 4100 watts per device.
Also, what scope do you have?
Only a fast scope will show the true inductive spikes.
@@JohnB-2021 good morning John yes the scope is a Tektronix task 210. I have a 200 amp shunt and it reads up 40 amps all day long as soon as I push further like 50 amps it blows the igbt bricks.. the driver still works fine. I did notice that the 1k pulldown resistors were burnt out as well.
@@craigivas4037 You need to measure the voltage across the transistors.
Do they get hot?
Can you tel, when to use a gate to source resistor and when to use a resistor berween gate and IC ?
If the device that switches on the fet can’t switch it off, put a resistor from the gate to source.
If you want to slow the switch on time or you want to limit the current from an ic, put a gate resistor between ic and fet.
If you encounter gate loop ringing, increase the size of the gate resistor.
E quando a cituação é a seguinte :
- NÃO SE PODE MUDAR A TENSÃO DE DRENO ! pra começar , e ela é a tensão de uma bateria de 12v e nada mais além disso. como se cálcular os resistores corretos para o MOSFET IRF640 ?
Alguém saberia me informar isso com vídeo por favor ?
que velocidade você precisa?
@@JohnB-2021 60 ciclos por segundo .
@@enfoteceletronica2818 Just stick a 50 ohm in
I thought there was 1000ma in 1 Amp. Why do you call 0.1A 100ma?
I thought there was too, has it changed?
Sorry you are right. I am getting forgetful ay my old age.
@Jerry Could it be because .1 = 1/10 and 1/10 of 1000 = 100. ??
@@bengunn3698 Thank you Ben😄
I thought I was going mad.