thanks for all. to be honest, i still cannt understand the deep principle of this kind of opto driving circuit design.
OK, the basic elements are
1) the LED (part of the optocoupler)
2) the transistor (to switch the LED on and off
3) the series resistor (to limit maximum current)
4) the parallel resistor.
I presume you are fully comfortable with 1, 2, and 3.
the parallel resistor is rarely used in this application, but may be used in certain circumstances.
Because the LED represents a very high impedance until the voltage across it reaches Vf, as soon as any current flows at all, Vf will be the voltage across the LED (with minor changes since Vf depends on If to a significant extent where If is small).
Thus if the voltage across the LED and the series resistor rise above Vf, there will be some light emitted from the LED.
Regardless of how good the transistor is, there will always be some leakage, and therefore some emission from the LED. This may not be significant, but on the other hand it might be...
The fix for this is to place a resistor across the LED. The two resistors act as a voltage divider now, and for any particular current the voltage across the LED depends on the parallel resistor (until that voltage reaches Vf). Thus at small currents the voltage is far lower than Vf and therefore no current flows through the LED and consequently there is no emission from the LED.
So the resistor across the LED ensures that the led will completely extinguish when the current through the switching device (the transistor in this case) falls to some low value, rather than zero (which may be unachievable)
The same issue can be discussed as a voltage rather than current issue, in which case the parallel resistor increases the voltage across the LED/series resistor combination required to turn on the LED.
Resqueline observed that this is useful for a LED driven by a capacitor. In this case the advantage is also that the LED will turn off completely at some calculable time rather than the brightness following a curve which is asymptotic to zero.
do u have general or standard circuit design when u deal with opto design. thanks for ur share if u like that...
Yeah, essentially exactly the same as your original circuit with R1 removed.
R1 is there for a number of special cases which arise infrequently.