All Celera Motion digital encoders have quadrature outputs that are compatible with 422 line receivers. The 422 data transmission standard (ANSI TIA/EIA-422, formally) is a balanced scheme for transmitting digital data over extended distances with very good noise immunity. In this scheme, a single signal is taken by the driver D and two complementary or differential signals are generated. These are subsequently sent out over a twisted pair transmission line and received by the receiver R where they are again combined back into the original signal.
Such an arrangement does a brilliant job of eliminating common mode noise. As this noise component will be identical in sign and magnitude on both signals, the difference between these signals will remain the same. The 485 standard is similar in almost all respects with the only distinguishing factor being that it supports multiple drivers, while 422 supports only one. For the purpose of this article, the discussion is limited to the 422 standard.
Generally, it is essential that 422 driver outputs (A+ and A-) do not exceed ±6 V with respect to ground. The differential voltage between them has to be greater than ±2 V but not over ±10 V. Within the time the signal pair reaches the receiver, the differential voltage must be greater in magnitude than 200 mV (input sensitivity) to obtain valid state changes.
The function of cable termination in any data transmission scheme is to eliminate or at least minimize signal reflections. Signal reflections are caused by impedance mismatching. If a signal, traveling down a line with a certain characteristic impedance (normally around 100 Ω) encounters a different impedance at the far end, it will be reflected back to the source. This reflection subsequently meets another impedance mismatch back at the source generating additional reflections and so on. Four types of termination techniques will be discussed in this article: No Termination, Series Termination, Parallel Termination, and AC Termination.
A general criterion is that if a system does not behave like a transmission line, no termination is required. This condition is defined as a data rate of less than 200 kbps or a signal rise/fall time of more than four times the propagation delay induced by the cable. According to the 422 standard, the input resistance of the receiver will be around 4 kΩ. Given these conditions, there will be some signal reflection at the receiver end but it should not be large enough to generate invalid data.
The benefit of this unterminated technique is that it reduces the amount of current required to create a signal at the receiver – thus, minimizing the driver’s power consumption. This is by far the simplest and least expensive solution, assuming cable lengths will be short and data rates will be low.
In a step towards reducing reflections while still keeping power dissipation low, another available option is the source or series termination technique. Series resistors Z are positioned at the driver output in this arrangement.
The resistors are selected in such a way that their value as well as the driver’s output impedance matches the characteristic impedance of the transmission line. Similar to the no termination example, a reflection will still be produced at the receiver end but will encounter proper termination once it gets back to the driver removing any additional reflections. Consequently, data rates will still need to stay low.
By far the most commonly used termination method, parallel termination, comprises of a single resistor connected across the differential inputs at the receiver. “Z” is the resistor value which is selected to match the characteristic impedance of the cable as best as possible (±20%). Due to this, the cable appears purely resistive eliminating signal reflections. While higher data rates and longer cable runs are supported by this technique, the driver’s power draw is increased due to the current now passing through the resistor.
When both signal quality and power consumption are major concerns, AC Termination provides a compromise between the unterminated, series, and parallel schemes. By including a capacitor in series to the termination resistor, the DC current draw is considerably minimized while still keeping the signal reflections low.
During a state change, the capacitor behaves like a short, making the termination look like the parallel example. During steady state, the capacitor charges up and acts like an open and thus the line appears unterminated. A major drawback of this technique is the reduction of data transmission rates owing to the resulting RC time constant. The capacitance “C” should be selected so that the resulting RC time constant is low with regards to the unit interval. Consider the following example:
A multiple twisted pair cable (Belden 9831) with a typical propagation delay of 1.6 ns/ft and a characteristic impedance of 100 Ω. To select the suitable value for C, the following equation should be used:
C ≤ (round trip cable delay)/characteristic impedance
For a 100 ft-long cable, one obtains (100 ft x 2 x 1.6 ns/ft)/ 100 Ω or ≤ 3,200 pF whereas for a cable measuring just 20 ft long, the same equation gives a value of C ≤ 640 pF. Apart from this, the following rule-of-thumb can be used for choosing a maximum data rate:
RC time constant ≤ 10% of unit interval
Working the 100 ft example backwards yields a switching rate that should not go beyond 312.5 kHz.
Several amplifiers and motion controllers are designed to accept TTL level differential encoder inputs. As such, they usually have input circuitry with pull-up resistors to 5 V. Any Celera Motion encoder with 5 V outputs will be completely compatible with this kind of termination. However, a number of Celera Motion encoders have digital outputs operating at 3.3 V as they are driven directly out of an FPGA. Pulling these signals up to 5 V is not advised because of the limited current capability of these outputs. Celera Motion suggests that these outputs can be terminated using either the parallel termination or no termination techniques explained above. The datasheet has to be verified for a particular encoder in order to determine the output voltage specification.
Evidently, there is no single termination technique that will be well suited in every condition. As so often in the case, the particular application dictates the design options. Factors like desired data rate, cable length, component cost, and power constraints will play an important role in the scheme selected by users. Other methods such as Bi-Directional, Alternate-Failsafe, and Power termination were not discussed for the sake of brevity, plus they are not relevant to encoder applications. The following table provides a quick, at-a-glance comparison of the techniques discussed in this article.
This information has been sourced, reviewed and adapted from materials provided by Celera Motion.
For more information on this source, please visit Celera Motion.