Gate drive optocoupler question.

[SeanHatch]

I'm designing a SMPS where the switch is a MOSFET. I want to isolate the gate of the mosfet from its control circuitry, so I started looking into gate drive optocouplers.

The problem is that I only have a 5V supply available to me, and all of the gate drive optocoupler IC's datasheets suggest using 10 - 15 V for Vcc. What will happen if I power it with only 5?

My guess would be that the switching times are slower. This would be OK since my SMPS isn't that advanced...it only switches at about 4kHz.

Any comments would be appreciated, thanks!

 

[Nigel Goodwin]

Why do you want to isolate the gate?, it's not commonly done.

 

[SeanHatch]

The loads I am driving with the power supply are non linear...IE change alot. For example the load could be 100 ohms now, and change to 10K very quickly The transients caused by theses changes usually shutdown the controller.

Another person working on the project is working on filtering these transients, but I am trying to isolate the circuits.

 

[jpanhalt]

I checked my files and came up with these possibilities:

1) Logic level mosfets -- probably poor choice for high power
2) Gate drive transformers (see: various IR app. notes)
3) Use the optocoupler to drive a complementary emitter follower which drives the gate (see: Zetex ap. note 18)
4) Use the optocoupler to drive a gate-drive IC
5) Switch to IGBT and use optocoupler per Fairchild app. note AN-7511

I wish I could include the links, but unfortunately I saved the documents in PDF in the days before I kept an index of the links. If one or more of the above possibilities appeals to you, let me know and I will look for the link or send the pdf directly to you. John

 

[Styx]

*if* the control really is that susceptable, just isolating the gate-drive aint gonna fix it, your gonna have nice switching currents on the rails.

It might be worth seeing if you can stabalise the controller. Ie more decoupling.

 

[SeanHatch]

*if* the control really is that susceptable, just isolating the gate-drive aint gonna fix it, your gonna have nice switching currents on the rails.

It might be worth seeing if you can stabalise the controller. Ie more decoupling.

On what rails? The circuits would be completely isolated--two separate supplies. Transients on the power supply would have to travel back through the optocoupler (or whatever) to even come in contact with teh rails. Is this possible? Am I not understanding you?

 

[SeanHatch]

i think im just going to buy a couple and see what happens. thanks for your replies.

 

[Styx]

On what rails? The circuits would be completely isolated--two separate supplies. Transients on the power supply would have to travel back through the optocoupler (or whatever) to even come in contact with teh rails. Is this possible? Am I not understanding you?

I'm designing a SMPS where the switch is a MOSFET. I want to isolate the gate of the mosfet from its control circuitry, so I started looking into gate drive optocouplers.

The problem is that I only have a 5V supply available to me, and all of the gate drive optocoupler IC's datasheets suggest using 10 - 15 V for Vcc. What will happen if I power it with only 5?

My guess would be that the switching times are slower. This would be OK since my SMPS isn't that advanced...it only switches at about 4kHz.

Any comments would be appreciated, thanks!

&&

The loads I am driving with the power supply are non linear...IE change alot. For example the load could be 100 ohms now, and change to 10K very quickly The transients caused by theses changes usually shutdown the controller.

Another person working on the project is working on filtering these transients, but I am trying to isolate the circuits.

The fact you are asking to about isolating the gate and that the control is having issues IMPLIES that at the moment there is no isolation (clarify). Thus if there is no isolation the source of the FET must be at the same reference as the controller. Hence my statement about firming up decoupling of the controller

If you plan to use an opto-coupler to drive the FET, are you also planning on isolating the power rail that the FET will be switching?

 

[HS3]

The problem is that I only have a 5V supply available to me, and all of the gate drive optocoupler IC's datasheets suggest using 10 - 15 V for Vcc. What will happen if I power it with only 5?

I do not understand this. Like what optocouplers in particular? Can you post a diagram? I think you may be mistaking things. A general purpose optocoupler just has an LED and a photodiode or phototransistor which will conduct when the LED lights up. There is no requirement for any particular Vcc (and what is Vcc anyway?

The requirement for 15 V is for driving the Gate of the Mosfet and for this you have to check the specs of the MOSFET. It is not only a matter of switching speed but also the MOSFET may not turn on all the way. You need to check the specs but if they say 10~15 volts I would go with 12 minimum unless you know exactly what you are doing.

I recently designed a 230V half bridge and designing the high side control was tricky. There are integrated solutions but I finally designed my own, very simple, solution with an optocoupler. But the gate voltage is about 13 volts which come from a floating capacitor.

Since the current you will need is minimal you can add a SMPS booster to get the ~12 volts you need.

Again, concrete examples and diagrams would help.