How To Observe at the GBT with VLBA + Mark5A and VLBA + S2 backends and Reduce Data

Test GBT-VLBA-Mark5A / GBT-VLBA-S2 observations

Summary:

We have tested single-dish GBT observing with high temporal resolution utilizing VLBA+Mark5A and VLBA+S2 back-ends. Observing with VLBA+Mark5A does not require any additional steps during set up in comparison with standard GBT observing in the VLBA mode. Observing with VLBA+S2 requires some recabeling as well as manual simple configuration of the S2 terminal. Both back-ends have produced high quality data. Results of data processing are presented and discussed. We conclude that both Mark5A and S2 recording systems can be successfully used with the GBT in experiments which require high temporal resolution.

Introduction

Both Mark5A ans S2 back ends are standard VLBI recording systems which are available in many radio observatories around the world. This makes the use of them relatively easy and reliable. They can be utilized not only for VLBI experiments but for any kind of experiment which calls for high temporal resolution. Particularly, for pulsar observing.

We have performed a test with GBT (February 5, 2005, 16:45 -- 20:00 UT, observations of two strong pulsars with well known parameters, PSR1937+21 and Crab pulsar) in order to check reliability of use in a single dish mode of available VLBA+Mark5A and VLBA+S2 GBT back ends. We have choosen to use pulsars in this test because they provide the best possibility to fine check parameters of the system used for hight temporal resolution observing.

Reguest for observations: GBT_VLBI_DAS_test.request

Used configuration

Mark5

Back-end: VLBI DAS Mark5A
Band: S
Observing frequencies: 2100 & 2068 MHz

Polarizations: LCP & RCP
Channels: 4 (aka W, X, Y, Z) of 16 MHz each both for RCP and LCP, i.e.
- center freq. 2100 MHz:
  - 2100--2116 MHz, upper sideband (direct spectrum), W
  - 2084--2100 MHz, lower sideband (inverse spectrum), X
- center freq. 2068 MHz:
  - 2068--2084 MHz, upper sideband (direct spectrum), Y
  - 2052--2068 MHz, lower sideband (inverse spectrum), Z
Sampling: 2bit

So, we have 512 Mbit/s for pure data rate.

S2

Back-end: S2
Band: S
Observing frequencies: 2100 MHz
Polarizations: LCP
Channels: 2 (aka X, W) of 16 MHz each for only LCP, i.e.
- - 2100--2116 MHz, upper sideband (direct spectrum), X
  - 2084--2100 MHz, lower sideband (inverse spectrum), W
Sampling: 2bit

Note: we use opposite designations for the 2100MHz-sidebands for Mark5 and S2

Supported files

VLBI key file (for sched program): GBTtest.key.

Station catalog to be used with the key file. A standard station catalog is corrected in order to support writing to Mark5 disk: stations_test.dat.

As a result of compiling the key file in sched, we get the card file which was used to control the VLBA+Mark5A back-end during the observations: tgbt_05acrd.gb.

Test Data Processing

Mark5A data

Decoding data

The Mark5 data respresent itself raw stream as if it have been recording to the tapes. To investigate Mark5 data format and then decode our data we used Mark4 memo #230.3and more earlier #230.2 by Alan R. Whitney which are the same in part concerning VLBA tape format. Also, Mark5 memo #005.2 by Alan R. Whitney was used to go into Mark5A data recording. However, I'd like to make some notices:

All parity bits were removed in our data that is in contrary to info in Mark5 memo #005.2. Parity bits should be removed only in Mark5B system rather than Mark5A which we used.
All pure data were modulated by a fixed pseudo-random sequence. In memos #230.3 and #230.2 there are only words that data may, optionally, be modulated and nothing about are they usually really modulated, and what is default.
Barrel-roll was not used. In log information barrel-roll was set to auto-mode, but the header information of sequent data frames shows that there was no changes of track numbers for all bit-streams.

