[sdiy] midi optocouplers

[harrybissell]

imho its too slow.

I don't know if this is correct of not, but a 32KHz signal would have a
half cycle time of 16us... in effect the frequency is really twice
32KHz.

With the rise/fall times you have there the opto would be in an analog
range all the time, never high or low. That would mean like zero noise
margin...

You could try it if you have one, but I would not count on it.

The lower the value of resistor on the drive (LED) side, the better off
you are
except for the power you need to drive it.  On the receive (transistor)
side...
a smaller resistor will give faster response, but you need a higher
drive current
or good CTR (current transfer ratio - ie gain, beta, etc).  Larger
resistor values
will be slower.

Why not a part with digital output drive ???

H^) harry   (who just got burned on the PC816/817. The proto had gain
range
"B"... 1st production had gain range "A" <lower> and the optos would not
saturate)

Karl Ekdahl wrote:

> A quick question: would a PC814 opto with rise/fall time max of 18uS
> be good for midi? I know the usual suggestions are 6N13X with way
> lower fall/rise but according to my calculations; 1/32150=31uS is
> absolute minimum rise/fall, so about half that should be good, right?
> Also, what resistor to VCC should i use (And why)?
>
> thanks
>
> Karl
>