[Everybody] TTL Cable length

[L.Belanger]

Hello,

Since that subject hasn't been discussed alot in this forum (beat me if you find a topic with this subject ;) )... i'm curious to see
what are the max TTL wires length you have achieved without any problems ?

by TTL i refer to 5V logic communication...

Thank you

PS: that thread could be useful for further reference :D

[Guest]

Google is your friend,

w w w.interfacebus.com/Design_Termination.html

Also a book on high speed logic design is a good idea to own.

HTH - Steve H.

[ELCouz]

wow nice find... bookmarked !!

back on topic,

I had never gone past 2 meters ,,, so i guess its the farest i did get without problems :P

[libor]

There's no such thing as a 'maximum TTL cable length'. stretching a cable length to its limits 'competition' has no use.
Think of your cable as a series resistor, a parallel capacitor, an impedance, an inductance, an antenna (both receiving noise and adding it to your signal and also acting as a transmitter of your signal attenuating it) and a delay box (making a problem int to way communication) in one piece.

Getting thru your data depends on these values (all depending mostly of the length, shielding, cable quality, twisted-or-not, bent-or-straight, surrounding environment etc) and most importantly on the speed you are trying to communicate. At a speed of 1 bit per second a cable of several kilometers may work, at a speed of 10 Gbs you are not likely to reach 10 cm.
The transmitting protocol/algorythm (the sampling rate - majority voting/or not, timing tolerance, error correction algorythms, etc.) used can also play a large role in how error prone a communication link is.

A long cable may work on your desktop (at a certain speed), but the same cable may pick up enough noise in an industrial environment making it useless. Or it can work at night (when the nearby AC motor is not running) but fail in the morning. ..or it can work when wound on a bobbin but fail when stretched out. ...etc..etc.

[SET]

Once looked at a system where 2 boards communicated using plain old CMOS logic parts (4000 series) - might have been ok if it wasn't installed on a ship and the boards were hundred meters apart

Needless to say it didn't work very well...

Did suggest to them to add differential drive and receive buffers (RS485/RS422 style), then you can do megabits/sec over long distances.

[ELCouz]

[Guest]

It all depends on your application environment.

Inside an instrument is one thing.

Routed on customer disconnectable cables is another (potential ESD problems).

Routed under the hood of a car is another (Watch those spark plug wires).

Routed by a FM radio station is another (can you say antenna?)

Routed in a robotic assembly line (Large currents = large pickup).

Routed at the end of a runway (watch those radar pulses)

So you see - with proper termination you can run a very long distance with straight TTL - the practical limit will have more to do with the secondary environment.

HTH - Steve H.

[SherpaDoug]

I do 1000 meters regularly in a coax on the bottom of the ocean at 1200 baud using nothing more than a PIC pin, a DC blocking cap in case of mis-wiring, a RC low pass filter and a series termination resistor.
Note that this is a really benign environment.
_________________
The search for better is endless. Instead simply find very good and get the job done.

[L.Belanger]

wow thanks for the (very informative)replies,,

i had an idea for long distance (+100m,home environment but near powerlines) without interferences ... maybe running RS232 through a fiber optic link.

i will investigate more on this ...

anyway, anyone have tried this ... any bad/good experience to share ?

Thank you

[Guest]

Oh - home project - Just use your best termination technique, perhaps some Zener diodes at each end for ESD and just go for it. If you have problems you can always add a line driver and receiver.

The best way to learn is to try things out - you already knew enough to suspect that a long cable might be an issue, but with proper termination this can be overcome.

Have fun!

HTH - Steve H.

[L.Belanger]

Let me explain what specific application...

My uncle own a "sugar house" (sugar shack... where they collect maple sap from trees...then make the maple syrup you know) what even you call it in the U.S. :P

He's currently converting is sugar house to an modern one (adding pipes to trees to get the sap) and he asked if it was possible to gatter statistics ex: sap flow (GPM, with flowmeters,, 3 or 4) from each section and precise temperature (DS1820 will do the job)

So let me recall the environment ... its in the wood far away from cell phones antennas ,, nearest city is like 150km ...oh yeah weather is cold (on average minus 32F , on winter ... and alot of snow) so cables will be underground of course ...
what else i must check, to be careful ??
Cables to each flowmeters might exceed 200m from the "sugar house"

can i use repeaters to help ,, or it will eventuality pickup too many things thus lowering the SNR to be repeated ?


thank you !!!

[Guest]

You don't need repeaters.

Look at the RS485 Transceivers at Linear Technology. These bad boys can run a long way - with proper termination and good cable.

You don't have to implement a full RS485 system to use these, you can think of them as serial line drivers and receivers.

Don't worry, be happy.

HTH - Steve H.

[SherpaDoug]

I would suggest RS485 over cheap twisted shielded pair wire. That way you can use the old "Phantom power" method to power the sensors. With RS485 one signal lead is always at 0V. So if you run the shield at supply voltage (usually +48V in the classic phantom power, but you will probably use less) then you get sensor power between the shield and a diode AND of the data lines. You only need enough local capacitance to power the sensor node for the duration of a reply transmission.

No batteries and cheap wire, whats not to like! Just label the line phantom power so some poor schmuck troubleshooting the system ten years from now won't be surprised to find +DC voltage on the shield.

P.S. I believe phantom power originated with vacuum tube amplifiers on stage microphones in the 1920's or 30's. They ran a differential audio signal with the center tap at ground and ran +48V for both plate and filament voltage on the shield. The extension to RS485 is obvious using diodes instead of a center tapped transformer to extract the power.
_________________
The search for better is endless. Instead simply find very good and get the job done.

[L.Belanger]

Thanks for the suggestions/info .. however i have thought about using shielded outdoor cat5 (Foil Twisted Pair,,UV Protected,,,1000ft for $170 U.S not bad!) so i will have plenty of wires in the future.
+i will not have to use phantom power (this technology is still used in some amplified mics,, i still have the phantom 48v switch on my mixer ^^)

i've read about RS485 protocol but a question come in mind... is it possible to have multiple devices on a single RS485 line or its only master / slave device only per cable ?

let say i use the cat5 network cable which have 4 twisted pairs wires ... on distance such as 200m,,, using AC (around 9VAC = nearly 12VDC) voltage then a bridge to convert it to DC... i know that AC voltage perform way better on distance than DC but i'm worried about the AC 60hz which will travel along with the RS485 wires. I've read that having twisted wire help to cancel each other interferences ... please correct me if i'm wrong on this.

thank you

[L.Belanger]