Digits A11, A12 & A13 were wrong in our data. There was fan-out 1:4 in our data, and these digits were wrong in all of decimated tracks, except the first ones.
For best results one should use values +3.3359, +1, -1, -3.3359 in the VLBA bit-correspondence scheme as noted in VLBA memo No.5 by L.Kogan. Unfortunately, for now it is not implemented and we used more simple +3, +1, -1, -3 scheme. However, in our subsequent processing we changed the decoding values from +/- 1 and +/- 3 to real values that correspond to the current +/- sigma levels in accordance with the technique developed by Jenet & Anderson (1998). This technique is implemented in bitstat program. It can be used for data correction as well as just for checking the percentage of bit-couples that give +3, +1, -1 or -3 values.

For decode our Mark5A data we wrote the program mk5decode. It decodes only the data recorded with mode "512-8-2", i.e pure data rate of 512 Mbit/s for 8 channels using 2-bit sampling. Small description of possible options one can get just typing "mk5decode" or see the code of mk5decode.cpp.

Thanks a lot to Walter Brisken who help me with data format especially with demodulating pure data using pseudo-random modulation pattern.

Results

Full receiver bandpass form together with interferences in our observations for both circular polarizations: LCP-band.ps, RCP-band.ps. There are based on data from scan2 (duration 50.8375 secs) with frequency resolution of about 15.26 Hz. Number of individual power spectra were averaged is 774.

The amplitude of Pcal signal in our receiver bandpass for both polarizations: LCP, RCP. The average fine structure of one of the Pcal harmonics in LCP is here. The value 0 of relative frequency represents the frequency of 2070 MHz. The number of samples were averaged is 422. The frequency resolution is 3.81 Hz.

The phase characteristic curves were built for both polarizations based on Pcal harmonics phase: LCP, RCP. There are shown together with S2 Pcal phase curves here.

During the most part of our observations we had noise-diode turned on. The switching frequency was 80 Hz, so the period is 12.5 ms. With this period we folded the noise-diode meander during of about 15.5 secs (1239 individual meander periods) for all sidebands in each polarization. Here are the examples for LCP-X (2084-2100 MHz) and RCP-Z (2052-2068 MHz). The averaged noise-diode amplitude in all frequency channels was 2.0 K. The average system temperature appeared to be of about 23 K. Anomalous significant peak is present at the rise edge of the noise-diode signal. Its fine structure are here: LCP-X, RCP-Z. The peak temperature is more that twice larger than our system temperature, i.e of about 125% (52K) for both LCP-X and RCP-Z with FWHM of about 120 ns.

Giant pulse from B1937+21 - the first one ever detected using VLBA Mark5A back-end - in all Mk5 frequency channels: W, X, Y, Z. This pulse appeared to be the strongest one with peak flux density of about 2000 Jy (150 sigma).

Dispersion measure of the B1937+21 was measured to be 71.017 pc/cm^3.

For the moment scan2, scan3 and 10 min of scan4 were processed. In scan2 we found 5 events above the threshold level of 20sigma, but all of them are false. In scan3 45 events were detected, and only 9 of them can be the true GPs. Number of false events is in good consistency with chi-squared distribution with 2 degrees of freedom.

Average profiles of regular pulses for the B1937+21 with phases of detected events: scan2 | scan3. Flux density of average regular profile for the B1937+21 was found to be of about 65 mJy.
The set of all detected events: scan2 | scan3_true | scan3_false.

There were processed of about 2.5 min of the Crab data in Mk5 records. Example of one of the Crab giant pulses in all frequency channels for both polarization: W, X, Y, Z. The values of flux densities are mainly determined by Crab Nebula background emission which was accepted as 710 Jy.

Here are profiles (lcp | rcp) of one of the GPs for the Crab Pulsar in the beginning of scan6 reconstructed in consequent 2-MHz frequency channels within the whole synthesized 64-MHz band for each circular polarization. Strong scintillation together with intrinsic spectrum changes are seen.

S2 data

The record for S2-RT was "32x4-2", mode "16x8-2" was used for playing our data back from S2 tapes to disk volume.
Description of S2 data format, available modes, user manual, etc. are here.
To decode S2 data program decoder was written. Small description of options can be found by typing "decoder". File "bit.info" should be present which describes the bits placement within data word. This bits placement is determined by used record and playback S2 modes.

Results

The form of receiver bandpass from S2 data without interferences: S2-band.

The amplitude of Pcal signal in both S2 frequency channels is here.

