Wierd Opto Behavior

[Jon Slaughter]

http://senduit.com/e1e047

This is from a 6N138. I used a PC900 configured exactly the same way(but
obvious changed in pin configuration) which gave much better results.

Whats causing this behavior? The rail voltage is 5V.

 

[pimpom]

Methinks we need more info..a circuit perhaps?

That's what I thought too.

 

[Jon Slaughter]

That's what I thought too.

Basic circuit. As I said, the PC900 works just fine. There is no load and
obviously the drive is not the issue. Hence on the output it's obvious(just
a resistor in the standard configuration). i.e., the circuit is the obvious
one... the minimal one to get any useful output from the device.

i.e., it's what you call a "test circuit"... which, IMO, is what you should
assume when no other information is given.

 

[Jon Slaughter]

Three "obvious"'s when nothing is obvious.


When no context is given, what we assume is that the problem is
underspecified.

Or perhaps your simply trying to overanalyze the issue? It's not
complicated. You can easily gather basic information from the original post
to remove most doubt. First, it's for digital logic by the pic. Second, the
PC900 works in the same circuit. Third, it's most likely not some complex
circuit(just by shear probability). This is enough information, if you have
seen this behavior before, to know what it is caused by. If you haven't seen
it before then obviously you will want more information because you'll not
have a clue what is causing it.

In any case I'd imagine that if I showed the circuit(which I won't because
it is too obvious) you still wouldn't know what is causing it and would then
want to know what kinda scope I'm using, what resistor types, tell me to
remove the capacitors(which there are none but that won't stop you), A photo
of the actual circuit since surely I must have wired up something wrong,
etc...

If I give you too much information it will simply give you more
opportunities to be wrong. The key here is that the PC900 works fine in the
exact same circuit. This alone is enough to deduce the behavior(since the
6N138 is similar specced and generally used as a replacement).

The point is that if you didn't get an idea of the problem in the first post
your not going to get it without knowing way more information than I'm
willing to sit down and type in.

Don't be ashamed of not knowing though. It's not that big a deal. I'm sure
you could figure it out if you had the problem but I'm not willing to be
asked what color socks I was wearing when I took the pic. AGAIN, either you
have seen such behavior before and have come clue what is causing it(one may
think initially it is capacitance but then it should most likely cause the
same problem with the PC900) or you don't have a clue.

If you want to know what the circuit probably looks like then do a google
image search for "Opto circuit" and pick one at random(at random) and you'll
probably have a 19/20 chance of getting the circuit I'm using.

 

[Tim Williams]

Circuit?

Since, in subsequent replies, you appear to be reluctant to provide one,
savor my suspicion: are you using the common ("emitter") pin for voltage?
Is there any common mode voltage (LED to output side), and if so, how much?
Is the base pin shielded, or are you doing something bizarre with it?

I've tried using 6N136's at 1000V/us common mode (300V step) and signal
quality fails. Fairchild I think rates theirs to 10kV/us but only for a 10V
step, useless.

Tim

 

[Grant]

....

The point is that if you didn't get an idea of the problem in the first post
your not going to get it without knowing way more information than I'm
willing to sit down and type in.

The problem is obvious, the solution likely depends on context.
....

If you want to know what the circuit probably looks like then do a google
image search for "Opto circuit" and pick one at random(at random) and you'll
probably have a 19/20 chance of getting the circuit I'm using.

In that case you already have your answer too.

Grant.

 

[Jon Slaughter]

Circuit?

Since, in subsequent replies, you appear to be reluctant to provide
one, savor my suspicion: are you using the common ("emitter") pin for
voltage? Is there any common mode voltage (LED to output side), and
if so, how much? Is the base pin shielded, or are you doing something
bizarre with it?

No, nothing bizarre... your standard basic isolated newbie digital comm
stuff that you can find when searching.

I've tried using 6N136's at 1000V/us common mode (300V step) and
signal quality fails. Fairchild I think rates theirs to 10kV/us but
only for a 10V step, useless.

What I see in the image is we have a capacitor rise time to about 3/4Vcc
then a short time later we get another capacitive effect with a much slower
rise time to Vcc.

I've never seen something like this before and only could understand it if
it was intentionally done.

Think of it as a problem on an exam: "How can a PC900 give almost ideal rise
time characteristics but a 6N138 give the wierd behavior in the image"?

What would be your answer? If you say "Common mode voltage" then that would
be wrong would it not? Since the PC900 would experience the same issue(more
or less). You say "It has something to do with the base pin shielding"...
well then why doesn't the effect show up with the PC900? etc...

Now, I am going on the assumption that the 6N138 is a direct replacement for
the PC900... I have read this from about 10 different sources showing the
6N138 used with the same results. This is the issue: I dropped in the 6N138
and didn't get the same results with contradicts what those 10 sources have
said. It has nothing to do with the circuit and giving the circuit will only
get more dead end answers. Again, I'm not saying the circuit may be
excerbating the problem but it is not the root cause of the problem. It may
be simply that all opto's behave this way more or less and the circuit I'm
using magnifies the problem. But if someone knew this answer they would say
that instead of asking for the circuit.

It seems that many of you guys don't want to make the assumption that I'm
using a simple and common circuit but don't mind the fact of making the
assumption that I wired of the circuit wrong. If your not going to allow one
assumption then don't allow the others as both are just as important for
finding the real answer. (Tim, I'm not talking about you here since you
obviously gave some potential answers)

 

[pimpom]

No, nothing bizarre... your standard basic isolated newbie
digital
comm stuff that you can find when searching.


