If you throw a pebble in a still pond, the waves will emanate from the disturbance and lap onto the shore, If you're standing at the shore and see the waves coming, you can easily see each wave and identify every peak and trough. Now suppose instead that the pond is a pool, finite in size, with cement walls. Each time a wave reaches the side of the pool, it reflects back toward the source. Eventually there will be a row of waves coming towards the edge of the pool, and a row moving away from the edge. When those reflected waves reach the other side, they again bounce off the side and rebound in the opposite direction. The result is a cacophony of motion barely resembling the original set of waves. A DMX signal is a type of wave. It is an electrical impulse that propogates down a data cable at the speed of light. The cable has a characteristic impedance defined as voltage divided by current, for an infinitely long length of cable at a given frequency. The impedance changes with the frequency of the applied signal, so you have to be careful to specify the frequency. Did I mention that this formula applies to an infinitely long length of cable? Otherwise, the impedance would vary according to who long the cable is. By using an infinitely long cable length, you avoid that issue. Impedance has two components; the magnitude, or resistive component, and the phase angle, or capacitive component (at least in the case of a shielded cable - it could be inductive in other cases.) No here's where the pebble and the pond come in. When you apply a signal to an infinitely long length of cable, in theory (because every time we try this in real life we run out of money), the signal never reaches the end of the cable. Therefore, it never reflects back down the line. It's like a very large body of water in which the waves are never reflected. But when we cut the cable to finite length, then a certain percentage of the signal is reflected when it reaches the end of the line. Then, when the reflection reaches the source, a percentage of it again reflects and bounces back the other direction. Eventually you have a mix of the original signal and a lot of reflections. It becomes very difficult for the receiver to distinguish the original data from the reflections, and you get signal errors. But when you terminate the line correctly, the calbe looks to the signal source just like an infinitely long length of cable. Instead of reflecting, the signal energy is absorbed by the terminator. The receiver sees a much cleaner signal and the errors are significantly reduced. So how do you terminate the data line correctly? A proper terminator consists of a resistor that is equal in value to the characteristic impedance of the cable. At high frequencies, the capacitive component becomes insignificant, so that all we need be concerned with is the resistive component. In DMX512/1990 Digital Data Transmission Standard published by USITT, they specify that DMX data cable should be twisted pair, low capacitance, shielded cable with a characteristic impedance of 120 ohms. That last bit of information is important because it tells us the value of the resistor we need to build a data terminator. A normal DMX data signal is transmitted as a data positive and dtat negative across a twisted pair of wires (pins 3 and 2, respectively). The receiver only responds to the difference between the two signals. This balanced data scheme helps insure the integrity of the data because any interference picked up by the data oline is likely to be picked up equally by both wires. Thus there is no differential at the receiver, and it does not pick up interference. The data line should be terminated any time there is an open data line tht is not terminated by a load. For example, if you have a system with a data splitter that splits the data between automated lights and dimming, then you should install a terminator at the end of the automated data line, and at the end of the dimmer data line. A terminator can be easily and inexpensively built by soldering a 120-ohm, 1/4-watt resistor across pins 2 and 3 of a male XLR connector. I use DMX terminators on all my data systems - they create strong data signals without any work, time or sweat! Subscribe to PLSN. Talk
Lighting at PSW's Lighting Forum,
moderated by Craig Leerman, or at the PLSN
forum. Send the author
an e-mail and no one gets hurt. Comments, questions, and insights can be directed
to rcadena@austin.rr.com. |