Long ground wire

[[email protected]]

Let's say you have a circuit that uses an opto-isolator to trigger a
microcontroller with simple logic high/low. (npn output)

The opto will have a steady DC voltage (through a resistor of course)
on it's led but will use a switched ground to trigger the led.

Now, let's say the pair of wires going to the switch is a couple
hundred feet or even more.

Everything works fine -- but of course, the led isn't going to be
falsely triggered by noise, etc -- it needs it's 5 ma or so to "kick
on".

Wouldn't it still be good practice to do something to eliminate or
reduce the effects of the long ground wire (acting as an antenna?)
which would connect directly to the ground plane of the PCB?

Certainly a well gorunded shield covering the pair would (probably?)
suffice but would you also want possibly an inductor of some sort
between the PCB and the ground (common) wire?

To keep noise out of the PCB ground circuit?

Actually what I have breadboarded works perectly, no issues, but I
don't have a very noisy environment either.

Thanks for any help.

 

[Rich Grise]

Let's say you have a circuit that uses an opto-isolator to trigger a
microcontroller with simple logic high/low. (npn output)

The opto will have a steady DC voltage (through a resistor of course) on
it's led but will use a switched ground to trigger the led.

Now, let's say the pair of wires going to the switch is a couple hundred
feet or even more.

Everything works fine -- but of course, the led isn't going to be falsely
triggered by noise, etc -- it needs it's 5 ma or so to "kick on".

Wouldn't it still be good practice to do something to eliminate or reduce
the effects of the long ground wire (acting as an antenna?) which would
connect directly to the ground plane of the PCB?

Use a twisted pair. This could be shielded, but it's probably not
necessary.

To protect the LED from glitches, just put a .01 ceramic cap in parallel
with the LED.

Good Luck!
Rich

 

[Ken S. Tucker]

Let's say you have a circuit that uses an opto-isolator to trigger a
microcontroller with simple logic high/low. (npn output)

The opto will have a steady DC voltage (through a resistor of course)
on it's led but will use a switched ground to trigger the led.

Now, let's say the pair of wires going to the switch is a couple
hundred feet or even more.

Everything works fine -- but of course, the led isn't going to be
falsely triggered by noise, etc -- it needs it's 5 ma or so to "kick
on".

Wouldn't it still be good practice to do something to eliminate or
reduce the effects of the long ground wire (acting as an antenna?)
which would connect directly to the ground plane of the PCB?

Certainly a well gorunded shield covering the pair would (probably?)
suffice but would you also want possibly an inductor of some sort
between the PCB and the ground (common) wire?

To keep noise out of the PCB ground circuit?

Actually what I have breadboarded works perectly, no issues, but I
don't have a very noisy environment either.

Thanks for any help.

Everything works good until a lightning storm.
Ken

 

[mkr5000]

Using a twisted pair.....and like I say, all is fine.

Of course, the opto is the way to go with something like this....I'm
just sketchy on what happens when the entire ground plane of the PCB
has a few hundred foot long "leg" on it.

I'm talking about the IC pin grounds and everything else....I didn't
know whether a choke in series would be a good idea?

Of course, plenty of .1 caps on the supply buss.

thanks.

 

[mkr5000]

Phil --
That's brilliant.

You're saying just supply direct DC to the LED and then parallel 2
resistors to the ground connection on the board (or even one would be
an advantage).

I'm just being a perfectionist and although this circuit has worked
fine for a year or so, I was always concerned about this....at least
doing it that way would do no harm and you'd have the peace of mind of
a "buffer" from your ground to the outside world.

You're a smart guy.

Have you seen this done in any equipment?