# Trouble with optocoupler 4n35

I'm trying to ride a full bridge inverter using SPWM to control the MOSFETs, the signal SPWM that I use to control the MOSFETs comes from an NI elvis, but it only provides 5V and I know the fact that the VGS from the MOSFETs must be higher than the VDS and I'm trying to convert at least 20VDC (I want 50VDC but that's another problem with the bridge), so im need a minimum voltage VGS of 20V. I know there are drivers as the IR2110/12 use to control the mosfets but at least for me it doesn't work, so i'm using the optocoupler 4n35 for amplify the input signal of SPWM,it works well while i'm not exceeding the value of 10Vcc, it amplifies the input signal well, but when i'm using more than 10Vcc in the optocoupler the output signal cross for zero and it goes to negatives values and it doesn't work for control the MOSFETs. Can someone give me some alternative for my problem o the 4N35 cannot be used for my aplication?

These are the output signals and the circuit that i'm using.

• What does "10 Vcc" mean? – The Photon May 19 '18 at 5:16
• Voltage between collector and emisor in the optocoupler – David Gonzalez May 19 '18 at 5:20
• Did you have the load attached when measuring the outputs? – The Photon May 19 '18 at 5:23
• Sanity check... Verify the oscilloscope (NI ELVIS) input channel you're using is configured for DC coupled mode and not AC coupled mode. – Jim Fischer May 19 '18 at 6:10
• @JimFischer Please note that the base is open; there is no feedback. This is a perfectly valid way to generate a positive output signal when the LED is on. – CL. May 19 '18 at 6:49

Driving a MOSFET directly from an optocoupler seems a poor idea, especially for PWM operation, because you won't get the switching speed you need.

Dependings on the MOSFET, the voltage and current, the grid intensity needed during switching can range from hundreds of mA to several amperes. This is required to shorten the switching time and thus limit commutation losses.

With your design, during tur-off, the current from the grid is limited to:

$$\frac{20V}{10k} = 2mA$$

This is orders of magnitude too low.

Also consider that, even during turn-on, optocouplers tends to be slowed down if the load resistance (10k here) is too high. I'm afraid the value of the resistor does not fit your needs.

I definitely think that using a MOSFET driver is a good idea. If IR2110 - like drivers do not work for your project, maybe you could consider components like MCP1406/07 ? They can be seen as a powerfull digital buffer/inverter, that can drive a MOSFET.

I think using the opto-coupler to separate the parts of the circuit is wise, but I recommend using a driver IC in between the opto-coupler and the MOSFET itself.

While a MOSFET does not require as high of a current as a thyristor or a BJT to turn on, it needs a small pulse of current to turn on properly and an opto-coupler is not meant to do that, there are many that should work if you provide 20V in their Vcc terminals, to name something IRS2104PBF could work.

So the signal goes through the optocoupler and you just feed the output of your optocoupler to the driver IC.

if you do not need any isolation just skip the opto-coupler and use the driver IC straight up if you can provide the PWM signals cleanly

I do not know enough about your circuit so I cannot recommend a specific thing to be honest.