Multiple Sensor Array

Until the total resistance in the common wires approaches 10 ohms you're unlikely to notice any interferance between devices. (assuming 10 bit DAC)

Can i ask how you came up with that figure and why ?

TIA

Feed your A/D's Vref from the same source as the top of the voltage dividers. That obviates any effect from regulator variation with load or temperature (assuming it stays within the allowable window for Vref).

If you're feeding them daisy-chain style, as you've shown, rather than a star configuration, then you may see the effects of the voltage drop in the supply and ground rails. Whether that's significant is something you'll have to calculate, based on the number of bits you need, how many drops there are, and the contributions of other error sources.

E.g.,

V5 ___Rwire ___Rwire o---.----|___|----.----|___|---- ... ----. | | | .-. .-. .-. | | | | | | | |47K | |47K | |47K '-' '-' '-' | | | .--o .--o .--o | | | .-. .-. .-. | | | | | | | |RT | |RT | |RT '-' '-' '-' | ___ | ___ | o---'----|___|----'----|___|---- ... ----' GND Rwire Rwire

(created by AACircuit v1.28.6 beta 04/19/05

I've got seperate voltage divider feeding one channel so I can monitor the voltage used to calc the temperature of each channel.

Just the Thermistors are fed on a long wire pair (about 20 - 30 metres max ), everything else is built & powered from 1 circuit board next to the IO Device.

Any ideas on the best way to mount a PCB to DIN rail ?

Thanks

Hey Terminal, I have no answer to your question, but here's a screenprint from my temperature monitoring system. I have it set so if any sensor goes above 5*F it the display will turn red as an indicator there may be a problem. As you see I have 10 freezers functioning properly and 3 that are empty and off. The program is always running on my computer, so a quick glance several times a day lets me know things are OK. When a freezer is empty the program allows me to move that freezer number into the other group. I have two groups, Full Freezers and Empty Freezers, I could have many more.

Lets assume you load the 317 to 50%.. We need a Total R load of 10 ohms.

TotalUnits = (1/10)/(1/47k) = 4700;

I don't think you have much to worry about..

Plug in octal housing and a DIN relay socket?

I didived your 47K by about 500 but it's just a guess. figuring half the noise wouldn't make it back to the DAC

I was only figuring one device effecting the reading from another, upon re-thinking it I reaslise it's likely that several thermisters will show the same trend ans so their effects will add.

if you have multiple devices then divide the 10 by (n-1)

If you want to be sure do up a circuit model and apply kirchoff's laws.

another option is to run the bias current up the sense wire and switch that current to only power one thermistor at a time

Perhaps you could use a 4017 decade counter, put the 47K bias resistor in series with the VCC supply measure the voltage at the VCC pin and connect the thermistors between the outputs and ground.

your micropcontroller would need to emit clock and reset pulses to drive the 4017 and read one anaglogue input...

HEF4017 needs 3V to operate reliably so the bias resistor may need to be reduced until there's sufficient current through any thermistor for 3V drop.

If you can run each thermistor on a separate loop that will reduce crosstalk too, another option is to use thicker wires for the +5 and the ground return,

or even put the 5V supplky closer to the thermistors.

• ----------------------[7805]--+... 9VDC (isolated) | | - ------------------------+ [R1] others | | wired in --------------------------------+ parallel sense 1 | | like original | [T1] diagram analogue ground | | --------------------------+-----+...

for the 9V you could use something like a YCL101 from a old thin coax ethernet card. or a transformer with two secondaries... it's important that the 9V is lsolated from the main supply so that thermistor current doesn't flow in the analogue ground wire