Typical established satellite transponders are often configured with channel bandwidths of between 26 and 72 MHz depending on the satellite system, although bandwidths between 5 and 120 MHz are not uncommon. This bandwidth was considered to be more than acceptable when the satellites were launched, but with the increasing demand for Internet traffic, digital TV and other digital services, operators are being forced to fill the available bandwidth to the limit. The consequence of this is that as signals occupy more of the available bandwidth they deteriorate because the transmission path, including the satellite transponder, uplink and downlink, degrades the signal. It becomes necessary to apply compensation for this degradation if the data rate is to be maintained; in order to do this effectively, the impairments must be measured.

Satellite in-orbit testing is carried out for several reasons. In its basic form it is to verify the integrity of the communications payload and the antenna platform following launch and prior to the release of the satellite to the customer. Regular checks are also carried out for the purpose of acceptance testing or anomaly resolution. Measurements can then be compared with forecasted values or previous results.

Fig. 1 Schematic of a typical satellite link.

Group Delay

One parameter that has proved difficult to measure is group delay over frequency, particularly through frequency conversion. Group delay is of prime importance in today’s communication systems.1 The requirement for distortionless transmission through a linear time invariant system is a flat amplitude response and a linear phase response. The components in a typical satellite link, shown in Figure 1, can only approximate these conditions. Group delay is a measure of the phase linearity. Flat group delay versus frequency implies linear phase. Figure 2 shows linear and parabolic group delay, which are typical of delays experienced in satellite networks. Parabolic delay is usually associated with bandpass filters found in satellite transponders and communication equipment. The sinusoidal delays are often caused by impedance mismatches in the system. Ideally, the group delay is flat, a straight line with no slope, so that all frequencies across the carrier bandwidth experience the same time lag through the link. If not, the recovered digits interfere with one another, making them difficult to distinguish and errors occur.2

Fig. 2 Group delay and the transmitted spectrum.

Microwave System Analyzer

The Aeroflex 6840 series Microwave System Analyzer (MSA) has become established as the ideal product for the measurement of group delay through frequency conversion components and circuits.1 It comprises a swept frequency-modulated source and a receiver, as shown in Figure 3, and measures group delay with the envelope or modulation delay technique. Since the group delay is derived from the modulation envelope and not the carrier frequency, the technique can be applied to measure frequency-converting networks. No external frequency converting hardware is needed because the source and receiver frequencies are independent. Figure 4 shows a typical amplitude and group delay response of a downconverter. The MSA can also carry out spectrum analysis, gain compression, third-order intercept, return loss/VSWR and cable fault location.

Fig. 3 Schematic of the MSA group delay measurement system.

Fig. 4 Measured amplitude response and group delay of a downconverter.

It is rapidly becoming the instrument of choice in the measurement of group delay and other transfer characteristics of satellite links from ground stations, either co-located or remote, through the in-orbit transponder. A set-up screen facilitates the selection of the input, output and/or conversion frequencies and levels (see Figure 5).

Fig. 5 Set-up screen for a downconverter measurement.

Transit Time

The transit time to and from a satellite can be considerable even for one in low earth orbit. For a geostationary satellite it is in the region of 250 ms. In practical terms this can mean that, since the source and receiver frequencies are synchronized, the receiver, which will have an aperture of perhaps 1 MHz, has moved beyond the received signal. It is necessary therefore to further offset the source and receiver frequencies to take account of the transit time.

The offset should be increased by

Foffset(MHz) = Sweep (MHz/ms) × Transit time (ms)

For example, the uplink (source) frequencies are 14,000 to 14,500 MHz and the downlink (receive) frequencies are 11,200 to 11,700 MHz; the satellite is in a geostationary orbit; the MSA is set to a sweep time of 10 seconds and an aperture (resolution bandwidth) of 1 or 3 MHz.

Transit time is 285 ms; sweep rate is 0.5 MHz/ms;

Foffset = 14.25 MHz.

The receiver should therefore be set to sweep between 11,185.75 and 11,700 MHz and the source to 14,000 and 14,514.25 MHz. The display will show the receive frequency range and the received frequency will be well within the resolution bandwidth.

In-orbit Measurement

Figure 6 shows the measured group delay characteristic of a satellite in geostationary orbit measured through a single ground station. Input (uplink) frequencies are 14.47 to 14.5 GHz and the output (downlink) frequencies are 12.17 to 12.2 GHz. Calibration was carried out at the input frequencies bypassing the antenna. (It is normal to calibrate at the source frequency rather than the receiver frequency to remove the delay changes inside the instrument through band switching and the frequency modulation hardware.) The setting up of the instrument and carrying out of group delay measurements are discussed in detail in Reference 3. In this case the sweep time was 10 s and the sweep rate was therefore 3 kHz/ms. The transit time offset is less than 1 MHz, so with an aperture of 3 MHz it can be ignored.

Fig. 6 Relative group delay of an in-orbit satellite transporter displayed using MiPLOT™ (courtesy of Loral Skynet).

Remote Ground Stations

This group delay test can be carried out across links where the ground stations are not co-located. The MSA acting as the source is located with a controlling PC running dedicated software at the link provider’s main station. A second MSA acting as the receiver is installed at the receiving end that could be anywhere in the world where the satellite has a transmission footprint. Using the GPIB interface to the local MSA and a serial connection via modems to the remote MSA, the instruments are configured to obtain a relative group delay measurement across the section of the link to be analyzed. The two instruments are synchronized over the frequency sweep. GPS receivers can be used at either end of the system to obtain a common time and frequency reference. Measurement data are then returned from the remote end to the local PC for review and storage of the results.


The Aeroflex 6840 series Microwave System Analyzer is the ideal single box solution for in-orbit measurement of group delay across satellite links because all units will readily cover all of the currently required bandwidths and any future increased bandwidths within a single unit. Models cover 10 MHz to 20, 24 and 46 GHz, and are priced from $67,000.


1. A. Jones and J. McManus, “The Measurement of Group Delay Using a Microwave System Analyzer,” Microwave Journal, Vol. 43, No. 8, August 2000, pp. 106–118.

2.  S. Back and M. Weigel, “Degradation of Digital Satellite Signals by Group Delay,” World Broadcast News, November 1999,

3.   “Measurement of Group Delay Using the 6840 Series Microwave System Analyzer with Option 22,” Aeroflex Application Note.

Aeroflex Inc.,
Plainview, NY
(800) 835-2352