Recent Status of the Key Stone Project VLBI System(KSP)

Yasuhiro Koyama (koyama(AT)

Kashima Space Research Center
Communications Research Laboratory
893-1 Hirai, Kashima, Ibaraki 314, Japan

1. Introduction

In addition to the three VLBI stations for the Key Stone Project (KSP) at Kashima, Koganei and Miura, the fourth VLBI station at Tateyama on the tip of Boso Peninsula became ready for the VLBI observations in August, 1996 (Figure 1). Immediately after the first check-out observations at Tateyama on August 20, 1997, daily operation of the Tateyama station began on September 1, 1996. It is a notable achievement that the new VLBI station started regular observations without major problems. This was possible because the construction and the preparation of the KSP VLBI stations were well organized and carefully designed. Since the Tateyama station started participating in daily KSP VLBI experiments, these experiments have been carried out with its full 4-station-6-baseline configuration with on a daily basis. In this report, the current status of the KSP VLBI system as of the time when 10th IERS TDC Meeting was held on March 14, 1997 will be mainly described. In addition, the improvements of the KSP VLBI system since then, and the future plans for the project will also be reported in the last section.

Figure 1. Configuration of the Key Stone Project VLBI Network.

Table 1. Percentage of the successful experiments for each station
Month Kashima Koganei Miura Tateyama
September,1996 96.7 96.7 83.3 93.3
October,1996 93.5 96.8 96.8 93.5
November,1996 80.0 93.3 93.3 83.3
December,1996 90.3 96.8 100.0 100.0
January,1997 96.8 90.3 100.0 100.0
February,1997 100.0 100.0 100.0 100.0
March,1997 100.0 100.0 100.0 66.7

2.Daily Operation

The daily observations of the KSP VLBI system are performed automatically. All four stations are equipped with two engineering workstations, one of which is for the control of the observation system and the other is for continuous data acquisition including the calibration data and the system monitoring data. These workstations are controlled and organized by the central control workstations located at Koganei central station and at Kashima sub-central station. There are two central control workstations to reduce the possibility of failure of the daily experiments. Since these two central control workstations are functioning identically and independently, the observations at remote sites are still under control even if one of the two central control system has a problem. The computer network of the KSP Network is also highly reliable since there is a redundant dial-up network line between Miura and Tateyama and it acts as a backup route when one of the three network lines connecting four VLBI stations loses connection for any reason. Table 1 shows how the reliability of the KSP VLBI system has been improved. The percentage of the successful experiments are counted for each station and tabulated in Table~1. In this table, an experiment is considered to be successful for one station, if the site coordinates of the station can be estimated with sufficient reliability. The redundant central control system at Kashima started full operation in January 1997 and it improved the success percentage. Especially in February, it is noted that all experiments were successful during the month.

Every morning, the observation tapes are replaced by human operators at each station during one hour from 9:00 o'clock by Japanese Standard Time (JST). After all the observation tapes are set in the Digital Mass Storage (DMS24) system, all tapes to be used in the next observation schedule are mounted into the DIR1000 data recorder and the quality of the tapes and whether the tapes are write-enabled are investigated, and in the end, all the tapes are rewound and set in place for the next experiment. This procedure is also performed automatically, and it tells system operator at Koganei station by e-mail if there is any problem. When the system operator receives an alarm message by e-mail, the cause of the problem is investigated and then resolved, if it is possible to do so from Koganei remotely. If an on-site repair is found to be necessary, related hardware vendors and CRL staffs are notified with detailed information. For example, a PC at Tateyama station which controls the Antenna Control Unit failed to keep the internal clock, although the PC clock is usually synchronized with UTC by using a GPS receiver signal. This problem caused data loss for the Tateyama station for three consecutive experiments but the system was recovered within three days. On the other hand, this case can be considered as a demonstration of the robustness of the KSP VLBI Network in the sense that an unpredictable failure at a remote site can be recovered within three days. In addition, the overall software system is being implemented so that the same type of problem can be resolved from central control system at Koganei or at Kashima.

3. Quality of the Observed Data

Figure 2. Three indices to evaluate the data quality of Kashim11-Koganei Baseline.

Figure 3. Three indices to evaluate the data quality of Kashim11-Miura Baseline.

