James said:
mooseo wrote:
...
...
I read the other responses and don't see any convincing arguments re
what the problem is exactly, which in part may be because your
description of what you are multiplexing and how is rather vague.
Re the schematic in
http://www.emesystems.com/OL2mhos.htm that you
referenced, did you build one whole circuit per sensor and then
attach each - (white) output to an A-to-D multiplexer input, with
all = (black) lines in common to the other multiplexer input, or
what?
Perhaps you could see if the problem still occurs with the following
hookup: Build one 555 circuit, but without the 2.2 uF caps at the
sensor lead connection points. Locally at each sensor, put a 2.2 uF
cap in series with each lead and waterproof the connections. Use
relays or manual switches to switch one pair of sensor wires at a
time into the connection points.
-jiw
Hi James,
Thanks for the suggestions. I'll try to provide some clarification.
I've been doing some experimenting with your ideas to see how they
work.
I've tried a number of different multiplexer scenarios, all based
around a complete circuit per sensor. The simplest setup has been to
connect all of the black wires together as a common ground with the
power supply and the A/D board. The red wires were all connected to the
positive of the power supply, and the white wires went through the
multiplexer to a single channel on the A/D. This caused all of the
afformentioned problems.
I've also tried switching the red and black wires through the
multiplexer with the white wires going to different A/D channels...
this still had the same problems. When I try this with a "monkey relay"
(i.e., running them all through a breadboard and physically plugging
and unplugging the red and black wires while the white stay connected
to the multiplexer) everything works fine. This suggests that using
mechanical relays to switch the power should work (I'm going to try
this next week), but running mechanical relays at 10Hz will probably
cause problems in the not-too-long term. Solid state relays are a
possibility, but I don't have any around that I can breadboard up.
I've tried your suggestions of switching the probes into a single
circuit, but with no luck. I tried it in a couple of configurations,
both with the capacitors at the circuit end of the wires (which was the
default) and out at the probes as you suggested. In the first case, the
voltage signal drifted all over the place... in the second case (with
the caps out at the probes) the voltage didn't respond to the probes
being submerged... this was with the probes running through the
multiplexer, so I'm guessing that the resistance and capacitance of the
circuit become highly variable.
At this point, a relay solution might be the best bet, but I'm still
waiting on an order that will have some inductors to use to try to
isolate the AC noise off the ground. And I'm very much open for other
things to try.
Thanks again for everyone's suggestions.
mike