The Effect of SQUINT on Telemetry Download

Performance - AMSAT Oscar 40 S2 Downlink

by Colin Hurst VK5HI

Introduction

This article contains the preliminary results of a downlink evaluation that the author has been conducting on AMSAT Oscar – 40 (AO 40). The initial request to embark on this exercise was made by two members of the AO 40 Command Team, Graham Ratcliff VK5AGR and James Miller G3RUH. During the early life of AO 40, the command team quickly identified that at certain angles of SQUINT while using the S2 downlink, decoding telemetry was more difficult. Their request was to develop a S2 downlink performance profile, which would identify optimal squint operation and potentially provide an additional tool in the determination of AO40’s orbital attitude.

Conclusions

A number of simple conclusions can be drawn from the work carried out to date.

1. Optimal error free downlink performance using the S2 downlink for telemetry captures occurs at squint angles less than 35 degrees.

2. The results of this survey albeit specifically focusing on Error Free telemetry capture, do also provide an indicative guide on what performance you can expect through the transponder for SSB voice communication, when the S2 system is employed as the downlink.

3. The downlink performance profile is directly attributable to the following factors.

Shielding of the S2 helical by the other antennae on the spacecraft platform and

Spin rate.

Quantification of Results

For this evaluation to be credible, only telemetry downloaded from a specific QTH could be used for the purpose of quantitative analysis. A group of stations within a short distance of each other could assist with the collection of data, to spread the load. Consequently when data is recorded from a point source the results presented can have a number of discontinuities. In time, these discontinuities “humps and depressions” will be smoothed out, as more telemetry is acquired. For example, at this point in time, little telemetry has been obtained at very low squint angles. This survey must be a long-term exercise to ensure sampling of telemetry for all the permutations possible with an elliptical orbit, namely high squint and low range, high squint and high range, but to name a few. To date the evaluation comprises in excess of 50 orbits, 35,000 CRC’d A Blocks and megabytes of “filtered” telemetry.

Factors that have influenced the results have included:

The initial telemetry was obtained using an experimental 8 turn helical; hence range and squint significantly influenced the quality of the downlink performance. A 750-millimetre dish replaced the 8 turns helical to improve station performance; consequently the influence of range became less significant. Finally a 0.7 dB NF preamplifier was added to optimise the receive performance (measured system temperature of 110 degrees Kelvin).

Also influencing the results are “real world” factors that are out of the authors control,

Microwave oven interference (neighbours).

“Black hole attenuator” from 300 degrees to 310 degrees azimuth (trees).

Despite these factors and the relatively small telemetry sample, tangible results have been obtained.

Squint Angle

At this point we should quantify the term Squint angle.

Squint angle is the angle subtended by the main axis of the satellite antenna and the observer. At a Squint of zero degrees the observer’s antenna and the spacecraft’s antenna are optimally aligned, providing maximum gain and minimal spin modulation effects. A squint angle of 90 degrees has the observer looking side-on to the satellite antenna (side lobes only).

Results – Using CRC’d A Blocks

The term CRC’d used throughout this article refers the past tense of CRC (Cyclic Redundancy Check), viz: Cyclic Redundancy Check-summed. A detailed explanation of the CRC is included in the AMSAT PHASE 3D TELEMETRY document. (Refer References). The CRC characters (2 bytes) are appended to each 512 Byte telemetry block. Testing with the CRC characters provides confirmation of an error free telemetry block.

On AO40 the telemetered A Blocks account for 50% of all telemetered blocks. The general download sequence is for blocks A, E, A, K, A, L, A, M, A, N repeated; however when the command stations are accessing AO40 the telemetry defaults to A blocks only.

In the first instance a simple performance measure was evaluated, by comparing CRC’d A Blocks to the Squint Angle, as shown in Figure 1, where the CRC’d A blocks are shown on a logarithmic scale.

In presenting these results we need to define the term downlink performance lobe. This term should not be interpreted as, nor confused with the actual side lobes of the S2 helical antenna.

Figure 1

Figure 1 clearly identifies a main downlink performance lobe and a secondary downlink performance lobe. The main lobe drops off very quickly around 35 degrees squint and the secondary lobe peaks around 62/63 degrees squint.

