Yes, my mistake. It was late when I drew the circuit.
By the way, your alternating voltage from H3-1 and H3-2, what frequency is it? Just thinking that the MOSFETs are switching at that frequency, and if the gate turn off was too slow it would generate heat in the MOSFET.

Martin, there is just a problem with your design that doesn't fit with my current one: on this circuit, which is complementary to other circuits designed to be scalable in this system, I have no access to the other magnet's pole. In other words, in your last uploaded design, I have no access to the power coming from H3-2. So... unfortunately, I don't see a way to make it work unless to re-design the whole system from scratch, which at the moment, I don't plan to do. That's why I am using the power coming from the 32v rail for handling the MOSFETs gates via the Zeners

Does this make sense to you? Thanks again

 

Because you're applying PWM signal to the MOSFETs they are switching on and off. Is it because they're not switching off fast enough? That 10k gate - source resistance may be causing the transistors to be in the linear mode for a while.

 

My suggestions regarding reducing switching times would start with push/pull MOSFET drivers instead of those relatively slow optos and a pull resistor.

But I'm going to refrain from giving advice here for these reasons...

1. I put my foot in my mouth the last time I tried to help you with this project.
2. You have many people already helping you.
3. You stated the reluctance to redesign your project from scratch.

 

I am sorry, but I wasn't probably clear enough, the PWM is applied at the H-bridge level (via L298Ns), not at the MOSFET level. Does this change the paradigm?

1. I put my foot in my mouth the last time I tried to help you with this project.

I am sorry, but I can't find any posting from you on this thread. Are you referring to something else? Please, let me know.

3. You stated the reluctance to redesign your project from scratch.

About my reluctance to redesign the project from scratch, unfortunately, that's due to the fact I have been working on this current configuration for months. A lot of hard work to find a scalable system that works. So, before starting over, I'll have to evaluate if it is really worth it. But please, any suggestions and ideas are very welcome, I can always find a way to adapt what I have to it.

Thank you again for your thoughts and understanding.

 

Because you're applying PWM signal to the MOSFETs they are switching on and off. Is it because they're not switching off fast enough? That 10k gate - source resistance may be causing the transistors to be in the linear mode for a while.

Thank you for the suggestion. As I wrote above, the PWM is applied at the H-bridge level. Can that be still an issue?

 

Is the PWM on the source side of the MOSFETs?

And JFTR the other time I tried to help was I believe your first post on this project, where you were just trying to figure out the matrix.

 

Looking at post #29 it looks like you have the bridge connected to the source of each MOSFET, each time the PWM turns "off" you could be losing gate charge.

 

Is the PWM on the source side of the MOSFETs?

And JFTR the other time I tried to help was I believe your first post on this project, where you were just trying to figure out the matrix.

Yes, it is on the source side of the MOSFETs.

Oh, yes, of course, now I remember! The big discussion about the magnet matrix Here it is:

https://forum.allaboutcircuits.com/...to-route-power-to-hundreds-of-magnets.171045/


Did something happen that I let you down? I hope not, I can't really recall. If so, my apologies for anything that could have said or wrote. I can't recall much right now, but I remember you and the other guys gave me such an amount of great information and suggestions that could let me continue this project forward to this moment

I appreciated your help, always!

Thank you again.

 

Looking at post #29 it looks like you have the bridge connected to the source of each MOSFET, each time the PWM turns "off" you could be losing gate charge.

Yes, you are right, H-bridge is connected to both sources. Ok, so you say that I could be losing "gate charge." Does that mean that the gate itself could get off in some way? And if so, how to prevent that from happening? Sorry for so many questions

Thanks again!

 

No, it was me that screwed up in that other thread.

Anyway, I see a possible discharge path for the MOSFET gates thru the other MOSFET gate when power is removed from the source by the bridge.

If that is happening then the MOSFET gate must recharge each pulse instead of remaining on.

But I'm not sure of that path, maybe someone else can confirm.

 

No, it was me that screwed up in that other thread.

Anyway, I see a possible discharge path for the MOSFET gates thru the other MOSFET gate when power is removed from the source by the bridge.

If that is happening then the MOSFET gate must recharge each pulse instead of remaining on.

But I'm not sure of that path, maybe someone else can confirm.

