Advertisement
Articles
Advertisement

Testing with noise in the new world of cable

Mon, 04/30/2001 - 8:00pm
John Kenneally, Vice President, Sales and Marketing, Noise Com Inc.

They say that you cannot stop progress. Progress in the cable TV world today means many things: digital cable, DOCSIS data services (including high-speed Internet service), telephone service and programming on-demand. Delivery of these new services demands new equipment to improve traffic handling and new signal types and quality to accommodate the increased data rates.

In general, all this means digitally-modulated, QAM signals which require new cable modems, cable modem termination systems, and improved fiber optic link equipment to transmit these signals over long distances with a minimum of distortion.

Verifying the performance of this equipment will present new challenges to equipment manufacturers and test equipment vendors. The complexity and number of tests to be performed will increase greatly. So will the role of noise-based test equipment. For test purposes, the nature of digitally modulated signals is closely emulated by AWGN (Additive White Gaussian Noise). This article reviews some of its uses in cable TV testing.

CNR generation

Bit error rate (BER) vs. carrier-to-noise ratio (CNR) is a key performance parameter for any data communications system. This parameter determines how accurately data can be transmitted and received for a given channel condition (noise level) and signal type. In the upstream channel, the noise level determines the data rate and transmit frequency the modem will use to send data to the headend. BER vs. CNR is a required performance test for downstream equipment as well. The PHY-07 DOCSIS modem interoperability test calls out specific BER vs. CNR performance levels for both 64 QAM and 256 QAM signals. AWGN must be generated and combined with the signal in the correct ratio to properly evaluate BER vs. CNR.

Figure 1 represents a general procedure for setting CNR with a typical noise generator. CNR (dB) is set with the adjacent channel or other interfering signals off. According to the following procedure:

  1. Measure Cp, carrier power, at the monitor port with noise off.
  2. Adjust carrier power to desired level using the signal attenuators in the noise generator.
  3. Measure Np, noise power, at the monitor port with carrier off.
  4. Using the equation below, solve for CNR.
  5. Adjust the noise attenuators until the desired ratio is achieved. An optional filter with calibrated Noise bandwidth can be used to improve CNR accuracy for a specific frequency range.
  6. After the correct ratio is achieved, the carrier and adjacent or interfering signals can then be turned on to perform the BER test.

CNR = Cp- Np+ 10log(NBW/SBW)

Where:

Cp= Measured power of the carrier or signal (dBm or dBmV).

Np= Measured power of the noise (dBm or dBmV).

NBW = Noise bandwidth of the noise source, or calibrated filter NBW (Hz). Used to determine No.

SBW = DUT system bandwidth.

New two-way services such as data (DOCSIS) and telephony make increased demands on return path data throughput. Increased data speeds make the noise levels in the return path a more critical issue. The nature of the noise in the return path adds to the complexity of this problem. Signals in the return path are transmitted from every customer location. They are often burst transmissions, which must begin and end within their time slot allocated by the headend.

The large number of customer locations cannot be controlled nearly as carefully as the headend. Return signals from many customers are combined and amplified on their way back to the headend. Because of this, return paths can suffer ingress noise and burst noise. These differ from AWGN in that they are not necessarily white (constant level vs. frequency) or continuous. The interfering noise and signal pattern in the return path can change over a short period of time.

For test purposes, this is sometimes emulated by burst (pulsed) noise. In most cases, this type of noise interrupts signal reception, requiring the receiver to use the error correction coding in the signal to reconstruct, or the modem at the customer location to retransmit data lost during the noise burst.

NPR

Another consideration for the return path is noise and interference generated by the link transmit equipment. In modern cable TV networks, this is typically a hybrid fiber/coax link. The major noise contributor in these types of links is typically the transmit laser. The noise generated by this equipment can have a detrimental effect on CNR and C/IMN (carrier to intermodulation noise), which in turn, increases the BER of the received signal.

The extent of this effect with various loading levels (input power) must also be determined in order to define the useful dynamic range of the link. One key to getting meaningful results from this type of testing is to use a stimulus signal which stresses the UUT (unit under test) in a similar way as the signals to be used in operation.

Digitally modulated signals, spectrally and in terms of Cumulative Distribution Function (CDF), look more like noise than CW. This makes them ideal for loading cable links for testing. The standard industry test of choice for return path devices has become Noise Power Ratio (NPR). The test is described below.

The "noise mask" with notch is depicted in Figure 2. The channel or UUT, usually a transmit laser, is loaded with noise including a stop-band (notch). The non-linearities and/or noise generated by the UUT partially fills in the notch. The difference in level between the notch band and the pass-band at the output of the UUT is the NPR.

The ultimate purpose of the test is to determine the useful operating range of the channel. Figure 3 shows a typical NPR vs. loading level.

At low loading levels, the NPR is dominated by the noise floor of the channel. At high loading levels, the NPR is dominated by non-linearities and intermodulation products. From this information, it is possible for the user to determine the useful operating range of the channel.

The signals in the forward (downstream) path are changing from NTSC analog video type, which somewhat resembles a CW, to digitally modulated signals for digital cable and data services— which act more like AWGN. This trend increases the validity of NPR testing in the downstream path devices vs. CW CSO and CTB testing. Many customers for these devices are demanding these tests already, and the trend continues to move toward NPR.

Future test solutions

Cable TV is poised to become the delivery medium for home Internet and telephone as well as broadcast and movie entertainment. As demands on cable TV equipment increase as a result of increased traffic and services, performance margins will shrink. There will be a need to more closely emulate the conditions encountered in real cable TV links. Test equipment companies will need to provide solutions with the accuracy and flexibility to test all possible conditions the UUTs might be subjected to.

While the test methods discussed cover many of today's equipment manufacturers' needs, they are continually updating their designs in order to accommodate new services. They will be looking for new solutions to verify the performance of their equipment under real world conditions.

The trends toward increased use of digital signals, fiber optic links and digitizing the upstream spectrum for transmission to the headend are driving equipment manufacturers and network operators to look for new test methods that will be meaningful in predicting the performance of new equipment. This will place even more demands on the test equipment vendors to more closely emulate real-world signal degradations and interference, and to detect how much of these undesirable effects are generated by new equipment itself. In order to properly simulate the equipment under test, they are likely to turn increasingly to broadband noise testing solutions and test equipment which can directly emulate an entire spectrum of digital signals.

To guarantee quality and reliability of new services to be provided, network operators and equipment manufacturers will be increasing the amount and complexity of their testing. They must choose wisely from among the new test methods and equipment to verify performance.

E-mail: Jkenneally@noisecom.com

Topics

Advertisement

Share This Story

X
You may login with either your assigned username or your e-mail address.
The password field is case sensitive.
Loading