This simplistic approach plots CRC’d A Blocks against Squint Angle in degrees. However this method of reporting is flawed in that the trend is “biased” towards lower squint angles due to the operational requirements of AO-40. The previous statement borders on the paradoxical, however one of the mission objectives for AO-40 is to maintain an Alon/Alat attitude of 0,0 subject to housekeeping manoeuvres for eclipse periods etc. Therefore at a nominal (or close to nominal) attitude the number of telemetry blocks downloaded per orbit equates to hours at low squint angles and minutes at higher squint angles, hence a “bias” towards lower squint angles.

This bias was notably evident prior to AO40 commencing its magnet torque manoeuvres towards Alon 270 and Alat 0 for the Arc Jet Manoeuvres. In this instance those manoeuvres afforded the author the opportunity to analyse downlink performance through the range of squint angles up to and including 120 degrees. Consequently, this torque activity has resulted in only a small number of telemetry blocks at very low squint angles, (less than 10 degrees), and this can be noted in Figures 1 and 4. That small sample in the low squint region has not unduly influenced the final results presented.

Having recognised that the above method was flawed an improved analytical approach was researched in order to identify performance during the periods of non-error free telemetry reception. The non-error free periods are clearly identified in Figure 2. Figure 2 highlights the downlink performance on what would be defined as a good error free pass.

On this particular orbit, in the range of squint angle from 33 degrees to 53 degrees, very few CRC’d A blocks were obtained, despite the range distance being of the order, low to medium. Figure 2 also confirms the requirement mentioned previously that telemetry capture and integration over an extended time is required to ensure a spread of results that cover the combinations of parameters.

After researching potential techniques, the author adopted an approach suggested by James Miller G3RUH, which was to test for Bit Error Rates.

Figure 2.

Results – Using Bit Error Rate

Following on from James’ suggestion a little research using a Hex Editor to evaluate the telemetered A Blocks and with reference to the AMSAT PHASE 3D TELEMETRY Specification a method to check Bit Error rates was derived.

With reference to Figure 3, which shows the Replay Dialog of an A Block using the P3T AO40 Telemetry Program written by Stacey Mills W4SM.

Each telemetry block from AO 40 consists of 512 bytes plus a 2 byte CRC (Cyclic Redundancy Check). The first 256 bytes of each A Block are identical except for the UTC time banner (highlighted in red by the author), with the qualification that no uplink commands are uploaded during the orbit under analysis. Analysing these 256 bytes provides us access to 50% of all blocks of 50% of bytes per A block, which equates to 25% of all downloaded telemetry.

Figure 3.

Further analysis indicated that filtering of totally corrupted telemetry blocks was necessary to provide improved accuracy. Hex Bytes #1A9 [seconds], #1AA [minutes] and #1AB [hours] (also highlighted in red) in each block are identical to those displayed in the UTC time banner. Filtering of each block is made on the basis that

If the respective Hours, Minutes and Seconds do not check in the block

then negate that block from analysis.

This filtering also ensured that the Mean Anomaly byte #1A5 would more likely be valid. The MA byte was used to cross reference the corresponding MA versus Squint from a tracking program, modified to output a comma-delimited file.

A suite of software was written to perform the Bit Error Analysis and output results as comma-delimited files for importation into a spreadsheet / graphing program.

Figure 4 details the raw data obtained from the Bit Error Analysis.

Plotted against the Squint Angle in degrees on a logarithmic scale are Total Bits Assessed and Error Bits

Figure 4.

For each item the raw values are shown with a superimposed weighted average. The graph covers the range of squints, 0 to 120 degrees. As would be expected we note increasing error rates with increasing squint angles.

In order to display this information as performance profiles the performance charts Figures 5 and 6 were developed.

Figures 5 shows the Error Free Profile whereas Figure 6 defines the Non-Error Free Profile. Both graphs are Error Rates (logarithmic scales) versus squint angle. The scales for each graph show % error bit per total bit per degree of squint.

Figure 5.

From this chart it can be readily identified that for the Error Bits / Total CRC’d Bits curve from a squint of 0 to 25 degrees we have a slow roll-off in performance. From 25 degrees onwards to 45 degrees we drop into a significant trough There is a secondary lobe around 62 degrees.

Figure 6.

From figure 6, we note a similar performance curve to that of Figure 5. The main difference is an improved trail of data through to a squint of 120 degrees.

Results – Moderated

If we moderate the values displayed in Figures 5 and 6 we are able to produce the Downlink Performance Curves as shown in Figure 7.

The curves clearly identify equivalent performance out to a squint angle of 35 degrees, at which point the error free performance drops off markedly. Hence we can conclude that optimal error free telemetry capture is available in the range 0 to 35 degrees squint angle.

