|
Archives
By Alexander Woerner, Rohde & Schwarz
As new services place ever increasing bandwidth demands on cable?s hybrid fiber/coax (HFC) networks, operators are increasingly turning to quadrature amplitude modulation (QAM) as the preferred digital modulation format for cable transmissions1, 2. Because of the high data throughput requirements of HDTV, cable modem and telephony services, 256 QAM with its higher data rate of 38.81 Mbps is usually preferred over 64 QAM, which offers only 26.97 Mbps throughput. With the increased data rate also comes increased requirements concerning RF signal quality and its measurement.
Modulation error ratio (MER)
MER is the single best value to describe the quality of a QAM signal in terms of distortions. It?s also called the ?signal-to-noise ratio (S/N) of the QAM world? to be compared with an NTSC signal. From a single number in dB one can discover how far away the digital cliff is. At this threshold or with even more degradations, the forward error correction (FEC) is no longer able to compensate for broken bits in the data stream, and errors start sneaking into the services, such that TV pictures will become blocky or frozen, telephone calls will be lost and data downloads interrupted.
With a 64 QAM signal, the cliff effect starts around 22 dB, and with 256 QAM around 28 dB. The mentioned MER values are absolute minimums to receive, demodulate and decode a QAM signal with a receiver or set-top box. At this point the sensitivity is so high, that a single dB more or less separates a totally perfect service from an almost unusable one. Therefore a margin of around at least 3 dB should be added on top of those values to leave enough room for unexpected noise.
These MER minimums are for the end of the cable network. The QAM signal usually runs a long way from its insertion point at a modulator in the headend through a cable network with several line amplifiers, optical transmission links and other interfaces, where more degradation has to be accounted for. Along the way the QAM signal quality decreases through noise, ingress, nonlinearities, laser clipping and other detrimental effects. That is basically the reason why the MER at the cable headend needs to be substantially higher, so that the resulting MER at the customer site still has enough margin above the minimum required signal quality for proper decoding. A typical QAM modulator nowadays provides an MER greater than 40 dB.
Headend testing
To ensure that such high MER values in excess of 40 dB are maintained, it is crucial to have a test set capable of measuring within the same range. While an MER of around 35 dB is more than sufficient at the customer premises for reliable decoding, the same MER at a cable headend won?t provide enough margins for all potential impairments within a cable network. Frequent severe service losses can be expected if MER is not measured with high enough margin at the headend.
The limit of any MER measurement is given by the inherent degradation of the signal itself within the analyzer. A very sophisticated internal RF design with extreme low linear and nonlinear distortions is required to obtain the highest possible MER reading far in excess of 40 dB. Measures to achieve this include low phase noise in the local oscillator (LO), very precise digital signal processing (DSP) using field programmable gate array (FPGA) technologies and root cosine Nyquist roll-off filters. This can be topped by excellent external and internal shielding, whereby not only every separate module is completely enclosed in an RF chamber, but also partial circuitries on single cards are separated by shielding walls bent around the components and thus ensuring that crosstalk effects are limited as much as possible.
If the MER at the output of the cable headend is monitored carefully at and above 40 dB, any degradation in the modulator or in the headend itself can be detected early enough to initiate remedial action, while still leaving enough signal quality margins for uninterrupted service enjoyment by customers.
Eye monitoring
The constellation diagram for visual identification of impairments is limited to static or slow?but not too slow?changing effects. Periodically occurring interference cannot be detected for instance. Furthermore it shows only a small part of the actually present in-phase (I) and quadrature (Q) data, because it records and displays samples only within a narrow time window. Numerous samples are not detected at all, causing large time gaps.
A technique called ?eye monitoring? has been developed to overcome the limitations, where I/Q data of numerous consecutive constellation diagrams is computed in the following way: The I component of each of these diagrams is shifted by 90 degrees and then displayed in vertical direction together with the Q component over. The result is an I/Q projection of the constellation diagram, basically a constellation diagram versus time, allowing reliable detection of periodic interference even if the variations are only very slow or the impairments are very short in nature.
A large number of detectors operating in parallel is required to permanently monitor and store the entire range of I/Q samples. Interference with very short pulses of only 10 ns, for instance, can be recorded even if just a single symbol is impaired in transmission. Neither bit error rate (BER) nor MER measurements is able to detect those short interferences, because they do not significantly contribute to the results.