Starting on March 1, 1996, the VLBI observations are performed with 16 channels each of which have 2 MHz of video band width. Total speed of data recording is 64 Mbps, which is 1/4 of the maximum recording speed of the KSP VLBI system. 10 channels are assigned for X-band and the other 6 channels are assigned for S-band. Figure 2 and Figure 3 show the various aspects of the data quality of the observed data for two baselines, one is between Kashim11 and Koganei stations and the other is between Kashim11 and Miura stations. In each figure, (a) the formal error of the baseline length estimates expressed by the standard deviation, (b) root mean squared of the residual of the observed delay after the least square estimates, and (c) the ratio of the number of delay observable which are not used for the least square estimates to the total number of the data included in the Mk-3 database for each individual experiment session, are evaluated and plotted.

These figures indicates the data quality of the data obtained in the daily KSP VLBI experiments. The most notable point is the evident improvement in the formal error and RMS residual delay for both of the baselines from February 22, 1997. Before that day, the formal errors of the baseline length estimates were around 4 mm but improved to approximately 2 mm. The RMS delay was improved from the level of 60 psec to about 20 psec. This was a result of the improved short time phase stability of the phase calibration signal at Kashima station. Before the improvements, the temperature inside the receiver unit located below the 11 m antenna dish was controlled in the range 25 +/- 1 oC. But it was found that this temperature control had been causing the short time periodic variation of the temperature inside the unit. The phase of the phase calibration signal is quite sensitive to the ambient temperature and the phase variation causes error in the observed delay. If the variation is slow, the error of the time delay can be removed as the clock difference between two stations through the VLBI data analysis. If the variation is faster than the interval of the clock parameter epochs which is 1 hour for the KSP data analysis, the delay error can not be removed and it degrades the estimated results and increases the residual delay. At present, the receiver units of all four stations are not temperature controlled and the observations are performed in the time range from 00:00 to 05:30 local time when the temperature is most stable during the day. An improvement of the temperature control system of the receiver unit with slow and smooth temperature control is considered and will be implemented in the future. The phase calibration signal unit will also be improved with a small temperature control system inside the case of the unit to stabilize the phase of the signal.

The other characteristic thing of concern in Figure 2 and 3 is the high percentage of invalid data to the total number of the available data for the Miura station. This is shown in Figure 3 (c) only for the Kashim11-Miura baseline, but it is also true for the other two baselines which include the Miura station. This is caused by weak phase calibration signals recorded in specific channels at Miura station. The band-width synthesis software automatically rejects a data channel if the amplitude of the phase calibration signal recorded in the channel is below 1% of the total power of the channel. At Miura station, the amplitude of the recorded phase calibration signal is not uniform for all channels. When the system noise temperature is high at low elevation angle or due to rainy conditions, up to four observation channels are removed from the band-width synthesis software. In this case, much data becomes invalid because of false detection of side-lobe of the delay resolution function. The cause of the weak phase calibration signal was diagnosed to be a problem in the Input Interface (DFC2100) of the KSP data acquisition system. At present, solutions are being considered and the problem will be solved in a near future.

4. Recent Improvements and Changes

After the 10th IERS TDC Meeting was held on March 14, 1997, following new changes and improvements have been applied to the KSP VLBI system.

First of all, real-time VLBI test observations were performed for four consecutive days starting April 7, 1997. The observations were done using the maximum data rate of 256 Mbps for 22 hours from 10:30 local time each day. The tests were nearly totally successful and the results obtained from the real-time VLBI data were consistent with the daily tape-based VLBI experiments performed at 64 MHz prior to April 6, 1997. At present, minor problems and software developments for the automated processing are in progress. The 256 Mbps real-time VLBI experiments will be routinely started near future.

In the data analysis, a program dbupdate was revised to obtain the nutation offsets from the IAU80 model provided in the IERS EOP Bulletins and include the offsets in the Mk-3 database. These offset values are now used in the data analysis.

Lastly, the phase calibration signal unit at Tatayema station was replaced on April 8, 1997. Before it was replaced, the phase stability of the unit was insufficient. Now, the phase stability has been improved and the formal errors of the baseline lengths of the three baselines towards Tatayema are at same level as the other three baselines. The power of the phase calibration signal at Miura station was increased and since then the ratio of invalid data to the total number of available data decreased. Although the Input Interface of the Miura station must still be repaired to make the phase calibration signal level uniform among the 16 observation channels, the data quality of the Miura station is now satisfactory.

5. Future Prospects

At present, the daily VLBI experiments of the KSP VLBI system are producing daily baseline lengths and horizontal station positions with formal errors close to 2 mm. Vertical station positions are estimated with formal errors of about 10 mm. The system development of the automated data processing system for the real-time VLBI observations are in the final stage, and the daily real-time VLBI experiments will start on a daily basis very soon.

Updated on June 9, 1997. Return to CONTENTS