I'm trying to control a heavy load using ESP 32. An optocoupler EL817 is used for galvanic isolation. An S8050 bipolar NPN transistor is used to amplify the control signal. At the end is the Ne-MOS FQP50N06. Without the Gate-Source pull-up resistor, the mosfet is always half open. When I apply a signal to the optocoupler input, the transistor opens and allows current to flow, but heats up very quickly. I assumed that this was due to the incomplete opening of the transistor. I measured the voltage between the Gate and the Source - 3 volts. Not enough to fully open the FQP 50N06. I changed the transistor to 70T03GH, which has Vgs = 3 volts, but it still heats up a lot. What is the problem?
I think you need a different MOSFET. From the data sheet, it looks like this one barely turns on at 4v (Gate-Source).
I’ve redrawn the schematics to something more normalised.
Now it becomes clear what's happening.
When the optocoupler starts to conduct the gate voltage starts to rise.
When the Vgs reaches the threshold voltage, the source current starts to rise.
The gate voltage will rise up to battery voltage minus twice Vbe.
When the source currents wants to rise so much that Vgs sinks below the threshold voltage, the MOSFET will pinch.
Voltage between Drain and Source will be a bit higher than the threshold voltage, resulting in heat generation.
It's possible that your drawing is not accurate.
The NPN transistor is not necessary.
You might want to use a P-MOSFET for high side switching.
So, basically, the load is in the wrong place...
@quarklabs
What do you measure on the Base of the transistor?
In this configuration the 1k is not need.
You don't need the transistor.
1 Like
I will use this scheme in the PID adjustment. Due to the high frequency, the optocoupler will not have enough current to charge the gate of the transistor and its transient state will cause heating. It's not difficult for me to add two extra transistors to make this process more efficient and high-quality.
The turn-off time of that two-transistor circuit must be very slow (bad), because of the slow discharge of the gate through that 100k resistor.
The circuit from @jim-p (post#6) is much better, assuming you actually need opto isolation.
Otherwise a logic level fet, directly driven by the Arduino pin is the best solution.
Leo..
That PNP transistor is doing nothing.
The opto-isolator can't pull it's base low to turn it on.
The opto-isolator can only pull things high, or let them float.
Same bad circuit.
The gate only gets discharged through the 100k resistor.
The PNP transistor does nothing here, because it gets no base current (to ground).
Don't re-invent the wheel if you only have a basic understanding of electronics.
Again: why do you think you need opto isolation.
An opto coupler is inherent slow-ish.
Your previous threads (Peltier-cooling a DSLR camera) doen't seem to need it.
A FQP50N06 is not a logic level fet, so the wrong choice for simple direct-drive with an Arduino pin.
Leo..
But to gain this knowledge, I need to reinvent the wheel... but I can't invent it because I need knowledge?
Because with an optocoupler, the probability of burning down the microcontroller is zero (except for attempts to throw the circuit onto the power line wires).
The chance that happens (with the right gate resistor) and a 12volt supply is near zero.
Study this page, Circuit#1.
Leo..
100 kohm != 4k7...
This value is not critical, but 100k is really high... larger fets will have larger gate capacity. It will take a while to fully shut down the fet...
This will cause heating of the fet..
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.