CP-1-K, 20-21 January, 1998 --------------------------- TX down 2148-2155. No signal at remote sites until 0050 when RC restarted. Stop 1600 UT (SP-GE-CP1 continued immediately). CP-2-E and CP-6-B, 23-27 March, 1998 ------------------------- Report of the CP2/CP6 at Tromso, 23-27 March 1998. M. Rietveld was co-ordinator. The CP2 and CP6 experiments were both started on time at 1000UT on 23 March 1998. Conditions were quiet. The Tromso antenna did not start moving until 10:06 UT since it was not enabled. CP6: Antenna vertical, 2 klystrons feeding whole antenna. AllX from whole antenna. A correlator error on VHF, after a crowbar at 22:06 UT on 26 March was not noticed and corrected until 23:30 UT. The data was integrated "manually" from the disk sets since INTSP did not function and time was not spent debugging it. The Tx power was calculated from the sum of the A and B klystrons in the integration program. CP-2: The ND integration/analysis was performed in real time, the latter using alt-anal. On 24 March at 13:51 UT Birger noticed that Lo2, ch 3 was wrong by several MHz. It appeared faulty and was replaced by a lab synthesizer (Hewlett Packard set to 150.0 MHz, 6dBm and locked to external 10MHz reference.) I had not checked the LO2 frequencies so this LO may well have been faulty from the start of the experiment. The colour plot of the long-pulse density data shows a discontinuous increase at about 1350 UT. Comparison with fof2 will have to be done to see if the fault was there from the beginning. As a result of this (and previous instances of LO2 failing) the operators will now do a PRINT-RECEIVER and and DISPLAY-ALL-LO2 and see if they agree whenever they write the UHF transmitter log. There is a special entry in the log now to indicate that the check was done. General: ========== The SPARC transfer was reliable throughout the experiment but there were data gaps in it caused by human error. The first occurred when the data were copied to EXABYTE using sd during the experiment. This has the effect, when certain options are not used, that the data and the directory are removed so that no more data can be transferred to that directory. The second gap occurred when the system manager inadvertently removed all the x permissions from all the directories in /data. This also means that no data could be transferred. Thankfully the data could be ddrawed from ND data tapes. I think the data saving procedures (particularly the use of sd) need revision/improvement. Data should not be automatically removed from disk unless necessary. Only one tape copy of the data exists at the time the data are removed and this is not too secure. A GUISDAP data analysis was also done of CP-2, but the first pass has a data gap due to missing SPARC data at that time. The data gap was later recovered from a ddrawed ND data tape, but the analysis has not been re-done. The NCAR-format data were saved to 6250 bpi Tapes, including the CP-2 velocity data. The remote site data were not analysed or handled by me. An attempt was made at "ddncar"-ing the data but it is not clear which version of ddraw works correctly. The one I tried (ddncar3 from HQ's machine) creates a new subdirectory for each file it writes. Colour (postscript) plots were made, and the postcript files are available on the web. Michael Rietveld (15 April 1998) CP-3-F, 27-30 April, 1998 ------------------------- Data gaps due to crowbars at 0832-0836 (28 April), 1114-1119 (28 April), 1651 (28 April), 1219 (30 April), 2251 (30 April). CP-7-G, 26-28 May, 1998 ----------------------- Scheduled start (10 UT) was delayed, initially by tests on the phase matching of the two halves of the VHF transmission lines, then by TX problems that turned out to be due to a loose cable at the UHF end of the cap-bank, causing arcing. The experiment was eventually started at 1330 UT (26 May). CP-7-G-V is experiment SP-EI-CP72-V (or SP-EI-VHF2MS-V using 35.4 kHz filters). Recent work on the VHF receiver meant that the configuration needed to be ALLX. Also, the attenuator settings needed revising. The signal path for X was set to 0 dB (previously 12 dB when it was ALLY) and the channels were adjusted by inspection of the oscilloscope signals. Channel 4 (actually using channel 3) needed 11 dB, channel 6 needed 5 dB and channel 8 needed 3 dB. Channel 8 seemed noisier than the others so it was connected to channel 1 instead, but there was little obvious improvement. Attenuator settings in the earlier version for channels 4, 6 and 8 were 14, 20 and 20 dB, respectively. Several gaps due to TX-related problems: 1709-1724 (26 May) due to a temperature sensor fault. 