What is the reason for the slow switching speed of mos?

[crystal2080]

The single-chip microcomputer drives the IO port, and the reaction time of the top and bottom edges of the optocoupler 3 pins is basically within 1ms, but the top and bottom edges of the mos tube switching waveform are basically about 40ms. The drain in the figure has been connected to the +KM and 10K resistors through the wiring terminals. The switching response time Why is it so long?

[R.G.]

I can only guess, but I think you're not driving the optocoupler LED with enough current.
Your microcontroller can only pull up its output pins to a little less than 3.3V. The EL817 LED needs 1.2V typical to conduct. That leaves only 2.1V through the 1K resistor. So the current into the EL817 LED is about 2.1ma. This is a small current to drive a MOSFET gate.
Try reducing the 1K resistor to perhaps 220 ohms, or smaller. Most microcontrollers can drive as much as 20ma from their pins, and the PIC family has internal current limiters so they can driver LEDs directly with no resistor. I think a lower resistor might speed things up.

[antonis]

I can also only guess so I'll add Mosfet inter-electrode capacitance as part of your delay issue..

[anotherjim]

I agree with RG, the LED could be driven harder. A resistor as low as 100R will give 20mA so 220R seems a reasonable minimum.
There is 10k to discharge the MOSFET gate capacitance (which is a slowdown). If the opto output is driven well, 1k can work there and it could be a minimum 220R too.

[Rob Strand]

The single-chip microcomputer drives the IO port, and the reaction time of the top and bottom edges of the optocoupler 3 pins is basically within 1ms, but the top and bottom edges of the mos tube switching waveform are basically about 40ms. The drain in the figure has been connected to the +KM and 10K resistors through the wiring terminals. The switching response time Why is it so long?

If your +KM is high voltage it can slow down things quite a bit.

If you are having switching time issue on both edges you have two problems.

The high to low transition at the MOSFET drain is set by how current is fed into the opto.

The low to high transition at the MOSFET drain is set by the turn-off time of the Opto and also the 10k resistor on the gate.

You first need to workout the turn-off waveform at the gate isn't slowed down by Opto's, especially the turn-off.   If that looks Ok then it's likely the 10k is too low.

Now the problem: In order to speed-up the drain low to high transition you need to decrease the 10k but that robs the drive with from the opto on the low to high transition.

You may be able to drop then 10k then boost the opto current a lot get it to work but sometimes it just cannot be done.

The old-school solution to this an active driver like this one,



When the MOSFET is turned off the pull-down current at the gate is much increased as the 10k at the base of Q3 now looks like 100ohm or so at the gate.

When the MOSFET is turned on  Q2  can provide a greater drive current to switch the MOSFET on quicker.

Another old-school solution was to use paralleled inverter CMOS gates instead of the buffer.

[bluebunny]

The post at the top appears to be cut'n'paste verbatim from an eight year-old fixya post... 

[Rob Strand]

The post at the top appears to be cut'n'paste verbatim from an eight year-old fixya post... 

Not another one.

We needs to eradicate those FR's.

[niektb]

funny though cause his other topic doesn't seem to be copy-pasted