We can also conclude that error free telemetry is reliably available in the ranges 0 to 40 and 55 to 70 degrees squint angle.

Although these results have been compiled using telemetry utilising the S2 Middle Beacon, which has a power output +10 db relative to the General Beacon, they should be analogous to that required for SSB voice communication through the transponder, on the basis that operators maintain their output signal at a level no greater than the GB (–10dB to MB). Without delving deep into link budget theory, but considering the pluses and minuses of the key factors, the S2 Middle Beacon is +10 dB relative to the General Beacon, the S/N ratio for error free telemetry, the required S/N for SSB reception, etc the downlink performance profile for SSB voice communication would in the first instance approximate to that of the Non-Error Free curve in Figure 7.

In the absence of any other performance profile it would be the ideal starting point for the serious satellite communicator planning to maximise his operating window, when AO-40 is on the DX horizon.

Figure 7.

AO 40 S2 Helical Antenna - Observations

During the capture of the telemetry a mental note was made of the actual signals at differing squint angles. Over time a pattern emerged that clearly suggested that the location of the S2 antenna on AO40’s platform relative to the other antenna and the spin rate were the two main factors affecting downlink performance. The net result of these two items is varying degrees of shielding of the S2 antenna to the observer and the rate of change of shielding. The following two images, Figures 8 and 9 were obtained from the Amsat-DL site, and have been suitably edited to show the antennae.

The complete suite of images is available at www.amsat-dl.org/launch/index.html

Figure 8.

This image shows a close up view of the S2 Helical, adjacent to one of folded dipoles of the V Band array. The 5 turn helical is the slender black column.

Figure 9.

This image shows the S2 helical relative to all the other antennae on AO-40.

When you study these images you can readily visualise the levels of shielding that take place per revolution of spin, at differing squint angles. The most significant obstacles are the 400N Engine, the S1 Band dish (top right), the L Band dish (foreground) and the immediately adjacent V Band dipole. Excellent diagrams are included in the AMSAT Journal, November 2000 that will assist with the identification of the items shown in Figures 8 and 9.

The effect of the spin rate can be very interesting. Listening to Orbit 334 (22^nd July 2001) when the squint angle was in the range 60 to 65 degrees, it was observed that all the A Blocks failed to CRC checksum whereas all the K, L, M and E blocks passed the CRC checksum. The spin modulation was such that the deepest of the fades from the shielding occurred in the middle of the A Block and was of sufficient duration to lose data lock. Each telemetry block is nominally 13.4 seconds in duration. For each A Block to fail the spin period equates to 26.8 seconds or 60 / 26.8 = 2.24 rpm, assuming the block failed at the same byte position. The telemetered spin rate for Orbit 334 was 2.2 rpm. Spin rate determination can also be accurately made using a stopwatch and listening to the characteristic spin modulation. Just record the time duration for 10 spin cycles and divide by 10.

Future / Ongoing Analysis

Suffice to say, telemetry capture will continue to obtain a greater sample to smooth out the “humps and depressions” especially in the lower squint region.

The survey to date has focussed on ascertaining a performance profile exclusive of variables such as range distance, “black hole” influence etc.

Time permitting, an analysis that groups error data as a range of distances eg: 10,001 to 20,000 kilometres plotted against squint will be undertaken to identify whether unique profiles can be established on the basis of range distance.

References

AMSAT PHASE 3D TELEMETRY:

Peter Guelzow DB2OS

Karl Meinzer DJ4ZC

James Miller G3RUH

Stacey Mills W4SM

Version Release 1.7, 15th March 2001.

Download available at www.amsat-dl.org/p3d/tlmspec.txt

AMSAT JOURNAL. Special Phase 3D Launch Issue. November 2000.

VK5HI Station Profile – S2 Reception

750 mm Solid Aluminium Dish

3 turn Helix feed (7.3 degree pitch)

DB6NT MKU 232 A2 preamplifier (mounted direct to dish feed)

SSB UEK-13 Receiving Converter (located indoors).

FT736R Receiver.

Measured system temperature 110 degrees K.

VK5HI Station Profile – Telemetry Decoding

G3RUH Modem feeding to P3T Decoding Software (W4SM)

AO40Rcv Version 1.40 Software Demodulator (AE4JY).

Both decoding modem systems run in parallel.

Up / Down microphone control (FT736R), connected to either modem.

Colin Hurst VK5HI (email vk5hi@amsat.org)

29^th July 2001.