0
\$\begingroup\$

Helo everyone,

I have an instrumentation cable (Twisted pair) that has aproximatelly 135nf/km capacitance and 0.6mH/km of inductance, I would like to know if I can use it as RS485 cable with 9600 baudrate speed. Calculating the impedance sqrt(L/C) it's aprox. 65ohm. I Know that higher capacitance has worst performance, but I would like to know What else that makes it not suitable in a physical way.

\$\endgroup\$
6
  • \$\begingroup\$ Is it 9600 baud or rather approx. 10M baud? You use unusual unit prefix here, please clarify. \$\endgroup\$
    – Turbo J
    Commented Nov 22 at 11:36
  • 2
    \$\begingroup\$ Do you mean 9.6 Mbaud, which is not a 'standard' speed, or 9600 baud which is? 9.6 M baud will be very constrained by length. For 9600 baud you could almost get away with a piece of wet string. \$\endgroup\$
    – Neil_UK
    Commented Nov 22 at 11:36
  • \$\begingroup\$ corrected to 9600 baudrate \$\endgroup\$ Commented Nov 22 at 11:45
  • 2
    \$\begingroup\$ Pretty much anything goes at these low speeds. \$\endgroup\$
    – Lundin
    Commented Nov 22 at 12:03
  • 2
    \$\begingroup\$ 67 ohms sounds rather low for twisted pair. Are these measured or specified figures for your cable? \$\endgroup\$
    – Neil_UK
    Commented Nov 22 at 13:46

2 Answers 2

1
\$\begingroup\$

The characteristic impedance is 67 Ω based on your numbers and, as such you need to terminate the cable at both ends with the appropriate value resistor (a resistor of 68 Ω would do).

However, the formula you used is for high frequencies (usually above 100 kHz) and you are operating at much lower frequencies. But, at 5 kHz, the wavelength is 60 km so, I don't think you'll get any extra problems due to reflections. Even if you take into account harmonics of the square wave shape of the data, I can't see any extra problems over and above getting a decent enough signal amplitude.

Another answer that gives more details about characteristic impedance.

\$\endgroup\$
3
  • \$\begingroup\$ Two 68 ohm resistors for termination is DC load of 34 ohms. Isn't RS-485 specified down to 54 ohm under normal use? \$\endgroup\$
    – Justme
    Commented Nov 22 at 13:45
  • \$\begingroup\$ @Justme TBH I haven't checked. If it is a DC spec you should also factor in the loop resistance of the cable as well and, that could easily push it over 54 ohm. I'll hold my hands up and say I wasn't aware of it. It says this on the wiki page (en.wikipedia.org/wiki/RS-485): RS-485 standard conformant drivers provide a differential output of a minimum 1.5 V across a 54-Ω load but, I don't know if that takes into account long loops? \$\endgroup\$
    – Andy aka
    Commented Nov 22 at 13:50
  • \$\begingroup\$ It simply says a driver must have strong enough output. Basically it means for 68 ohm termination resistors, your cable must have at least 40 ohms DC resistance before the 68 ohm resistor. Does not sound reasonable. It's best to stick to 120 ohm terminations for RS-485, however in many applicatios 100 ohm CAT5 is used, but they may still be terminated with 120 ohms. \$\endgroup\$
    – Justme
    Commented Nov 22 at 15:24
0
\$\begingroup\$

The low speed setting indicates that you're using a very long cable; please state approximate cable length.

Speed: In general the only reason for lowering the speed is to use lengthy cabling.
The speed that provides a stable connection depends on the actual electrical environment (not only cable parameters), the cable layout, positioning relative to disturbance/interference and length.

At least for short cables e.g. 10 meters or so
the actual physical layout of the cable might not matter very much.
RS485's "differential signaling" generally reduces interference to nothing.

Ref: https://en.wikipedia.org/wiki/RS-485

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.