VLBI System Documentation

Mark IV Field System

System Setup and Tests

N. R. Vandenberg, D. B. Shaffer

NVI, Inc./GSFC

Operations Manual

NASA/Goddard Space Flight Center Version 8.2

Space Geodesy Project September 1, 1993

• 1.0 Purpose
• 2.0 Standard S/X Geodetic Frequency Sequence
• 3.0 Wideband S/X Geodetic Frequency Sequence
• 4.0 Astrometry S/X Geodetic Frequency Sequence
• 5.0 Losing Channels Gracefully
• 6.0 Track Assignment Table for Mode C
• 7.0 Track Assignment Table for Mode A
• 8.0 Cable Delay
• 8.1 Counter Setup
• 8.2 Checking Cable Sense
• 9.0 Phase Calibration Signal Injection at I.F.
• 10.0 Overall System Test
• 11.0 Checking Coherence
• 12.0 Checking Frequency
• 13.0 Clock Synchronization with GPS
• 14.0 WWV Synchronization
• 15.0 Head Position Locations

This manual describes system setup and system tests that apply to any geodetic VLBI system, Mark III, VLBA, or a combination backend. This manual also contains tables of standard set-ups and track assignments.

Refer to the Mark III Setup manual and the VLBA Setup manual for setup information that applies to your specific equipment. Those manuals cover the individual module setups and checks.

Refer to the Phase Calibration System manual for more details on the phase cal theory and monitoring. Refer to the pcalr manual for how to set up automatic monitoring.

The tables in this section give the frequencies for the standard S/X geodetic frequency sequence. Channels 1-8 are for X-band, 9-14 for S-band. The table on this page gives the RF (sky) frequencies. The table on the following pages has BBC and VC frequencies for various standard L.O. and upconverter frequencies.

Standard Frequency Sequence

Channel	Sky Frequency	Relative Frequency
X-band
01	8210.99	0
02	8220.99	10
03	8250.99	40
04	8310.99	100
05	8420.99	210
06	8500.99	290
07	8550.99	340
08	8570.99	360
S-band
9	2217.99	0
10	2222.99	5
11	2237.99	20
12	2267.99	50
13	2292.99	75
14	2302.99	85

Standard S/X Frequency Sequence

BBC and VC Frequencies

X-band
LO= upconv=	8080 479.9		7600 0		8080		8100
Channel	BBC Freq.	Patch	BBC Freq	Patch	VC Freq	Patch	VC Freq	Patch
01	610.89	A	610.99	A	130.99	1L	110.99	1L
02	620.89	A	620.99	A	140.99	1L	120.99	1L
03	650.89	A	650.99	A	170.99	1L	150.99	1L
04	710.89	A	710.99	A	230.99	1H	210.99	1L
05	820.89	A	820.99	A	340.99	1H	320.99	1H
06	900.89	A	900.99	A	420.99	1H	400.99	1H
07	950.89	A	950.99	A	470.99	1H	450.99	1H
08	970.89	A	970.99	A	490.99	1H	470.99	1H
S-band
LO= upconv=	2020 479.9		2000 500		2020		2000
Channel	BBC Freq.	Patch	BBC Freq	Patch	VC Freq	Patch	VC Freq	Patch
09	677.89	B	717.99	B	197.99	2L	217.99	2L
10	682.89	B	722.99	B	202.99	2L	222.99	2L
11	697.89	B	737.99	B	217.99	2L	237.99	2H
12	727.89	B	767.99	B	247.99	2H	267.99	2H
13	752.89	B	792.99	B	272.99	2H	292.99	2H
14	762.89	B	802.99	B	282.99	2H	302.99	2H

The tables in this section give the frequencies for the wideband S/X geodetic frequency sequence. Channels 1-8 are for X-band, 9-14 for S-band. The table on this page gives the RF (sky) frequencies. The table on the next page has lists of BBC and VC frequencies for various standard L.O. and upconverter frequencies.

