Surprisingly no comments on my circuit proposal on post #61. Could even replace the zeners with fixed resistors as they are only there to fix the gate/ source voltage.

 

Surprisingly no comments on my circuit proposal on post #61. Could even replace the zeners with fixed resistors as they are only there to fix the gate/ source voltage.

Hi Martin. I actually inquired about your proposal on post #63:

https://forum.allaboutcircuits.com/...ener-diode-explode.179806/page-4#post-1641136

And I am intrigued by the simplicity of it. I haven't tested it yet, but I will. Thank you!

 

And I am intrigued by the simplicity of it. I haven't tested it yet, but I will. Thank you!

Sorry, I missed that amongst the other posts.

 

Sorry, I missed that amongst the other posts.

No problem! I am just wondering two things about your circuit:

1. Does it work even if H3-1 and H3-2 switch polarity? I can't visualize that in my mind.

2. If I apply a varistor to protect the MOSFETs as we discussed (by clamping it to the drain and source), will I need to do anything to protect the Zener as well from possible voltage spikes coming from the magnet?

Thanks again!

 

>"by clamping it to the drain and source"

I was saying to put it across the coil as a replacement for a flyback diode. Sorry if I wasn't clear

 

Like this...

 

1. Does it work even if H3-1 and H3-2 switch polarity? I can't visualize that in my mind.

Yes, it should. Looking again at my circuit, I would alter it slightly, see attached image.

2. If I apply a varistor to protect the MOSFETs as we discussed (by clamping it to the drain and source), will I need to do anything to protect the Zener as well from possible voltage spikes coming from the magnet?

I would connect a varistor across the coil L1 in my diagram. Stops the transient there in its tracks.

 

Like this...

Oh, yes, that's actually the best way to do it from what I have read around as well. Unfortunately with my current setup, it is very hard if not impossible to put a varistor that way since the MOSFETs in this circuit control one pole of the magnet, and the other pole is controlled from another circuit where some solid-state relays are located, on a separate board (that's due to the overall design of the project, to be modular and scalable).

On that other board, I have protected the relays with the same kind of varistor placed across the relay's outputs, so I was thinking to do something similar here by putting the varistor across the drain and source of the MOSFETs which would make things easier (if the protection is still guaranteed though!).

Here is what I have done with the relays of the other circuit:



And here is what I was thinking to do with the MOSFETs:



Even though the varistors I'd use would be the ones below that should protect a little bit more if the max needed voltage is 32v and MOSFETs max allowed voltage is 40v:

https://www.mouser.com/ProductDetail/576-V39ZA05P

Your thoughts on that? Do you see anything wrong by clamping the varistors on the MOSFEts that way?

Actually I found the suggestion to do that as explained on this page:

https://www.electronics-tutorials.ws/resistor/varistor.html



So I guess it should work?

 

Yes, it should. Looking again at my circuit, I would alter it slightly, see attached image.

I would connect a varistor across the coil L1 in my diagram. Stops the transient there in its tracks.

Awesome circuit, it looks great to me. I'll make a prototype and test it! Thank you!

Just one question: I guess the value of R1 is still 220 ohms, right?

 

Looks good - Let me know how it goes

 

Looks good - Let me know how it goes

Thank you! I will!

 

Just one question: I guess the value of R1 is still 220 ohms, right?

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.

 

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.

I don't think that's going to be a problem. The switch is in the 50ms range and it is not much repetitive (2-5x every 2 seconds max). But I can test and see. Thanks again!

 

Ok, here is an update. I have soldered the varistors as discussed and replaced the Zener with the ones suggested by @click_here. And here are my results:

Zeners are now ok, no problems with those.

Instead, it seems I still have problems with the N-Channel MOSFETs. They become too hot during operation despite the clamped varistors, and I think that shouldn't happen. I am afraid these MOSFETs are not ok for this circuit.

I am using these N-channel MOSFETs right now:

https://datasheet.lcsc.com/lcsc/1811141141_HL-Haolin-Elec-5N04_C237240.pdf

But I am unable to understand what's their power dissipation, which is not specified on the datasheet. The previous ones I used in a previous prototype, never got hot and never had any problems even without the varistors:

https://www.mouser.com/datasheet/2/196/Infineon_IRFZ44NS_DataSheet_v01_01_EN-1228416.pdf


Are you able to understand where is the problem with these N-Channel MOSFETs compared with the other ones besides the fact they support a much lower voltage and amperage? And why are they becoming so hot? I am afraid they can break easily because of that, and I'll still have shorting issues with them.

Remember: what I need is a max voltage of 35v and a max continuous current of 1-1.5 amps.

Thanks again!

 

Do you have access to an oscilloscope?

It might be worth looking at what is actually happening

 

Do you have access to an oscilloscope?

It might be worth looking at what is actually happening

Unfortunately not. But I am wondering if the small size of those components is playing any role. If you looked at their specs, it looks like they should work just fine, but there must be something else that prevents them to work without overheating compared to the other MOSFETs I used in the previous prototype.

Since I am going to try the circuit suggested by @Martin_R above with breadboard and the previously used components, if that works, I'll remake the whole PCB with the same components so I'll be sure everything works. My choice to use these small MOSFETs was just for the sake of having a small size circuit.

 

Any reason to go from a reliable part to a cheap unknown one?

Good question... that was just for the sake of their small size. By looking at their specs, I thought they would have worked just fine, but I was clearly mistaken. Do you think that fact no power dissipation is shown on their specs is a sign that they may be weak on that front? Or do you see any other difference between those and the previously used ones that could cause them to get that hot?

Thanks again.

 

Another no-name lists PD as 1.2W.

http://www.21yangjie.com/style/pdf/low-voltage-mosfet/YJL05N04A Rev 3.0.pdf

Lower Drain-Source Voltage, higher Junction to Ambient and Drain-source on-resistance.

Thank you, I didn't think to look around for other specs, appreciated!

Ok, so... do you think 1.2W is enough to support 35v at max 1.5 amp? Or that's the problem?

Another detail about my application: often MOSFETs have to deal with PWM signal coming from the H-bridge, which's needed to regulate the power of every single magnet as needed. The fact is, when PWM is applied, the MOSFETs heat up. Otherwise, they stay relatively cool. And I have noticed that P-channel MOSFETs also heat up, not as much as N-channel, but still they heat up with the PWM applied.

Here are the P-channel I am currently using in this circuit:

https://datasheet.lcsc.com/lcsc/1809291523_Wuxi-NCE-Power-Semiconductor-NCE40P05Y_C163738.pdf

And those have a power dissipation of 2W.

For example, if I apply current for 3 seconds at max power (without PWM applied), the MOSFETs remain pretty cool. If I apply the same 3 seconds with PWM to have a different power applied to the magnet, they heat up considerably. What can that be related to?

Thanks again.

 

Long switching times would be my first bet.

 

Long switching times would be my first bet.

So... how would you suggest fixing that? Thank you!