Coax Termination into 2-Wires 2

Keith,

A 4 - 20 mA signal is as close to DC as you can get. The signal typically represent temperature, pressure and such "0 - 10 Hz" signals. So there is no need to worry about signal degradation. The only thing you have to consider is isolation resistance and compliance voltage.

The coax cable per se does not imply that there are any HF signals involved - it was probably used because someone thought it was needed, without any further consideration.

So, all you have to do is connect your new 18 AWG wires to the coax and continue from there.

I am afraid I can’t agree with skogsgurra on this one. Whilst the signal won’t degrade, the signal-to-noise ratio could get a lot worse in a hurry if the wires are not dealt with appropriately.

Although there may be no hf signals involved (by design), 50 feet of wiring makes an excellent antenna. I would suggest that you don’t even consider replacing the coax with ordinary wire. The system evidently worked before and all you can possibly do is make it worse! You must expect the outer sheath of the coax to be carrying RF current. This will be coming from radio/tv pickup, mains wiring etc. The idea is to make the jumper from the screen to the system as short as possible. 3 inches would be hopeless. Ideally you would make a 360 degree connection to the coax by using a suitable housing. Failing that use the skill of an electrical technician to make a robust joint to the screen using solder, heat shrink sleeving, ty-raps, and what ever other mechanical support you can get to give a reliable short connection (lets say less than half an inch ideally).

The system may not respond to signals above a few tens of hertz, in theory. In practice out of band interference can make measurement equipment do undesirable things.

logbook,

The coax will act as an antenna even underground? There are only a few feet of conduit above ground at either end. Are you saying that those few feet of are enough to get the tv, etc. signals? Thank you for your input.

Keith

Keith,

Wet earth will certainly attenuate RF signals so that only frequencies below a few hundred kilohertz will penetrate very far. This is all do with the “skin effect”, the conductivity of the soil, the frequency of the wave, and so forth. A metal conduit will also attenuate ambient fields by some amount.

A few feet of cable above ground could be plenty to get pickup; perhaps I am being over-cautious, but it really depends on the application and the sensitivity of the system. What if somebody is walking past and talking on their mobile phone? Without knowing the whole setup down to the last detail, it is best to play safe and assume a problem is likely. You can never PROVE that the system is robust against such interference. All that can happen is that it can be proved inadequate when interference occurs. Life is tough for us engineers!

logbook (and Keith),

A 4-20 mA input is band-width limited (filtered, if you like). The reason is that it shall be possible to run a 4-20 mA signal all over the plant without having to worry about interference. That's why the 4-20 mA current loop was created - to NOT be sensitive to interference. And it has proven itself in hundred of thousands of installations, perhaps millions, over the last three or four decades. So, if there is no obviuos reason (find out how the other 4-20 mA loops in the plant are wired) you should not complicate life. As you said; life is tough for us engineers...

Presumably the co-ax is carrying the 4-20 mA loop with the outer braid acting as the return path. As long as this is earthed at each end the, basically DC, signal on the line won't be bothered by RF interference. 4-20 mA transmitters and receivers have filters to remove such interference.

However, if you want to replace the co-ax with any other cable, I would suggest cat 5 twisted pair. As long as your transmitter and receiver have common mode chokes at their output and input, respectively, you should have no problem at all with interference, which will be picked up by the cable in common mode. The chokes will pass the DC, differential, signal but reject any common mode signal.

Oddly enough, experience has shown that shielded twisted pairs can have worse noise problems than unshielded. The shield acts as an antenna and curents are set up in it. Since the shield is not terminated in its intrinsic impedance at each end, these curents will form standing waves in the shield which readily couple into the twisted pair. With an unshielded pair, this phenomenon cannot happen, so noise immunity can be better than with shielded pairs. In any case, whether the pair is shielded or unshielded, any points of imperfection in the twisted pair will convert the common mode signal coupled from the shield or directly from the interferer into diferential noise in the pair, hence the importance of using cat 5 or better cable. However, since this interference will be at RF, it will have no effect on the 4-20 mA loop.

On the other hand, for links using 10 base T or 100 base T, this is a real problem, as it is for DSL links.

Overall, if you want to replace the co-ax, I would suggest cat 5, shielded, twisted pair and to earth the shield at one end only.

Coleagues of mine have used this approach in extremely noisy electrical environments with considerable success.

Regards

John