Wideband Frequency Sequence

Channel	Sky Frequency	Relative Frequency
X-band
01	8212.99	0
02	8252.99	40
03	8352.99	140
04	8512.99	300
05	8732.99	520
06	8852.99	640
07	8912.99	700
08	8932.99	720
S-band
09	2220.99	0
10	2230.99	10
11	2250.99	30
12	2305.99	85
13	2340.99	120
14	2345.99	125

Wideband S/X Frequency Sequence

BBC and VC Frequencies

X-band
LO= IF3= upconv=	8080 479.9		7600 -500		8080 500.1		8100 500.1
Channel	BBC Freq	Patch	BBC Freq	Patch	VC Freq	Patch	VC Freq	Patch
01	612.89	A	612.99*	A	132.99	1L	112.99	1L
02	652.89	A	652.99*	A	172.99	1L	152.99	1L
03	752.89	A	752.99*	A	272.99	1H	252.99	1H
04	912.89	A	912.99*	A	432.99	1H	412.99	1H
05	652.99*	C	632.99	C	152.89+	3	132.89+	3
06	772.99*	C	752.99	C	272.89+	3	252.89+	3
07	832.99*	C	812.99	C	332.89+	3	312.89+	3
08	852.99*	C	832.99	C	352.89+	3	332.89+	3
	*=no upconverter				+=with IF3
S-band
LO= upconv=	2020 479.9		2000 500		2020		2000
Channel	BBC Freq	Patch	BBC Freq	Patch	VC Freq	Patch	VC Freq	Patch
09	680.89	B	720.99	B	200.99	2L	220.99	2L
10	690.89	B	730.99	B	210.99	2L	230.99	2H
11	710.89	B	750.99	B	230.99	2H	250.99	2H
12	765.89	B	805.99	B	285.99	2H	305.99	2H
13	800.89	B	840.99	B	320.99	2H	340.99	2H
14	805.89	B	845.99	B	325.99	2H	345.99	2H

The tables in this section give the frequencies for the astrometry S/X geodetic frequency sequence. Channels 1-8 are for X-band, 9-14 for S-band. The table on this page gives the RF (sky) frequencies. The table on the following pages has BBC and VC frequencies for various standard L.O. and upconverter frequencies. This frequency sequence has a subset which also forms a reasonably good frequency sequence. This subset is used for VLBA stations which have only 8 BBCs available.

Astrometry Frequency Sequence

Channel	Sky Frequency	Relative Frequency
X-band
01*	8210.99	0
02	8220.99	10
03	8250.99	40
04*	8300.99	90
05	8420.99	210
06*	8540.99	330
07	8550.99	340
08*	8570.99	360
S-band
09*	2220.99	0
10*	2230.99	10
11	2250.99	30
12*	2310.99	90
13*	2340.99	120
14	2345.99	125

*=Subset of 8 frequencies for VLBA stations

Astrometry S/X Frequency Sequence

BBC and VC Frequencies

X-band
LO= upconv=	8080 479.9		7600 0		8080		8100
Channel	BBC Freq.	Patch	BBC Freq	Patch	VC Freq	Patch	VC Freq	Patch
01	610.89	A	610.99	A	130.99	1L	110.99	1L
02	620.89	A	620.99	A	140.99	1L	120.99	1L
03	650.89	A	650.99	A	170.99	1L	150.99	1L
04	700.89	A	700.99	A	220.99	1H	200.99	1L
05	820.89	A	820.99	A	340.99	1H	320.99	1H
06	940.89	A	940.99	A	460.99	1H	440.99	1H
07	950.89	A	950.99	A	470.99	1H	450.99	1H
08	970.89	A	970.99	A	490.99	1H	470.99	1H
S-band
LO= upconv=	2020 479.9		2000 500		2020		2000
Channel	BBC Freq.	Patch	BBC Freq	Patch	VC Freq	Patch	VC Freq	Patch
09	680.89	B	720.99	B	200.99	2L	220.99	2L
10	690.89	B	730.99	B	210.99	2L	230.99	2H
11	710.89	B	750.99	B	230.99	2H	250.99	2H
12	770.89	B	810.99	B	290.99	2H	310.99	2H
13	800.89	B	840.99	B	320.99	2H	340.99	2H
14	805.89	B	845.99	B	325.99	2H	345.99	2H

This section recommends what to do if you have too many dead video convertors (Mark III) or baseband (VLBA) and cannot swap in spares to make the full 14 channels. In any case, you should consult with staff at Goddard before dropping channels to confirm your choice. This is especially important because if two stations have bad channels then we would like to ensure that both stations drop the same channels!