Figure 1 shows an eye monitoring diagram of an almost undistorted 64 QAM signal of about 20 seconds. Each of the eight horizontal lines represents the discrete I and Q components of all symbols as they progressed over time during the 20-second measurement period.
Detecting distorted signals
An eye monitoring diagram of a distorted signal is shown in Figure 2. This sample was taken at a headend, when cable modem service was heavily affected and customers reported interruptions and loss of downstream Internet service. Neither constellation diagram nor MER or BER measurements indicated any kind of impairment. Every 1.5 seconds very short periodic interferences created vertical lines throughout the whole diagram, at which times the I and Q values were rather continuously spread over the whole range than fixed values. A short-term constellation diagram, if available, would have looked very cloudy and distorted.
Eye monitoring differs from the familiar eye diagram in its clear time assignment of the measured I/Q samples. Both diagrams graphically display the eye height, but only eye monitoring allows continuous detection. It can consequently be interpreted as a follow-on development of the eye diagram.
MPEG-2 protocol analysis
Whatever service is transported using QAM, an additional look into the data may give a different view into distortions and can unveil even more errors. The demodulated signal, usually an MPEG-2 transport stream, can be checked for protocol errors to ETSI TR 101 2903. This includes checks for synchronization, packet decoding errors, content identification, presence of service information (PSIP) and tests for the program clock reference (PCR) and presentation time stamp (PTS). Automated real-time video quality monitoring is beyond the scope of this standard, but becomes possible too, if video data is available in-the-clear (unencrypted).
An example in Figure 3 shows some typical errors marked by an asterisk (*). The transport error indicates packets having more errors than FEC can correct, while the continuity count error (?CONT COUNT?) shows errors in the sequence of packets, such as lost or mixed up packets. Both error types are in conjunction with an impaired QAM transmission at a given BER before FEC of 4.9E-04, which is above the quasi error free threshold (QEF) of 2E-04.
Other errors in this sample concerning PCR time stamps and their accuracy may affect the ability to properly decode the video signal, because PCR provides the essential synchronization time basis between decoder and encoder. PCR values originally are created upon encoding of each single video signal and typically recalculated at a transmission link to compensate for effects of remultiplexing and network delays. Therefore any errors may be caused either at the studio or anywhere later during signal transport. To locate the exact point of error insertion, transport stream monitoring has to be applied at different test points within the cable network.
Outlook
The ever-increasing demand for more bandwidth to deploy individualized services challenges both manufacturers and cable operators at the same time. To provide 256 QAM in a cable system requires a very clean design of the cable network, modulators operating far beyond an MER of 40 dB and test equipment that is capable of measuring within that range. Tools like eye monitoring and transport stream protocol analysis provide essential help during periods of troubleshooting and assistance to maintain the superior quality of service for customers.
Alexander Woerner is a manager of market development for Rohde & Schwarz Inc. Email him at .
Did this article help you? Email .
References
- ANSI/SCTE 07-2000 (Formerly SCTE DVS 031), Digital Video Transmission Standard for Cable Television.
- ITU-T Recommendation J.83 Annex B (1995), Digital Multi-Programme Systems for Television Sound and Data Services for Cable Distribution.
- ETSI TR 101 290 V1.2.1 (2001-05), Digital Video Broadcasting (DVB), Measurement guidelines for DVB systems, European Broadcasting Union.
Bottom Line
Pristine Plant Needed
The ever-increasing demand for more bandwidth to deploy individualized services challenges both manufacturers and cable operators at the same time. To provide 256 QAM in a cable system requires a very clean design of the cable network, modulators operating far beyond an MER of 40 dB and test equipment that is capable of measuring within that range. Tools like eye monitoring and transport stream protocol analysis provide essential help during periods of troubleshooting and assistance to maintain superior quality of service for customers.
Figure 1: Eye Monitoring of a 64 QAM Signal With Good MER
Figure 2: Eye Monitoring With Distortions Every 1.5 Seconds
Figure 3: Transport Stream Errors to TR 101 290
 Back to December 2003 Issue
Access Intelligence's CABLE GROUP
Communications Technology | CableFAX Daily | CableFAX's CableWORLD | CT's Pipeline
CableFAX Magazine | CableFAX databriefs | Broadband Leaders Retreat | CableFAX Leaders Retreat
|
