Conductivity sensors

All should have one common probe, like a ground.

Hi Homer, thanks... I'm not quite sure what you mean by this... all the circuits are on a single power supply, so they share a ground. The probes themselves are connected between pins 3 and 6 on the 555. Are you suggesting that I ought to link these circuits in some other way?

Find a design where one probe is connected to ground. Yes, you need to link them all that way.

When you multiplexed the powers, did you isolate the grounds as well. Did you have filters isolations for the power supply? Did you isolate the probes? Treat this as a very noise sensiitive circuit.

Hmm, I'm not sure I understand. Are you suggesting that I modify the design I have, to somehow connect the probes to ground, or are you just suggesting that I need to come up with a totally different design? If the latter, do you have any other suggestions about it, other than that one probe needs to be connected to ground?

Hi, Yes, I did switch both the power and the ground through the multiplexer. The other connection was the voltage out, which went into the A/D. I did not have any filters on my power supply. What sort of filters would you suggest? Maybe I'll try running it off a battery to see if that helps.

thanks, mike

Look for (moisture meters) and find a design where one probe is connected to ground or to the power supply. That way all of those can be commoned.

Inductors.

Yes, that should help.

I've just tried running the circuit off batteries... both 1 battery for the whole thing, or separate batteries for the multiplexer and the probe, with the grounds connected, but the circuit behaves exactly the same. Individually, each probe registers being submerged by increasing the voltage drop across the resisstor, but when 2 are submerged in the same liquid, both have the same drop as if they were not submerged at all.

any other thoughts? I can try getting some inductors, although it really doesn't look much like a noise response... more like the two circuits are inhibiting each other.

thanks, mike

Why connect the grounds? Even if you have to connect them, use inductors. The probes seems to be AC shorted.

Same difference, signal from one probe is just noise for another.

For the same reasons as you can turn on one light in your house without affecting the others even though all are connected together.

That may be true for 60Hz, but not for higher frequencies. At 10 Khz (and harmonics), signals can travel in different paths. Unfiltered power supplies are just AC shorts.

If all of your monitor units share one common ground, each can monitor it's probe separately. Of course the output must be isolated from each other.

However this won't work with the indicated circuit. He could, however, switch the CS lines as he reads each sensor module and this would avoid the need for isolation.

You are thinking digitally; however, his problem is analog. He already tried switching the power lines, so CS would not be any better. The problem is uncontrolled impedence paths, whether they are active (powered) or not.

I am assuming he has some sort of isolation on the outputs (otherwise how will it work?) So if he 'reads' one probe at a time it seems reasonable that he will get an isolated result. Still, a solution where one probe is grounded or high would be simpler.

I have the grounds connected simply because the A/D board is set up as single-ended, i.e., it's reading each channel relative to common ground. I can rewire some things to route the ground of each signal through the multiplexer and run the board in differential mode.

I see what you are saying about AC noise, although I'm new enough at this that I have a hard time wrapping my brain around AC. Why is it that the current would drop in a shorted situation? I would have guessed, though without knowing why, that it would have increased...

Thanks for giving me some things to try.

.

Looking again at the specs for the multiplexer (

) I notice that the isolation and crosstalk protection start to drop quite a bit at higher frequencies... I have no idea what the specs are on the multiplexer within my A/D board, but they could suffer similar problems. Might that be a source of "uncontrolled impedence paths?" I've never had to worry about such things before.

I've also never used inductors before. Do you have any suggestions for a general purpose model to start with?

Thanks, mike

I know you don't want to hear this. But the final solution might be optically isolated multiple A2D units. That's why the commerical units are so expensive.

Your 555 circuit works by rapidly charging and discharging by a reference voltage. If that reference voltage (from the probe) is noisy, the basic assumptions are off.

Hmm, no, I really don't want to hear that! Given the realities of my environment, I'm guessing these things will survive about 2 weeks before they get destroyed by waves, so cheap is essential.

I'll try some power filtering and see if that helps.

Otherwise, I guess it might be time to look for a new solution...