Now and then, there may be an experiment for which one of the stations will be forced to run with less than a full set of convertors (i.e. less than 14), despite the extra VC in the Mark III rack. In that case, there is a best choice for the video or baseband convertor(s) that should be omitted. This choice is based on the sidelobes in the delay resolution function and any loss in the precision of the group delay measurement. It is possible to drop one VC or BBC at each of X- and S-band without seriously degrading results. Not counting the loss of sensitivity from recording fewer bits, and the very real problem of an abnormal processing scheme. Hence, we do not recommend this as a long-term panacea for sick convertors!

Standard S/X: For the standard (also known as "narrow band") S/X frequency sequence, the channel to drop if there are only 13 modules available is #6, the sixth X-band channel, whose sky frequency is 8500.99 MHz. Dropping channel #3 is a plausible alternative, but don't drop both! Figures 1 and 2 show the delay resolution functions with channel #6 in and out.

If there are two bad converters, then the second one to drop is converter #11, which is the third channel in the S-band sequence, with a sky frequency of 2237.99 MHz. The delay resolution functions with and without this channel are shown in Figures 3 and 4.

The rms bandwidth actually goes up slightly with these deletions. At X-band, the highest sidelobe increases only marginally, to just over 0.60. At S-band, the sidelobe level increases from well under 0.60 to over 0.70. Thus, the choice is to delete an X-band channel first.

NOTE: Figures 1 through 4 referred to above are located at the end of this section.

Wide-band S/X: For the wide-band S/X frequency sequence, if there are only 13 converter modules, X-band channel #6 with sky frequency 8852.99 MHz is the recommended one to drop. When the wide-band sequence is being used at a VLBA station that is switching frequencies, channel #6 is not used and so dropping it would not affect the data to the VLBA stations.

If there are two bad converters, then the recommendation is to drop another X-band frequency, channel #4 with sky frequency 8512.99 MHz. This is the other X-band channel that is not recorded at VLBA stations.

Astrometry S/X: For the astrometry S/X frequency sequence, the design includes a 4-frequency subset at both X- and S-band for VLBA stations which have only 8 baseband converters. Therefore none of these four channels at either band should be dropped if there are only 13 converters available for an experiment.

With one bad converter, the channel drop is #2 at X-band, sky frequency 8220.99 MHz. If there are two bad converters, the two channels to drop are EITHER X-band #3 and #7 OR X-band #2 and S-band #11 (the third S-band channel). Either option results in approximately the same sidelobe level. If your station has poor sensitivity at S-band, then it is better to drop two channels at X-band rather than compromise the number of channels at S-band.

8.1 Counter Setup

An HP 5316A timer/counter is used to measure the delay in the cable from the antenna to the back end. The counter is normally read out automatically with the cable command by the Field System and the reading is logged. The switches on the counter should be set up as follows:

Power on

Reset/Local press to reset

Setup selection only

TI

A -> B

function

Blue button on right should be OUT

Gate time set to minimum

Attenuators, filter all set to OUT

AC/DC DC

Trigger on leading edge OUT

On the 5316A counter, all of the setup can be done remotely through a procedure that invokes the HPIB SNAP command. A possible command might be the following:

hpib=dev,AS0 BS0 TR1 AT+4 BT+4 FN2

where dev is the device name that corresponds to the address set up on the counter. Refer to the description of the ibad.ctl control file in the Control Files manual.

8.2 Checking Cable Sense

Before the experiment begins, you must check the sense of the cable cal by inserting a length of cable (or an additional connector) at the input to the ground unit.