1732 to about 1830 (26 May) - several major breaks due to crowbars. 2220-2307 (26 May) - crowbar caused ND problems. 0251-0304 (27 May) - HV trip followed by a crowbar. 1010-1048 (27 May) - crowbar caused ND problems (no sparc transfer on restart) so exited EROS, did RT EROSCL and restarted. 2328 - 0210 (28 May) - similar sequence of events - crowbar, ND hanging, no sparc transfer happened about 3 times. Major gap 0130 - 0210 UT. 0210 UT, restarted experiment but fire alarm sounded while attempting to raise HV (which was producing crowbars at the time). Transmitter hall full of smoke - switched off everything. The fault was a burned piece of rubber at the VHF end of the cap-bank. The area was cleaned up and the switching arm re-adjusted for smoother operation. 0812 UT (28 May) restarted the experiment, RF on at 0820 UT. CP-4-B, 23-24 June, 1998 ------------------------ Parallelled cap-bank. Start 10 UT. The cable connections for channels 3 and 4 were initially wrong, and rectified at 1035 UT. Long pulse densities from the two beams were rather similar but the B-side/phased/Y beam power profile results were significantly smaller than the latter and the A-side/unphased/X pp results sometimes larger than the acf densities. CP-1-K, 17-19 August, 1998 -------------------------- Start experiment 09 UT 17 August. Sodankyla alt code signal initially in gate 2; offset ppd shifted from -415 us to -455us to centre it in gate 4. Between 0930 and 0959 UT the antennas were directed towards Andenes (POINT-GEOG 68.44, 10.0, 250.0) for tests in conjunction with the MF radar there. The antenna was again field-aligned by 10 UT, where it stayed until the end of the experiment. Calibration of peak F-region densities from the long pulse with local digisonde foF2 values suggested a value of 1.24 for the system constant, which was therefore applied in the data analysis. A power failure in Kiruna caused loss of vector velocities and of recorded raw data between 0320 and 0524 UT on 19 August. However, the ND integration program continued to run and analysed data are available on the tape. CP-2-E and CP-6-B, 21-25 September, 1998 ---------------------------------------- As the VHF antenna was phased, CP-6 was a 1-klystron operation using the A-side ('X'). Sodankyla alternating code signal was shifted towards gate 3 so offset ppd was changed from -450 to -460 to centre it in gate 4. F-region peak was clearly higher than common volume at 220 km so this was changed to 278 km at 1824 UT (21 Sept) where it stayed until the end of the experiment (:ELAN files changed also). CP-6 was integrated off-line from the ND disk sets as INTSP would not start. On VHF, the peak power reading (on RTG and from integration program) was zero. This was corrected at 1248 UT (21 Sept) by switching the 'A' and 'B' peak power input cables to the camac adc. The system temperature in the south position of CP-2 was systematically higher than the other 3 positions (which were all very similar). It changed slowly and irregularly with time, being typically 10-40K higher, generally lowest at night. The background spectra there showed a DC peak, maybe 30% higher than the normal spectrum. However, the SE and FA position showed a similar, but smaller peak, with nothing at vertical. On VHF side, ND started producing data files without 30-min eof marks after a crowbar at 1926 UT (22 Sept). This could have been noticed by checking the V-EROS console, or by noting that the 'RESTART' displayed by PR-EXP was not correct (should be the start of the next do-loop. The problem was because AUXRUN had stopped and was remedied by issuing the RT AUXRUN command. This was rectified at 0633 UT (23 Sept). From about 1300 to 1411 UT (23 Sept) eiscatt was down following a power cycle on one of the tape drives and subsequent difficulties in rebooting. Raw data were not dumped to the sparc for this period but the 1/2" copies of the relevant data sets were made on the ND. Several crowbars occurred, mainly during the first 3 days and mostly on 23 Sept. Sept 24 and 25 were very stable, transmitter-wise.