The phase characteristic curves of phase calibration signal for both frequency channels in S2 are presented here. Also, there are shown together with Mark5 pcal phase curves here.

The strongest GP from B1937+21 from S2 data together with Mk5 data: 2100 - 2116 MHz, 2084 - 2100 MHz. The S2 & Mk5 profiles look very close to each other.

The profiles of one of the Crab giant pulses obtained from S2 and Mk5 data: 2100 - 2116 MHz, 2084 - 2100 MHz. Differences in profile shapes between S2 and Mk5 are caused by different values of DM were used in coherent dedispersion procedure (S2: 56.725 pc/cm^3; Mk5: 56.74 pc/cm^3).

Software

mk5decode - to decode Mark5A data recorded in "512-8-2" and "512-16-2" modes;
decoder - to decode S2 data;
bitstat - to correct decoded files for bit-statistics and check the percentage of bit-couples;
see - for binary data visualization.

Please, read: If you used the programs, files or any other information from this page, we are asking you to acknowledge use of it, both by giving the web-address of this wiki-page and by referencing the paper: Kovalev, Y.Y., Ghigo, F., Kondratiev, V.I., Bartel, N., Soglasnov, V.A., Popov, M.V. 2005, ``Observing at the GBT with VLBA+Mark5A and VLBA+S2 backends'',GBT Commissioning Memo 236; available at _http://wiki.gb.nrao.edu/pub/Knowledge/GBTMemos/GBT_Mark5A_S2.pdf

Thank you!

-- YuriKovalev - 13 June 2005 -- VladKondratiev - 24 June 2005 -- VladKondratiev - 4 October 2007