1. Record the counter reading by using the cable command in the Field System.

2. Describe in a comment what was the physical set-up for that reading.

3. Repeat steps 1. and 2. after you insert the additional length.

4. Repeat steps 1. and 2. after you remove the additional length.

In correlating the tapes it is useful to have a weak phase calibration signal recorded with the video signal in order to measure the phase of the local oscillators in the video converters. Such a signal may be injected into the RF via a front-end phase calibrator (only the S/X receiver currently contains one) or into the IF via a pulse generator and coupler. A recommended procedure for generating phase calibration signals in the IF is the following:

The output of an HP10511A spectrum generator is to be injected into the I.F. This will look just like a phase calibrator except that it will not calibrate drift and dispersion that might occur in the front-end. The simplest arrangement is to drive the spectrum generator from one of the 5 MHz outputs from the 5 MHz distributor and couple the output into the I.F. via a broadband coupler placed between the I.F. from the front-end and the input to the rack. The coupling should be adjusted to about 1% of the total power by placing an attenuator on the output of the spectrum generator. Check the injection strength by disconnecting and reconnecting the 5 MHz to the generator and observe the signal strength in the total power detector in the I.F. distributor. In order to place the calibration signals at 10 kHz in the video, the frequency converters should be set 10 kHz below an integral multiple of 5 MHz. The presence of the 10 kHz signal should then be checked using the 10 kHz phase calibrator viewing filter (see drawing 3HA3RAS A-6203 in the Delay Calibrator section of Mark III VLBI System Description manual). 5 MHz is a little out of the specified input frequency range for the HP10511A, but appears to be quite satisfactory. This method of phase calibration injection requires that frequency sequences be chosen with 5 MHz spacing instead of the more natural 4 MHz spacing. Thus the mode A frequency sequence will cover 13 x 5 + 4 = 69 MHz instead of 56 MHz. The HP10511A will run at a 4 MHz rate if there is a stable 4 MHz available, but will not run at 1 MHz without modifying the input circuit internal to the unit.

To get a pulse rate of 1 MHz the output of a spare delay calibrator "antenna" module can be used in place of the HP10511A. For this purpose the module can be driven with a 5 MHz signal directly from the 5 MHz distributor if the output from the "ground" module of the delay calibrator is not available.

Since the output of the pulse generator may cover a broader bandwidth than the receiver I.F. the nominal power level for the calibrator of 1% should be measured by the detectors in the video converters rather than by the I.F. distributor. A 1% power level determined by the I.F. distributor may be considerably diluted when measured within a 2 MHz bandwidth.

A good test of the of the entire system is to do the following:

1. Set all the converters except #15 to the same LO frequency, (say 220.00) and set #15 to that frequency plus 10 kHz (220.01). Inject the LO output from converter #15 into the other converters' IF inputs via the 60 dB attenuator, power divider, and alternate IF distributor inputs. (Be careful the converter inputs are connected on the IF distributor patch panel to the appropriate IF subchannel outputs.) Set the attenuation in the IF distributor to 0 dB.

2. Look at the converter USB video outputs as described in Section 6.0 of the Rack Wiring subdivision. All 14 outputs should contain a strong 10 kHz signal. The 10 kHz signal in the LSB channel should be at least 20 dB weaker than the signal in the USB channel. Run PCALR in the Field System and check that the signal amplitude and phase are detected and that the phase does not change from sample to sample by more than a "reasonable" amount ("reasonable" depends on the amplitude of the signal). The phase of the signal may change from track to track.

3. Record tracks 1-14 on a tape in mode A. These tracks are the USB outputs from the 14 video converters. Look at each track with the decoder to see that the time code is properly decoded and that the error rates are acceptable. Connect the decoder DATA output to a scope and trigger the scope with the decoder FRAME SYNC output. Each frame should begin with the sync block that contains the time code (see drawing 6056 in the Blue Books. The data following the sync block should look like a 10 kHz signal that is stable in phase except for phase noise consistent with the performance specification in Section 3.0 of the video converter section of the Blue Books.

4. Repeat steps 1-3 but with the LO frequency in converter #15 set to 10 kHz below the other converter frequencies (219.99). This will put the 10 kHz signal into the LSB channels. Record and decode tracks 15-28 in mode A.

