Best Approach for Long-Distance Sensors

The Project

So nervous, I have a project idea., I don't know if it's okay or not

I want to monitor multiple motion and/or ultrasonic sensors to build a simple proximity alert. I want to "deploy" sensors at key points along the perimeter of my home (outdoors). Later, I'd like to send z-wave notifications to my home controller and play a gentle alert tone for large signals but ... baby steps... For now I'll settle for a simple output (a lit LED will do) just to get things working.

The Problem

I'm aware of the components I'll need and have an idea for the sketch to drive it (either with one or as many sensors as an Uno 3 can support after shields) but I'm at a loss regarding distance. I have a choice between multiple Arduinos with their own sensors, each communicating back to home base OR one Arduino in the loft with multiple sensors and a hydra-like set of cables to each, as well as a SINGLE hardwired power draw. I like the latter approach as it's less wasteful in a number of ways and is certainly cheaper (especially considering radio shields).

The Actual Question

What I want is recommendations on how to approach (if it's even feasible) long-distance connection of the sensor components to the Arduino. Specifically:

Might some of the extra Cat-5 I have laying around be a good match for this?
If not, why (I'm trying to learn - a reference to good reading is all I ask)?
Is there an alternative I hadn't considered?
Am I trying to run before I can walk?
Mother will they tear your little boy apart?*
*Okay, so that's just a Pink Floyd reference. I'm just getting started with the Arduino as a new hobby (I'm a software amateur, no EE background or experience). I'm currently facing the problem of "not knowing what I need to know to ask a good question." Please forgive any perceived laziness and feel free to school me. :-)


Further research did turn up this thread where someone suggested this is feasible but the OP said one sensor wouldn't respond. A responder said one sensor's manufacturer suggested a low pass filter on the sensor side of the connection, as close as possible. Thoughts?

Also, it occurs to my network engineering background that if multiple cat-5 runs in different directions wastes at least 2 of 8 strands, maybe a 2-pair telephone data cable would also work just fine, assuming the tips are soldered to make them solid leading into the breadboard. Again, thoughts?
Thanks suggestions, thank you all guys.


  • The design basis of CAT5/6 cabling is one of electrically balanced and isolated signalling. These two factors provide a high level of electrical robustness. Wired Ethernet uses transformer coupling to address both.

    A typical board level sensor IC does not have these ruggedised attributes built into the IC. You would be advised to add electrical interfacing components to bring your intended comms up to this level to continue with unshielded twisted pair (UTP) cabling like CAT5.

    Alternatively, using twisted pair shielded cabling can assist in reducing the electrical interfacing demands. The more basic strengthened differential line driver arrangement, as opposed to transformer coupling, can be very effective here. Not using at least a differential driver/receiver into twisted pairs of wires will dramatically increase the level of unwanted crosstalk.

    Without shielding on the cable even a toughened line driver can suffer much worse effects than a bit of data corrupting crosstalk.
  • kwinnkwinn Posts: 8,440
    edited 2017-11-25 - 21:02:23
    For a typical house my suggestion would be to use the hydra approach as long as the sensor output is a digital signal. Each cat5/6 or phone cable could use one or two pairs to power the sensors and a pair for the digital signal. If noise is a problem an rs422/485 line driver/receiver can be added to the signal line.
  • temegun: Not sure if you had your question answered or not.
Sign In or Register to comment.