How to eliminate the Miller effect on BJT BC547?

Question

I have an application:

• The input in a 3V train pulse @ 100Khz
• I need an Ic to be 10mA for the LEF of the optocoupler.

I am using BC547C:

• Input Capacitance= 9pF
• Current Gain Bandwidth Product(hFE)=400.

The problem is the Miller effect, it will make (Rb+CM) act as a low pass filer at10Khz if m calculation is right. Miller effect info

How can I solve this problem Are their another way to get my desire output ?

Is this circuit with mosfet will work ? I could change the LM231 output from 5V to 15V.

The lm231 circuit I used is at page 10 in this datasheet

Answer 1

Sound like you want to drive an opto-isolator with an LM231 output.

The LM231 has an open-collector output and cannot source current- so it won't work with your circuit as shown. It doesn't have a quite enough current capability to be guaranteed to drive the 7mA you've chosen for the opto.

Possible solutions include:

1. Use a better (higher CTR) opto-isolator and drive it with 5mA. Then you need only connect the optoisolator LED to Vcc with a series resistor.

2. Use a PNP emitter follower as so:

^{simulate this circuit – Schematic created using CircuitLab}

As I mentioned in my comments, the 4N25 with a suitable load resistor (something like 7-10K) will be very slow and may not work up to 10kHz.

Edit: Your modified circuit will work, however use a 2N7000 or similar small MOSFET rather than an IRFB4310, use 470 ohms rather than 2K, and connect a 2K from the gate to Vcc (pullup resistor).

Answer 2

First off, to get 10mA through the opto-coupler LED you need to reduce the value of RLed. The LED drops about 1.2V, so RLed should be 380 Ohms or less.

Secondly, at 160k Rb may not provide enough Base current to get a reliable 10mA at the Collector. Using a much lower value would ensure that the transistor turns on fully, and also reduce Miller effect.

You could connect the LM231's output directly to the Base of the transistor, then it will turn off much faster because the comparator is providing a low resistance path to discharge the miller capacitance. With a 10k pullup resistor you should get rise and fall times of less than 200ns at the LED.

The 4N25 has inherently slow response, so your 100kHz square wave input will come out looking (at best) more like a triangle wave. This may be OK if you just want to get some kind of pulse through and aren't worried about distortion.