|
|
| |
Technical Bulletin |
|
| |
Shielded or Un-shielded Wire?
It is interesting to note that the EIA spec for RS-422/485 does not require shielded wire. When one reads the datasheets for RS-422/485 driver/receiver ICs, they don't require shielded wire either. Most telemetry cameras send data at the glacial speed of 4800 baud, but even if the data rate were higher, UTP wire has less loss than shielded wire.
So why do some manufacturers continue to spec it? There are several possible reasons:
- Their engineers aren't very experienced with transmission. All they know is that they want to protect their precious data, and shielded wire will have lower interference. This is partially true. Yes, shielded wire will sustain less crosstalk, but the interference is down in the millivolts, and the RS-422/485 signal is 8 volts peak-to-peak. A few millivolts won't affect such a large signal. What is bad is when a field engineer grounds the drain lead at both ends and large currents flow. The magnetic coupling in this case can be significant.
- The equipment has a processor and there's a lot of stray clock-noise inside their box. The engineers don't know how to make their equipment pass the radiated emissions "smog-test", but if they use shielded wire it acts as a distributed filter. There are several ways around this: a) Use three meters of shielded wire, then convert to UTP. The emissions remaining after the three meters are negligible. b) Put in a filter at the I/O connector. Place a 1000 pF capacitor from each data conductor to chassis ground. Better yet, use Metal-Oxide Varistors. These MOV devices are around 1000pF too, but they also protect against damaging transients. c) Design the equipment right from the beginning. d) Be a stinker. Ignore the emissions. Hey, when was the last time you heard of interference from a data line causing nearby equipment to fail? [NVT is not advocating this approach!]
- The equipment has a lot of sensitive CMOS that blows out when there's a lightning hit. The solutions in the previous paragraph work here; Use three meters of wire to shunt the fast rise-time energy onto the chassis, or use a filter, capacitor, or MOV. If you're really concerned, use an off-the shelf transient protector, but connect its drain wire to the equipment chassis.
- It's the way we've always done it. This is the same argument people in our industry use to continue using coax. By the way, does your ethernet LAN still use coax?
Theory and specifications are no substitute for real-world operation. NVT has been shipping product for seventeen years. There are thousands of installations that employ UTP for their transmission of RS-422/485 data. We know of no installation that ever experienced a problem with it.
Does this meet the needs?
|
|
| |
| |
| |
|