Attachment:	Action:	Size:	Date:	Who:	Comment:
GBTtest.key	action	4520	23 Feb 2005 - 02:44	YuriKovalev	VLBI key file
stations_test.dat	action	44370	23 Feb 2005 - 02:47	YuriKovalev	station catalog with Mark5 disks added
tgbt_05acrd.gb	action	3477	23 Feb 2005 - 02:47	YuriKovalev	GBT card file
GBT_VLBI_DAS_test.request	action	1818	04 Mar 2005 - 16:10	VladKondratiev	Request description
mk5decode.tgz	action	63002	31 Oct 2007 - 01:30	VladKondratiev	new version of mk5decode
scan2_prof.ps	action	51224	07 Mar 2005 - 19:40	VladKondratiev	scan2 average profile of regular pulses
scan3_prof.ps	action	52518	07 Mar 2005 - 19:40	VladKondratiev	scan3 average profile of regular pulses
scan2_falseGP.ps	action	106610	07 Mar 2005 - 19:41	VladKondratiev	scan2 set of false events
scan3_trueGP.ps	action	185837	07 Mar 2005 - 19:41	VladKondratiev	scan3 set of true GPs
scan3_falseGP.ps	action	713634	07 Mar 2005 - 19:42	VladKondratiev	scan3 set of false events
p135LCP.ps	action	43304	08 Mar 2005 - 17:53	VladKondratiev	Crab GP profiles in 2-MHz frequency channels (Mk5)
p135RCP.ps	action	43295	08 Mar 2005 - 17:54	VladKondratiev	Crab GP profiles in 2-MHz freq. chan. in RCP (Mk5)
LCP-band.ps	action	15903	10 Mar 2005 - 20:11	VladKondratiev	Full receiver bandpass in LCP in the observations
RCP-band.ps	action	15992	10 Mar 2005 - 20:11	VladKondratiev	Full receiver bandpass in RCP in our observations
pcal.lcp.ps	action	18191	28 Mar 2005 - 16:40	VladKondratiev	Pcal phase pass in LCP
pcal.rcp.ps	action	18472	28 Mar 2005 - 16:40	VladKondratiev	Pcal phase pass in RCP
meander.lcp.X.ps	action	3098263	29 Mar 2005 - 23:35	VladKondratiev	Noise-diode meander in LCP (2084-2100 MHz)
meander.lcp.X.peak.ps	action	18574	29 Mar 2005 - 23:36	VladKondratiev	Noise-diode meander rise edge in LCP-X channel
meander.rcp.Z.ps	action	3109524	29 Mar 2005 - 23:36	VladKondratiev	Noise-diode meander in RCP (2052-2068 MHz)
meander.rcp.Z.peak.ps	action	18554	29 Mar 2005 - 23:37	VladKondratiev	Noise-diode meander rise edge in RCP-Z channel
decoder.tgz	action	24963	21 Mar 2005 - 20:52	VladKondratiev	program to decode S2 data
s2-doc.tgz	action	6039430	21 Mar 2005 - 20:56	VladKondratiev	s2 manual, description of s2 modes, data format
lcp.pcal.ampl.ps	action	18766	28 Mar 2005 - 16:07	VladKondratiev	Pcal amplitude in Mk5 data in LCP
rcp.pcal.ampl.ps	action	19045	28 Mar 2005 - 16:07	VladKondratiev	Pcal amplitude in Mk5 in RCP
mk5strongW.ps	action	19630	29 Mar 2005 - 00:34	VladKondratiev	The strongest GP from 1937 in Mk5, channel W
mk5strongX.ps	action	19132	29 Mar 2005 - 00:35	VladKondratiev	The strongest GP from 1937 in Mk5, channel X
mk5strongY.ps	action	20158	29 Mar 2005 - 00:35	VladKondratiev	The strongest GP from 1937 in Mk5, channel Y
mk5strongZ.ps	action	19636	29 Mar 2005 - 00:36	VladKondratiev	The strongest GP from 1937 in Mk5, channel Z
lcp.pcal.ampl.zoom.ps	action	32196	28 Mar 2005 - 16:31	VladKondratiev	Pcal harmonic fine structure
pcal.phase.ps	action	19465	28 Mar 2005 - 16:49	VladKondratiev	Mk5 & S2 pcal phase curves
s2.pcal.ampl.ps	action	17338	28 Mar 2005 - 16:52	VladKondratiev	S2 pcal amplitude
mk5_crab_W.ps	action	23432	29 Mar 2005 - 00:30	VladKondratiev	Mk5 Crab GP example, channel W
mk5_crab_X.ps	action	23472	29 Mar 2005 - 00:30	VladKondratiev	Mk5 Crab GP example, channel X
mk5_crab_Y.ps	action	23238	29 Mar 2005 - 00:31	VladKondratiev	Mk5 Crab GP example, channel Y
mk5_crab_Z.ps	action	22760	29 Mar 2005 - 00:31	VladKondratiev	Mk5 Crab GP example, channel Z
phaseS2WX_1_25.ps	action	12660	28 Mar 2005 - 18:09	VladKondratiev	S2 pcal phase curve
strongW.ps	action	20435	29 Mar 2005 - 00:33	VladKondratiev	GP 1937, S2 & Mk5 profiles, 2100 - 2116 MHz
strongX.ps	action	20881	29 Mar 2005 - 00:34	VladKondratiev	GP 1937, S2 & Mk5 profiles, 2084 - 2100 MHz
crab_s2_mk5_W.ps	action	21266	29 Mar 2005 - 00:32	VladKondratiev	Crab GP, S2 & Mk5 profiles, 2100 - 2116 MHz
crab_s2_mk5_X.ps	action	22111	29 Mar 2005 - 00:33	VladKondratiev	Crab GP, S2 & Mk5 profiles, 2084 - 2100 MHz
s2_band.ps	action	17258	28 Mar 2005 - 18:42	VladKondratiev	S2 receiver bandpass
bitstat.tgz	action	14972	28 Mar 2005 - 19:03	VladKondratiev	prg to check & correct for bit statistics of data
see.tgz	action	65845	28 Mar 2005 - 21:21	VladKondratiev	program for visualization of binary data

Topic HowToObserveReduceMark5AandS2Data . { Edit | Attach | Ref-By | Printable | Diffs | r1.21 | > | r1.20 | > | r1.19 | More }

Revision r1.21 - 31 Oct 2007 - 01:34 GMT - VladKondratiev	Content copyright © 1999-2007 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.


NRAO Home > Green Bank \| Wiki Topic: GB > Data > HowToObserveReduceMark5AandS2Data

Changes \| Index \| Contents \| Search \| Statistics \| Jump to Topic