During an experiment, check periodically that the time code being written on each track can be properly decoded and that the parity and sync error rates are acceptable. A modified version of test step #3 above may be run during an experiment if there are phase cal rails at 10 kHz in the USB channels: Connect the input of the 10 kHz narrowband filter (it should be in the drawer in the tape drive) to the decoder DATA jack, and connect the filter output to a scope. Trigger the scope with the decoder FRAME SYNC output. You should be able to see a phase-stable 10 kHz signal standing well out of the noise.

L.O. coherency can be checked by substituting a separate 5 MHz input to the phase cal system or for the L.O. reference. This secondary standard could be a rubidium or cesium. Look at the 10 kHz phase cal signal on a scope triggered from either the maser or the secondary standard. Do not use INTERNAL scope trigger, use a signal referenced to the maser, e.g. from the FRAME SYNC output from the formatter. The signal might drift slowly, but should be otherwise stable.

A phase cal test performed without separate frequency standards will detect unlocked L.O.s but it will not detect malfunctioning masers. Hence, this two-standard test is mandatory. However this test will only detect a truly useless maser. The data quality will already be very poor, if not unusable, before this test can detect a problem.

The absolute frequency of the local oscillator can be checked by looking at an external source of known frequency. This will verify that the L.O. has locked up at the right frequency.

1. One method is to observe a satellite that has a downlink of known frequency. The DSCS satellites transmit at 2272.5 or 2277.5 MHz. The receiver L.O. lock points are far enough apart that just a crude check to a few MHz on the satellite frequency will verify the correct L.0. setting. To point your antenna at the satellite, you will need a current ephemeris and the program SATPS. Be aware that the DSCS satellites are not always on the air, so you might need some luck too. Also, there are other S- band satellites near the Pacific DSCS but they radiate at different frequencies.

2. An alternate method is to radiate a test signal at one of the video converter sky frequencies (sky frequency = L.O. + VC center frequency). This signal could be generated as a harmonic from a signal generator or synthesizer, driven from an external reference. For example, you could radiate 2223 MHz and expect to see it in the total power of VC10 (202.99). If you vary the frequency of the signal generator by a few hertz, then the resultant drifting signal will be seen through the phase cal filter. This method checks coherence as well as frequency.

The recommended method for synchronizing station clocks is to use a GPS timing receiver.

In order to correlate VLBI data it is necessary to know the offset between the formatter time and USNO time within a few microseconds. GPS timing receivers provide an excellent means for measuring this offset with an accuracy of better than a microsecond. This section, originally written by Richard Strand, provides a basic explanation of how to set-up the most common GPS timing receiver used at geodetic stations to measure the offset. The procedure may be slightly different for other receivers. Refer to your GPS receiver documentation for more information.

The FTS 8400 satellite timing receiver is used to measure the offset between a 1 pps signal and the GPS system 1 pps. The receiver is a stand alone computer based instrument requiring little operator intervention after initial installation and setup. It will automatically track a constellation of NAVSTAR/GPS satellites and calculate the offset.

The FTS 8400 is often rack mounted and weighs 45 lbs with batteries. It is 7 inches (178 mm) high, 18 inches (457 mm) deep and 19 inches (483 mm) wide. Power requirements are 120 VAC (90 to 130 VAC) or 240 VAC at 60 Hz (47.5 to 66 hz).

The FTS 8400 antenna is 6 inches (153 mm) high and 7 inches (178 mm) in diameter and should be mounted on the roof or other high elevation with an unobstructed horizon for optimum tracking conditions. The maximum distance the antenna can be installed from the receiver is 30 meters (using RG8 type cable) or 11 meters (using RG58 type cable).

The 8400 requires a 1 pps input reference. This is usually provided from the Maser time standard. This 1 pps signal cable will go to the FTS 8400 rear panel on the BNC port marked 1 PPS IN.

The 8400 will require an external 5 MHz frequency input from the Maser as a reference. This cable will go to the rear panel bnc marked FREQ IN.

