Ok my bad, I should have asked if you were using the enable pin or an input to provide the PWM?

If you are using the enable pin, then you still have to consider the possibility of other paths discharging the gates.

 

Ok my bad, I should have asked if you were using the enable pin or an input to provide the PWM?

If you are using the enable pin, then you still have to consider the possibility of other paths discharging the gates.

I am using the enable pin to control the PWM. So, if there are even more paths discharging the gate, the situation is even worse, right? And is there a way to prevent that?

 

Good, so that means the voltage goes to 0 with each cycle and not to the opposite voltage.

So as I said in a previous post I can see a possible path thru the circuit, but I can't be sure it exists, because quite frankly my hobby skills are not up to the task, so if someone else such as one of the pros can have a look that would be great.

If the gates being discharged is not the overheating problem, then you are going to have to look somewhere else. (of course you already knew that)

 

Good, so that means the voltage goes to 0 with each cycle and not to the opposite voltage.

So as I said in a previous post I can see a possible path thru the circuit, but I can't be sure it exists, because quite frankly my hobby skills are not up to the task, so if someone else such as one of the pros can have a look that would be great.

If the gates being discharged is not the overheating problem, then you are going to have to look somewhere else. (of course you already knew that)

Thank you very much for your good and objective advice. I'll wait for anyone to chime in with more thoughts on all this, in the meantime I'll consider buying an oscilloscope and I'll do some further tests.

Thanks again

 

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.

The reviews look good - Very much worth a try.

Also have a look at the 2nd hand market - There can be really cheap ones out there that people just want to get rid of.

An oscilloscope is a must have

 

The reviews look good - Very much worth a try.

Also have a look at the 2nd hand market - There can be really cheap ones out there that people just want to get rid of.

An oscilloscope is a must have

Ordered! Once I get it what do you suggest testing? I am thinking this:

1. MOSFETs gate level with and without PWM
2. MOSFETs source/drain with and without PWM

What else?

 

Start there - You basically want to confirm everything is working as it should. It might be worth it to take some time and sketch out what you expect.

One thing that we want to check is the rise time of the gates, noting anything that you didn't expect. When the voltage is rising and current is falling it can require a lot of power (see attached)

We also want to check the coils and see what sort of stress they are putting on the MOSFETs. Also confirm that the Arduino isn't doing any unexpected things (internal pull ups are turned off, ect...)

 

Start there - You basically want to confirm everything is working as it should. It might be worth it to take some time and sketch out what you expect.

One thing that we want to check is the rise time of the gates, noting anything that you didn't expect. When the voltage is rising and current is falling it can require a lot of power (see attached)

We also want to check the coils and see what sort of stress they are putting on the MOSFETs. Also confirm that the Arduino isn't doing any unexpected things (internal pull ups are turned off, ect...)

Sure thing, all that makes sense. Could you please explain the dotted curve in your sketch?

I should receive the scope today, so probably tomorrow I'll have some answers.

A good video about MOSFETs and PWM is this one, even though it is not exactly what I am doing, it gives good thoughts about the whole concept:

I tried to find resources or other videos online specifically about how to handle a PWM signal with MOSFETs triggered by an H-bridge or, anyway, from an external device, but I couldn't find anything. I'll keep searching.

Thanks again

 

That's the power dissipation curve.

 

I don’t understand your statement that the MOSFETS are controlled by the bridge. The MOSFETS should be what makes the bridge. Can you post a schematic of the brdge driving the MOSFETS?

Bob

 

>Could you please explain the dotted curve in your sketch?



It's got to do with multiplying the two linear functions together

I do really like the Great Scott channel, it is great!

 

Using the schematic in post #61 I can more easily illustrate an unwanted discharge path.

Note the pull resistors designed to discharge the gates and do their job when voltage is removed from the gates, but the problem is…this will also happen if voltage is removed from the source.

Normally a PWM design will either be asserted at the gate or the drains.

What I don’t understand is why you aren’t using a bridge on both sides of the coils, but I’m sure you have your reasons.

If your scope shows discharge during the PWM off periods you will need to do a redesign, with a more conventional approach, which shouldn’t be a problem…reversing the polarity of the coils is straight forward, and the PWM can still be done with the enable pins, and you should then move your suppression to across the coils.

And, if you use a separate voltage source to drive the inputs you won’t need all those Zeners and resistors.

 

That's the power dissipation curve.

