The Effect of Twisted Pair

In a little brochure titled "The Attachment to a General Catalog for Data Communication, LAN, Peripheral Device for Computers, Industrial Computers, and Softwares - Preserved Catalog Exclusive for Cables" which has been direct-mailed to me, there is a writing, "RS-232C Transmission Cable for Long Distance, Double-Strengthened Aluminum Shield, Countermeasures for RF1/EM1, UL, CL-2 Standards." The following statement of virtues is attached to the writing, though the Japanese translation from English is a little bit difficult to read:

The length of RS-232C interface is usually limited to 15 m at most. It is necessary to use a balanced interface like RS-422 or a local modem in order to extend it over the length. It is possible, however, in case of transmission distance of less than 150 m and the speed of 9600 bps, to extend the length by using special cables of low electrostatic capacity made in U.S.A. Long distance transmission cables of CC27 series are of very low electrostatic capacity because of their insulators of low dielectric rate. As a result, weakening of waveform at risetime and falltime is low, and direct current resistance and attenuation are both little because of the conductors of AWG 34 twisted pair wires. Meanwhile, the spurious (unnecessary radiation) towards outside is controlled to the minimum by using double-strengthened aluminum shield, which means it is strong against noises from outside, too.
It is followed by the following questions and answers:
Question: Why is it possible with long distance transmission cables to extend the length over 15 m stipulated in RS-232C standard? Answer: Ordinary computer cables have some problems such as weakening of output waveform of signals and errors due to crosstalks in case of transmitting high-speed data in a long distance, because they have big mutual-capacitance (electrostatic capacity). Meanwhile, long distance transmission cables have very little electrostatic capacity and low direct current resistance due to thick conductors of AWG 24. That is why it is possible to extend it even by 10 times of the stipulated length. Question: Are they strong against noises? Answer: They are of little unnecessary radiation (spurious) because of the shields of strengthened aluminum foil to cope with RFI (Radio Frequency Interference) and EMI (Electro Magnetic Interference), and are strong against noises from outside as a result. Question: Is the conductor made of twisted pair? Answer: Of course, it is. The wires are assorted by color, which makes it easy to distinguish them in manufacturing cables.

Now, the summary of the specification of "ANSI/EIA-232" is as follows:

The maximum output voltage  +-15 V (no load)
The minimum output voltage  +-5 V (3 kOhm load)
The slew rate less than 30 V/us

The maximum input voltage  +-25 V
Input resistance   more than 3 kOhm, less than 7 kOhm
Threshold voltage
 Logic 0 (ON)   more than +3V
 Logic 1 (OFF)  less than -3 V
Effective load capacity  less than 2500 pF

The maximum transmission speed 20 kbps
The area between -3 V and +3 V is a transient area without logic fixed. There is a rule that data signals must pass through the area in 4% of nominal signal interval (3 % in case of V.28) and control signals in a second, which fix the upper limit of output impedance of the driver.

There is no relation between capacitance and weakening of waveform in ordinary cable transmission. "RS-232," on the other hand, does not match characteristic impedance of the cable and impedance of the driver or the receiver, and therefore, the rise of waveform is controlled by the product of output impedance of the driver (exactly speaking, parallel composite resistance of output impedance of the driver and input impedance of the receiver) and load capacitance including the cable. So, the total capacitance is limited at lower than 2500 pF. Accordingly, it becomes possible to use it for a long distance by using cables of low capacitance.

However, the input impedance of the receiver is ranged between 3 kOhm and 7 kOhm. So, the direct current resistance in a round-trip is only 86 Ohm which is about 3% of 3 kOhm, even if 28 AWG conductor that has 1/2.5 cross-section area of 24 AWG in this advertisement is used for 200 m. Therefore, the low direct current resistance argued in the statement of virtues does not mean very much.

It is not necessary to think too much about the shield, for the voltage amplitude is big and the slew rate is limited. It is easy to imagine that cables without shield are used in most environments, when you see for example "10 Base-T" with more strict conditions being used without shield. Regarding shields, I have seen an advertisement of another manufacture, saying "our cables without shield are easy to do terminal manipulation such as connecting, dividing, etc." It is interesting to see various ways of marketing (Note 2).

Now, the "twisted pair" stressed in this statement of virtues is cables made of two wires twisted together in order to, most importantly, reduce mutual inductance between wires and crosstalks of signals. This manufacturing method is one of the most basic technologies to cope with crosstalks. So, it may be as important as the advertisement argues.

Despite that, though, most "RS-232" connections are by ordinary cables without twisted pair and do not seem to cause any problem.

As argued in the advertisement,

are twisted pairs effective with RS-232?
Otherwise, is it just a nonsense like sales talk of street venders in Japan?

This is the question, this time.

(Note 1) IBS Japan, 1997
(Note 2) 1995-04-27 NIKKAN KOGYO SHINBUN

Return to Home