Max. distance between a temperature sensor and microcontroller.

markuster

Hi,

I need to use a temperature sensor.
My problem is about the distance.

Do you know the max. distance between a temperature sensor
and a Basic Stamp or a Propeller or a SX microcontroller ?

I read about the DS1620 and about the Parallax temperature probe
but I can't find this information.

I need 30 feet from the Basic Stamp to thr sensor.

Thanks, Mark

Beau Schwabe

markuster,

What you have going for you with the DS1620 is that the clock frequency can be from 1.75MHz to DC.
What this means is that you can slow down your communication to the DS1620 based on your distance.

In a length of wire you have resistive, inductive, and capacitive properties. The resistive component in most cases is negligible. Your biggest culprit is going to be inductive and capacitive properties. The length of wire itself is an inductor, and the parallel relationship between the wire and the shield or neighboring wire is capacitive. Together they form a LOW-pass filter which prevents higher frequencies from making their way from one end to the other. As a result, the longer the distance, the lower the frequency you are able to transmit and still have a signal at the other end.

Also, because of that distance I would use termination resistors (1k or so) to ground on BOTH ends with every signal line. This helps to attenuate any RF noise caused by antenna effects due to the length of the wire. The resistors one both ends create a pseudo transmission line that can improve the overall performance.

WBA Consulting

I have never used the DS1620, but if you are still thinking of using the Sensirion SHT11, revert to this thread and the link I mentioned where I had an SHT11 working across 208 feet of CAT5 cable with the demo code and my custom code that logged every 30 seconds for a solid week. As Beau mentioned for the DS1620, you can also clock the SHT11 as slow as you want, so you can easily get past latency problems if you do run into them. You can see the schematic for my module that is still on my SHT11 Module Product page (under the Specifications link).
I only make them for custom orders now, but I can build them fairly quickly. I have a batch in the works using the SHT15 version and will be buying extra for the price break, so I will have some of those to sell within a couple weeks.

varnon

Don't forget about the DS1621, which only requires a 2-wire interface I2C. I thought thats what I was getting when I ordered a DS1620 from parallax. Not that one extra pin is a big deal, but I really wanted to start using I2C devices.

Beau Schwabe

Since (according to the datasheet) there are no maximum times for the Clock, the same would hold true for the DS1621... a slower communication speed would translate to a longer wire run between the I2C master ans the I2C slave. Just make sure that a slower I2C clock is not going to interfere with other I2C devices that might share the same line.

markuster

Thanks:

You have been very clear.

Just , about the thickness of the cable:
If I use a wider or thicker cable if will be better ? , or a shielded cable must be considered.
(or both , the width and shielded? )

Thanks, Mark

Beau Schwabe

A thicker wire will lower the IR drop (Current drop due to resistance) but won't do much in the way of antenna effects. In fact a thicker wire would have a higher capacitance resulting in a lower frequency low-pass filter that would be created, meaning that the transmission rate would/could be affected. To reduce antenna effects, a terminating resistor (of 1k or so) to ground on the signal line (on BOTH sides) is effective. The idea is not to have an extended run of wire susceptible to accepting stray RF. The resistors form a pseudo transmission line by loosely coupling the signal line to ground which can greatly attenuate any RF noise.