Thank you!

 

I don’t understand your statement that the MOSFETS are controlled by the bridge. The MOSFETS should be what makes the bridge. Can you post a schematic of the brdge driving the MOSFETS?

Bob

I am sorry, but I wasn't probably clear. Unfortunately, I am unable to upload right now the entire system schematics, but let me try to explain it with different words. What I meant is that the signal is coming from an integrated H-Bridge (L298N) controlled by an Arduino. Such H-Bridge generates the signal with or without PWM and the polarity according to what we want to do with the magnet (attract or repel). The fact is since the whole system is a matrix of magnets (and the final project would include 900 magnets), I can't have 900 separate H-bridges to control every single magnet... instead, I am able to control every single magnet by using relays (controlling the X-axis of the matrix) and MOSFETs (controlling the Y-axis of the matrix) to "redirect" the signal from a single H-Bridge to the magnet I want to fire up (in reality, if you look at the first schematics of I posted in my first post, the number of used H-Bridges is 9 with relative MOSFETs and relay circuits, so that I can control up to 9 magnets at the same time on a matrix of 900 total magnets). The sketch I have uploaded in the posting below can help to better understand the whole system.

So, back to the PWM problem we are discussing, that's generated by the source H-Bridge, but of course, it is handled by both relays and MOSFETs through their source/drain connections. I hope this is clearer.

 

Using the schematic in post #61 I can more easily illustrate an unwanted discharge path.

Note the pull resistors designed to discharge the gates and do their job when voltage is removed from the gates, but the problem is…this will also happen if voltage is removed from the source.

Normally a PWM design will either be asserted at the gate or the drains.

Awesome! That makes sense...

What I don’t understand is why you aren’t using a bridge on both sides of the coils, but I’m sure you have your reasons.

I am using it on both sides! Basically, the path the signal does is this, in both directions:



It is super-simplified of course, but I hope the path is clear. In other words, what we are discussing right now is the block of the two MOSFETs.

If your scope shows discharge during the PWM off periods you will need to do a redesign, with a more conventional approach, which shouldn’t be a problem…reversing the polarity of the coils is straight forward, and the PWM can still be done with the enable pins, and you should then move your suppression to across the coils.

What kind of redesign? Curious... thanks!

And, if you use a separate voltage source to drive the inputs you won’t need all those Zeners and resistors.

Ok, I could have that... but I guess the ground should be common with the rest of the circuit, right?

 

I haven't followed this thread, but I want to make a couple of comments.

a) Look at photovoltaic optocouplers; They are solar cells and FETS, so they GENERATE an independent supply

b) Look at an H-bridge driver that does, FWD, REV, COAST and BRAKE

 

I haven't followed this thread, but I want to make a couple of comments.

a) Look at photovoltaic optocouplers; They are solar cells and FETS, so they GENERATE an independent supply

b) Look at an H-bridge driver that does, FWD, REV, COAST and BRAKE

That's interesting... I guess the photovoltaic optocouplers need light right? If so, that may be a problem...

About your suggested H-bridge driver, how can I search for those?

Thank you for your thoughts.

 

I hope this is clearer.

Yes, I get it now, thanks.

Bob

 

Yes, I get it now, thanks.

Bob

Thank you!

 

Ok, guys, I have finally learned how to use this little scope and took some measurements for your review.

Of course, this scope is pretty cheap and the display is so tiny that it is difficult to see details, but I think it gives a sufficient idea of what's going on (I promise, I'll get a better scope as soon as I can!)

In the meantime, here is what I got:

Gate measurement with no PWM applied:



Gate measurement with 80% PWM applied:



Gate measurement with 20% PWM applied:



Gate measurement with 0.4% PWM applied (the minimum applicable):




Here is the drain without PWM applied:



Drain with 80% PWM applied:




Drain with 0.4% applied (minimum):



And finally, here is the source without PWM:




The source with 80% PWM applied:




And the source with 0.40% PWM applied (the minimum):




It's interesting how much the gate is affected by the PWM and that's probably the cause of heat. It is also interesting to note in the source measurement how much the voltage is "vibrating" even when the signal should be off. It looks like the Arduino is still modulating the signal even when it should be off... I'll have to review the program controlling it because I think I forgot to turn that off as well

Any other measurement do you want me to take? Please, let me know.

I am eager to know your thoughts about all this.

Thanks again.

Fab.