Opto Isolation issues

[Jon Slaughter]

I'm trying to figure out why different opto's for a midi isolation
bridge(4-20mA current loop with isolation) doesn't work for most of the op
amps I've tried.

The one that works, but not well is the TLP630. The rise time is aweful but
enough so that I can sharping using two mosfet to get it to work.

The others,

H11G3, 4N33 - Darlingtons
CNY17, MOC8205

The darlingtons are total crap. I may be using them wrong though but the
rise time is about 1/20th of the others. I'm not sure what to do with the
base but I have tried various things as seen in other circuits including

http://www.midi.org/techspecs/electrispec.php

Usually, such as connecting the base to V, complete ruins the signal and
draws a significant amount of current.

The other opto's don't work well either.

I'm not sure if this is normal behavior and they are just extremely slow or
if it the CTR and all the ones I have are too low(although this doesn't
explain why the TLP630 works).

Any Ideas what I'm doing wrong?

 

[Tim Williams]

Miller is a bitch. So don't let him get into your circuit.

High speed optos (e.g. 6N136) use a photodiode effectively cascoded into a
transistor. You can add base-emitter resistance to reduce CTR, improve
noise immunity and reduce fall time.

Even a 4N35 will do ~2us edges if you use the right combination of drive
current, B-E resistance and load current. Caveat: if CTR drops, it won't
turn on as fast, or at all.

Anything with "phototransistor" or "darlington" in it (let alone both!)
isn't good for anything above ~100kHz.

Tim

 

[Jon Slaughter]

Miller is a bitch. So don't let him get into your circuit.

High speed optos (e.g. 6N136) use a photodiode effectively cascoded
into a transistor. You can add base-emitter resistance to reduce
CTR, improve noise immunity and reduce fall time.

Even a 4N35 will do ~2us edges if you use the right combination of
drive current, B-E resistance and load current. Caveat: if CTR
drops, it won't turn on as fast, or at all.

Anything with "phototransistor" or "darlington" in it (let alone
both!) isn't good for anything above ~100kHz.

Tim

Ok, I think I got it. The opto is simply a bjt with an optically driven base
instead of electrically(I know it's obvious but just now sinking in that
it's this simple). All the standard rules of bjt's apply with the obvious
adjustment of the base current by the CTR. Obviously the opto's bjt's may
not be the same spec as discrete, etc but all the info should be in the
datasheets.

 

[Tim Williams]

Obviously the opto's bjt's may not be the same spec as discrete, etc but
all the info should be in the datasheets.

You wish. ;-)

Although if you're lucky, you might find a plot of CTR vs. Rbe. Not sure
I've ever see t_r / t_f vs. Rbe.

Tim

 

[Jon Slaughter]

You wish. ;-)

Although if you're lucky, you might find a plot of CTR vs. Rbe. Not
sure I've ever see t_r / t_f vs. Rbe.

Hence the "should"

 

[Joe G \(Home\)]

I'm trying to figure out why different opto's for a midi isolation
bridge(4-20mA current loop with isolation) doesn't work for most of the op
amps I've tried.

The one that works, but not well is the TLP630. The rise time is aweful
but enough so that I can sharping using two mosfet to get it to work.

The others,

H11G3, 4N33 - Darlingtons
CNY17, MOC8205

The darlingtons are total crap. I may be using them wrong though but the
rise time is about 1/20th of the others. I'm not sure what to do with the
base but I have tried various things as seen in other circuits including

http://www.midi.org/techspecs/electrispec.php

Usually, such as connecting the base to V, complete ruins the signal and
draws a significant amount of current.

The other opto's don't work well either.

I'm not sure if this is normal behavior and they are just extremely slow
or if it the CTR and all the ones I have are too low(although this doesn't
explain why the TLP630 works).

Any Ideas what I'm doing wrong?

Another idea is to use one of the Analog Devices iCouplers.

Joe

 

[JosephKK]

Hence the "should"

Still a false assumption. The very assumption that is reasonable to
fully characterize every possibly useful parameter over many
production runs is so beyond cost acceptable. They are in business
to make a profit, not to do and publish the results research that
will their competitors more than it could possibly profit their
customers.