Oh well, nobody screws up when good advice is given or attempted to be given!

I appreciate your reply and thoughts, and I think you may be right even though I have no idea how to figure it out and solve it.

Also, an update: I was actually wrongly convinced that the previous design didn't have hot MOSFETs. Today, just for the sake of removing any doubts, I took out the previous prototype made with breadboards and the following MOSFETs (without any varistor protection!):

P-channel:
https://www.mouser.com/ProductDetail/512-FQU8P10TU

N-channel:
https://www.mouser.com/ProductDetail/Infineon-IR/IRFZ44NSTRLPBF?qs=9%2BKlkBgLFf2g3fcs7l0nig==

And despite I can confirm that the P-channel MOSFETs above get a little warm (but nothing to worry about), the N-channel instead, get VERY hot! So... everything seems to behave exactly like the other, tiny MOSFETs I have in this second circuit prototype. The fact is that these ones are larger and dissipate more heat, the other ones are tiny and I guess they can burn more easily.

 

I hate to say it, but to dive any deeper you're going to need an oscilloscope.

You don't know what the voltage spike is when the coil is turned off, so you don't know what the minimum parameters needed for the MOSFETs

Have you considered an IGBT? Those things can handle a beating

 

@fablau

Is the waveform provided by the PWM attached to the source of the MOSFETs…

Pulse (positive or negative)…off (open)…repeat.
Or
Positive…Negative…repeat

 

I hate to say it, but to dive any deeper you're going to need an oscilloscope.

You don't know what the voltage spike is when the coil is turned off, so you don't know what the minimum parameters needed for the MOSFETs

Have you considered an IGBT? Those things can handle a beating

Ouch... an oscilloscope, pretty expensive... but I can consider it. Actually, I know there are good ones nowadays that aren't that expensive... You are right, there's no easy way to understand what's going on. But... for the voltage spike, isn't the varistor limiting that anyway? Just wondering...

As for IGBTs, I never thought about those. I see being pretty expensive and bulky. I'll make some research... thanks for the tip!

 

@fablau

Is the waveform provided by the PWM attached to the source of the MOSFETs…

Pulse (positive or negative)…off (open)…repeat.
Or
Positive…Negative…repeat

The circuit and PWM is controlled by an Arduino. When the PWM is applied, the signal is continuously running on/off/on/off...etc... as a square wave according to this document:

https://www.arduino.cc/en/Tutorial/Foundations/PWM

 

About buying an oscilloscope, do you think this would be sufficient for my research:

https://www.amazon.com/Oscilloscope-Multimeter-LIUMY-Professional-Automatic/dp/B08832JX4H/

Or is it too cheap? That'd be a good price range for me.

 

No, I want to know what the PWM looks like after the bridge.

In the bridge is the Arduino signal connected to a input that provides a push pull output?

Or does the Arduino connect to an input that provides an open drain output?

 

Ok, so I had a look at the bridge datasheet and the output is definitely push pull, so that is probably where your overheating problem is coming from.

Every cycle of the PWM is discharging the gates, and forcing the MOSFETs that you want to stay on to turn on and off (at least partially) and since they have to recharge thru either the opto or the pull resistor this is slow.

Your design needs to keep the MOSFETS controlled by the optos to remain in conduction all during the PWM cycles.

 

Ok, so I had a look at the bridge datasheet and the output is definitely push pull, so that is probably where your overheating problem is coming from.

Every cycle of the PWM is discharging the gates, and forcing the MOSFETs that you want to stay on to turn on and off (at least partially) and since they have to recharge thru either the opto or the pull resistor this is slow.

Your design needs to keep the MOSFETS controlled by the optos to remain in conduction all during the PWM cycles.

Wow, thank you! But please, let me understand what you mean with "Your design needs to keep the MOSFETS controlled by the optos to remain in conduction all during the PWM cycles." I don't understand what you mean. They are controlled by the OPTOs and are kept on by them during the PWM cycles. What am I missing?

Thanks again!

 

You're missing the fact that every cycle of the PWM is discharging the gates, by virtue of being connected to the sources.

Your MOSFETs are kept in conduction by keeping the gates charged, when you reverse the voltage on the source you discharge the gate.