Hi, If you're acquiring a 4-20mA signal across a 250 ohm resistor then
conditioning it (filter, scale, etc) with a reasonably standard op-amp
setup, what sort of provisions are likely to be required to ensure that
common serial protocols overlaid on the loop won't cause problems? The
application under consideration is fairly standard industrial, specified
accuracy around 0.5%, response times of a second or so. We don't know which
types transmitters may be encountered, so the question can't be any more
Asking the question another way, approx what sort of electrical specs apply
to these serial protocols - magnitude, timing? TIA
On Sat, 21 Oct 2006 17:23:35 -0700, "Bruce Varley"
Why don't you filter the resulting voltage drop across the resistor
and not the current loop itself? Doing it this way will not effect the
capacitance or reactance of the loop.
I did a quick search for the capacitance limits on RS-484 loops but
did not find any specs. I know I have them somewhere because this has
come up in the past when I had problems with a loop. The problem then
turned out to be a ground loop in a component that was not fully
isolated, not filtering, but I remember measuring loop capacitance and
digging up the spec somewhere.
I am cleaning out office files right now. I might run across the old
paperwork in the next day or so. I did a bunch of field tests on loops
then, trying to get a feel for how robust a loop like this really is.
The comm speed made a big difference. Do you know what that will be
and what protocol will be used?
On Mon, 23 Oct 2006 19:23:31 -0700, "Bruce Varley"
Is anything existing?
Do you know how many resistors are going to be in the loop?
0.5 % accuracy is pretty good for the loop. You need to make sure the
sensing device across the resistor is high impendence to get that
accuracy. You can put resistors in series to the sensing device that
are less than 1/1000 of the input impendence and not effect the
accuracy or the serial comm that is on top of the loop. This in
conjunction with a cap across the sensing devices input will average
the reading without affecting the loop.
You can only put around five sensing resistors in the loop before the
total loop resistance becomes too high for the power supply of the
transmitter to force 20 ma current through the loop. I am assuming you
know this, but it never hurts.
If the current loops are not totally isolated from ground, you will
have problems. I would suggest that you insist the loop devices are
isolated or have isolators installed between the loop and the devices.
AGM makes isolators. I can't stress this enough. I have spent days
trying to get rid of stray current in existing loops. I had one that
somehow got current from a dawn to dusk light. The loop was connect to
a computer and the values were being logged. When I looked at the
graph with the operators, it became obvious that the signal was wrong
only at night and the error cycled on and off. I happened to spy a
dawn to dusk light that would try to start and then the thermal
overload would trip the light. Ten minutes later, it would try to
restart. Watching the light and the graph on the screen confirmed the
connection. The light had ground faulted and was pumping current into
the ground. Enough was leaking through to the loop to make about 20%
error in the signal.
At least I was working by the hour then.
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.