Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

analog sensor cabling question

Status
Not open for further replies.

danrogers

Member
Hi all, I am putting a few temperature sensors around my house (for a thermostat project). The sensors will be up to 20-30 meters from the micro controller so I would like the smallest noise possible. I have two questions

1) I'm using 10k thermistors set up in a voltage divider circuit. The vcc is 5v. Am I better off having the voltage divider at the uC and just two wires running to the thermistor, or setting up the voltage divider at the thermistor and running 3 cables?

2) is there any advantage (noise) in using twisted pair cable? I have got plenty of cat5e that I could run to the sensors?

Many thanks!
 
current is less susceptible to noise pickup. Twisted pair cabling ensures that any noise that gets onto your cables, gets on to both, so the differential voltage across the cables remains the same. regarding 2 wires or 3: it makes no difference. If noise gets onto the junction of your resistor and thermistor, it will do it regardless of whether you have 1 wire or 2
 
For household heating, your concerns over noise aren't warranted.

If you're worried, use dirt cheap micro controllers with an ADC at the thermistor location, the best way to avoid noise in an analog situation is to get it into the digital domain as fast as is possible. Cheap digital thermal sensors are available, even on a low budget.

You may want to rethink your design.
 
@ Ian, I'm not sure how I would set up a constant current supply - I guess I would need some type of driver, would it have a 5v output and would vary the current to keep it at 5v? Not sure how it would work.

@ simonbramble, I was wondering about the effects of having a divider at the therm or at the uC, because if I went for at the uC I would be measuring the resistance of the 10k therm + the wires only. whereas the other method I would be measuring for a signal approx 2.5v + the resistance of the wires - is that correct? I will use the cat5e then thanks :)

@Sceadwian, That is true I don't need high accuracy at all, but I would like the various sensors to have a reading mildly comparable to each other. Maybe I'm worrying about nothing and they won't be too bad. I don't think I will bother with the digital solution because of its complexity to a relative noob like myself but thanks for the input
 
dan, I do think you're over analyzing it a bit. Build it, test it, if there's a problem you'll find out, given you have Cat5 wiring to work with you can't go wrong, you have 8 lines you can dynamically configure.

Remember if waste were an important consideration no one would use UTP cabling because only 2 of the 4 pairs of wires up to and including 100BT networking are actually used, the rest are wasted. Gigabit over copper and power over ethernet setups actually use these extra wires, which is one of the reason why 4 pair UTP is required for Ethernet. Cat5 for 100BT and Cat3 for 10bt, the only real difference is the number of twists per inch of cable and that only has noise reducing effects for differential signals (meaning not applicable in your case) Cat5e is nothing more than a high end quality approved Cat5 cable (oxygen free copper and excellent insulation materials)

Personally if it's not network based I prefer cat3, because the fewer twists per inch make the cables a LOT easier to deal with for the thumb fingered folks like me. For what you're doing you could use nearly anything.
 
Last edited:
Keep in mind that you should be able to lower the noise considerably by using a big capacitor across the sensor wires at the microcontroller. You could even use an additional resistor to keep the cap value within reason. Your temperature changes will be very slow, so the lag from a large time constant filter will not be noticeable.
 
I like Roff's idea. You could go another step since you have a micro and take 3 samples a few seconds apart and average them or throw out the high and the low.
 
I'd go with Roff's suggestion too. Even the cat5 cable is a bit of a luxury. Years ago I built a successful domestic heating controller using an ordinary silicon diode (fed from 5V via a 10k resistor) as the remote sensor, connected over some 10 metres to a comparator by normal mains flex, with a 22μF cap in parallel with the diode at the comparator input to suppress noise pickup.
 
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top