Memo to: Heard Island/ATOC File Subject: Digitization at PAFB 1st week of December 1992 Date: 11 December 1992 From: Kurt Metzger Introduction David Palmer of NOAA (Miami) and I spent the first week in December at the MILS laboratory at Patrick Air Force Base digitizing data from both 1991 HIFT and the 1992 ATOC-FACT experiment. This memo describes the data sets that were digitized. ATOC-FACT There were a total of 8 shots involved in this experiement. The shots were numbered from 1 through 8. The a/d clock was 1 KHz and was derived from the IRIG-B timing track on tape. The IRIG-B track was filtered using a 1 KHz bandpass filter having a nominal Q of 8 with the filter output then being zero crossing detected to generate the needed triggering waveform. a/d channel assignments were: 0 1 second square wave from small bcsg box 1 ascension phone 22 2 ascension phone 23 3 ascension phone 24 4 ascension phone 29 5 ascension phone 30 6 IRIG-E time code The one second square wave was not locked to the 1 KHz IRIG-B derived a/d clock. This track was included to assist in tracking down any potential problems with the playback clock. The IRIG-E track can also serve as an aide in any such effort (IRIG-E was in a sense locked to the IRIG-B clock). Shot 1 was digitized 4 times. The digitizing runs for this were S1R1, S1R2, S1R3, and S1R4. The run 4 data was the one taken as "the" final run. The other shot data sets were each digitized in a single run. The file names were of the form SnR1 where n is the shot number. (All of the sample files files have an .SAM extension.) S1R4 through S8R1 all used a gain of 4 on the a/d converter plus an external gain as tabulated below. shot start (nom) ch gain dB 1 328:13:22:20 20 20 20 10 0 2 335:11:44:40 20 20 20 20 20 3 335:17:33:55 20 10 20 10 10 4 335:22:16:30 20 10 20 10 10 5 336:08:20:30 20 10 20 10 10 6 336:09:28:55 20 10 20 10 10 7 336:13:45:10 20 10 20 10 0 8 336:14:55:05 20 10 20 10 10 The 0 dB gains were non-inverting, all other external gains were inverting. The data on each channel was inspected by plotting on a single sheet: a) 16+ seconds approximately centered on the expected arrival time, b) the self normalized magnitude of a 16,384 point fft, and c) the power smoothed fft result using a 1 Hz wide sliding window. From the plots it is apparent that there are impulses in many of the data sets. Some of these are hits from the FAA radar located at the south end of PAFB getting into the playback electronics and some of these are transients recorded at Ascension. If these are deemed to be causing any problems in the data analysis it should be relatively easy to detect and null these in the data. The data plots were viewed on site and used to evaluate the results of each digitization. It was also possible to make hardcopies of a subset on site. A complete set was produced and sent to Dave when I returned to Ann Arbor. The sample files and plot files were compressed using the pkzip program and copied onto floppy disk for Dave Palmer to take back to NOAA. This same data will be placed onto a DC2000 cartridge that NOAA can read and use. HIFT A number of HIFT receptions recorded at Ascension Island on analog tape were digitized. Tapes from two recorders (numbers 1 and 3) were involved. The purpose of doing this was to acquire data from hydrophones that were not digitized in real time, and to obtain data eliminating the clock uncertainties present in the two on-line data acquistion systems. These data sets should be useful for looking at the differences in the dopplers among the Ascension phones. There were problems with the tape system used on the T5200 used to digitize the data sets which limited its ability to always read back data tapes that it had recorded. It was not clear whether the problem was with the tapes involved or the hardware. Back at Ann Arbor I was able to mount and verify all data tapes. However since such was not possible at PAFB some of the data sets were re-digitized and several attempts were made at recording them. More analog recordings are available but we once again simply ran out of time. No plans are being made to digitize additional data until it proves out that the data sets we have already acquired are of sufficient interest to justify another trip to PAFB. The sample clock differs from that used in the HIFT field systems. The field systems used 4 times the 57 Hz carrier (228 Hz). The PAFB digitized data sets used a 250 Hz clock obtained by bandpass filtering (Q=8) the IRIG-B 1000 Hz carrier and dividing by 4. Another difference is the inclusion in the data data of a 1 second square wave (independent clock) and of the sampled IRIG-E channel (100 Hz carrier). These can be used to establish the reliability of the digitization and very precise sample timing. a/d channel Recorder 1 tapes Recorder 3 tapes 0 1 sec square wave 1 sec square wave 1 phone 23 phone 19 2 phone 24 phone 21 3 phone 29 phone 28 4 phone 30 phone 22 5 phone 25 phone 26 6 IRIG-E IRIG-E The 1 second square wave was generated by one of the portable BCSG boxes and was not locked to the 1 KHz playback clock. The data channels were analog filtered using 57 Hz, Q=4 bandpass filters. The gains were adjusted on each channel and were not measured. run4r1 026:10:30:30 m? run5r1 026:13:30:30 cw run6r1 026:16:30:30 p run10r1 027:04:31:00 p run11r1 027:07:30:30 m1 run12r1 027:10:41:00 m? run13r1 027:13:31:30 cw run14r1 027:16:40:02 p run18r1 029:19:42:30 m1 run19r1 029:22:42:00 m? run20r1 030:16:42:00 p Some of the recorder 3 data sets were digitized once and recorded onto more that one tape because of problems with the tape system. Three of the recorder 3 data sets were digitized twice to insure having a copy. The times of digitization for these data sets are shown below for identification purposes. Since all tapes apprear readable it would be interesting to see how well the results for two digitization agree. They should agree quite well. run4r3 run5r3 run6r3 12/03 09:18a run6r3 12/04 11:13a run10r3 12/03 12:36p run10r3 12/04 12:49p run11r3 12/03 01:52p run11r3 12/04 02:06p run12r3 (2 copies) run13r3 (2 copies) run14r3 run15r3 run16r3 run17r3 run18r3 run19r3 run20r3 As noted in the ATOC-FACT section the FAA radar was capable of placing short spikes into the data (1 or 2 values zapped every 12 seconds). I have not surveyed the data to see if and or how these affect these data sets. The filters may have have significantly reduced the effect. The digitized data sets are to be recorded onto DC2000 tapes for transmittal to Dave Palmer. The sampled data sets will also be placed onto optical disk and into the cartidge tape HIFT archive.