What I see in the image is we have a capacitor rise time to
about
3/4Vcc then a short time later we get another capacitive effect
with
a much slower rise time to Vcc.

I've never seen something like this before and only could
understand
it if it was intentionally done.

Think of it as a problem on an exam: "How can a PC900 give
almost
ideal rise time characteristics but a 6N138 give the wierd
behavior
in the image"?
What would be your answer? If you say "Common mode voltage"
then that
would be wrong would it not? Since the PC900 would experience
the
same issue(more or less). You say "It has something to do with
the
base pin shielding"... well then why doesn't the effect show up
with
the PC900? etc...
Now, I am going on the assumption that the 6N138 is a direct
replacement for the PC900... I have read this from about 10
different
sources showing the 6N138 used with the same results. This is
the
issue: I dropped in the 6N138 and didn't get the same results
with
contradicts what those 10 sources have said. It has nothing to
do
with the circuit and giving the circuit will only get more dead
end
answers. Again, I'm not saying the circuit may be excerbating
the
problem but it is not the root cause of the problem. It may be
simply
that all opto's behave this way more or less and the circuit
I'm
using magnifies the problem. But if someone knew this answer
they
would say that instead of asking for the circuit.
It seems that many of you guys don't want to make the
assumption that
I'm using a simple and common circuit but don't mind the fact
of
making the assumption that I wired of the circuit wrong. If
your not
going to allow one assumption then don't allow the others as
both are
just as important for finding the real answer. (Tim, I'm not
talking
about you here since you obviously gave some potential answers)

Did those 10 different sources explicitly say that the 6N138 is a
direct replacement for a PC900? Judging from a quick glance at
the datasheets, the PC900 is a Schmitt device while the 6N138 is
a linear device. That's certainly not direct replacement. The two
may function equally well and may indeed be interchangeable *in
some applications*, and those 10 sources may be talking about
specific applications in which the 6N138 can replace the PC900.
But I certainly would not say without qualification that one is a
direct replacement for the other.

Under some drive and load conditions, it's quite natural for the
PC900 to exhibit snappier switching because its internal Schmitt
trigger works on the threshold principle. OTOH, the 6N138 has a
roughly linear transfer characteristic.

I didn't have time to go through the 16-page datasheet of each
device in detail, but this much is clear: The two are not
identical devices and they will definitely exhibit differences in
behaviour under certain conditions. And most prominent among
those conditions are the associated circuits.

 

[Jon Slaughter]

http://senduit.com/e1e047

This is from a 6N138. I used a PC900 configured exactly the same
way(but obvious changed in pin configuration) which gave much better
results.
Whats causing this behavior? The rail voltage is 5V.

Ok, I was able to drastically reduce the behavior so it is close to ideal.
The problem seems to be because the base of the 2nd transistor in the
darlington pair is storing charge and cannot turn off quick enough. Adding a
resistor from the 2nd base(pin 7) to ground helped a lot. Strange all those
people claiming that the 6N138 was a direct replacement without mentioning
this(maybe it is suppose to be obvious since it is a darlington opto) ;/

Now I guess I know what slow base recovery looks like in a darlinton. Can
someone give a good explanation why is going on as I can't seem to find
anything online about it? (I just remember somewhere reading that one could
add resistors to improve the speed of darlingtons)

At least I think this is whats going on...

http://senduit.com/677d98

 

[Hammy]

Ok, I was able to drastically reduce the behavior so it is close to ideal.
The problem seems to be because the base of the 2nd transistor in the
darlington pair is storing charge and cannot turn off quick enough. Adding a
resistor from the 2nd base(pin 7) to ground helped a lot. Strange all those
people claiming that the 6N138 was a direct replacement without mentioning
this(maybe it is suppose to be obvious since it is a darlington opto) ;/

Now I guess I know what slow base recovery looks like in a darlinton. Can
someone give a good explanation why is going on as I can't seem to find
anything online about it? (I just remember somewhere reading that one could
add resistors to improve the speed of darlingtons)

At least I think this is whats going on...

http://senduit.com/677d98

This application note may be useful.

" Optocoupler Applications"

From California Eastern Laboratories

http://www.datasheetcatalog.org/datasheets/228/466209_DS.pdf

 

[Jon Slaughter]

This application note may be useful.

" Optocoupler Applications"

From California Eastern Laboratories

http://www.datasheetcatalog.org/datasheets/228/466209_DS.pdf

Cool, thanks!

 

[pimpom]

Ok, I was able to drastically reduce the behavior so it is
close to
ideal. The problem seems to be because the base of the 2nd
transistor
in the darlington pair is storing charge and cannot turn off
quick
enough. Adding a resistor from the 2nd base(pin 7) to ground
helped a
lot.

So it _was_ the circuit after all?

Strange all those people claiming that the 6N138 was a direct
replacement without mentioning this(maybe it is suppose to be
obvious
since it is a darlington opto) ;/

Do you mind providing the URL for at least some of those sources?

As I said in my other post which you haven't responded to, the
PC900 and the 6N138 are very different devices. The PC900 is a
Schmitt trigger device which is optimised for fast, clean
switching and will not do well at all in a linear circuit. The
6N138 is a linear type and its switching behaviour will depend on
the circuit which you stubbornly, even rudely, refused to
provide. The two are *not* direct replacements for each other
except under certain specific conditions.