The 8400 has a built-in keyboard and display for initialization, control and display of time, position and 1 pps offset. The keyboard consists of 16 keys which control all modes of operation and data entry. The six major functions are:

INIT Initialization data entries

CTRL Control of modes and tracking parameters

I/O Setup of I/O functions

DISP Display of measurements and computations

CALC Calculations of satellite visibility

TEST Software revision code

The top four unlabeled keys are utilized for menu selection. The CLR key returns the display to the activity message state. The ENT key terminates a data entry to allow review of data. Control parameter entry are done with numeric keys, +, -, N/W, S/E. All parameters are retained in a nonvolatile memory.

The 8400 may be turned on with the line toggle switch found on the back panel. Upon power-up the receiver will self test then search for a satellite. If the receiver has been moved to a new location in latitude or longitude by more then one-half degree, a new approximate position must be entered.

To enter approximate position:

1. Press INIT key. (displays menu)

2. Press 1. (APPOS) Approximate position entry

3. Key in LAT using the N/S key to terminate.

4. Key in LON using the W/E key to terminate.

5. Press ENT to enter this data.

6. Key in ALT in meters. The altitude should be within a few kilometers.

7. Press ENT to enter altitude data.

To enter cable delay:

1. Press INIT key. (displays menu)

2. Press CABLE.

3. Key in delay in nanoseconds. (1.5 nsecs per foot of cable)

When one satellite is acquired and lock is achieved, the data collection process begins. To see the activity messages press the CLR key. The second line of the display will show one of the following messages. In the messages, nn is the satellite number.

Message: Comments:

k SEARCHING SV nn Searches complete range of doppler/code phase. k = 0-9 progress through search activity.

xxx TRY FOR SV nn Searching specific range. xxx = PRE,EPH,ALM.

kk GET DATA SV nn Lock and collecting NAVDAT messages. kk = Ax bit sync, Bx byte sync, Cx data sync.

DATA DECODE SV nn Decoding and validating.

mm TRACKING SV nn Acquisition complete. Measurement in progress. mm = Elevation of satellite.

To display the time:

1. Press DISP. Display

2. Press 1.

To display maser 1 pps (formatter) offset to GPS:

1. Press DISP. Display

2. Press TI/FRQ (3). Time interval/frequency measurements.

3. Press TI (1). Top line displays time interval of FTS 8400 receiver 1 pps to maser 1 pps using the receiver as the reference (stop pulse).

The FTS 8400 has I/O options to interface with RS232 data lines and can track specific satellites at specific times. These and other options are described in the OPERATING MANUAL FOR FTS 8400 SATELLITE TIMING RECEIVER, part number 8400-m02 from FTS, INC. 34 Tozer road. Beverly, MA 01915. Phone (617) 927-8220.

The formatter clock must be synchronized to UTC, to an accuracy of a few microseconds or better. This section contains a description of a method whereby you can check that you're within less than a millisecond of the correct time. Using a GPS timing receiver is a much better and more accurate method for checking time. Refer to the previous section in this manual.

If a WWV receiver is available, and WWV is receivable the station time may be compared to the WWV seconds tick. The block diagram shows the appropriate set-up, and the figure shows a typical oscilloscope display.

The major uncertainty in WWV synchronization is short-wave propagation delay. For a given site, the propagation delay is generally quite stable and probably good to 200 s. The delay can be calculated by finding the great circle route distance between WWV in Fort Collins, Colorado (longitude, latitude: w=105.0, w=40.7 degrees) [or WWVH in Kauai, Hawaii (w=159.8, w=22.0 degrees)] and the station site (s, s); and then dividing by the speed of light. (We're assuming a spherical earth in what follows, which is close enough for a few hundred microseconds.) The angular distance between stations, Q, is (from spherical trig):

Q = arccos[sin(w)sin() + cos(w)cos(s)cos(w-s)]

The physical distance D between the station and WWV(H) is then

D = Q*R earth, where Rearth = 6375 Km.

Then, the propagation time delay T is

T = D/c, where c is the speed of light: 3x10⁵ km/sec.

Typical delays for WWV reception in California are 4-5 milliseconds.

The following page is a graphical picture of the Mark IIIA head position locations.