The Progress of CVN Software Correlator and Its Application Juan - - PowerPoint PPT Presentation

The Progress of CVN Software Correlator and Its Application

Juan Zhang, Weimin Zheng, Li Tong, Lei Liu, Fengxian Tong, Yun Yu, Ping Rui East Asian VLBI Workshop 2018(EAVW2018) 4-7 September, 2018 PyeongChang, Korea

Outline

• 1. Background
• 2. The progress of CVN software correlator
• 3. The applications of CVN software correlator
• 4. Future work

Background

CVN needs a general purpose software VLBI correlator: Deep-space, geodesy and astronomy

• Deep-space probe tracking need special features:

Real-time multiple objects processing (CE-5) Fast fringe searching and accurate delay model reconstruction GPU + CPU high speed software correlator.

3

Outline

• 1. Background
• 2. The progress of CVN software correlator
• 3. The applications of CVN software correlator
• 4. Future work

CVN software correlator structure

Station Shanghai Station Urmuqi Station Beijing Station Kunming Data preprocess Fringe searcher PCAL extractor Software correlation Output

NFS NFS

Spacecraft delay model reconstruction Postprocess &

• rbit

determination Correlator configure file Job file

5

CVN software correlator upgrade

 Update for the lunar mission:  phase center  2 phase centers for two-probe same beam VLBI tracking  Single object fringe search  4 objects fringe search GPU+ local correlation  Phase reference VLBI FITS-IDI  Specifications:  Input format: Mark4/Mark5B/VDIF  Output format: CVN/FITS-IDI/MK4  Graphical user interface  Applications  Probe phase referenced VLBI  IVS data processing MK4

6

Imaging results of DiFX and CVN software correlator

Image of J0856+2111. (a) DiFX; (b) CVN SCORR Image of 4C39.25 . CVN scorr(red line), DiFX (black line)

7

Geodesy upgrade: output format CVN--> Mk4

• Platform:192 CPU cores cluster
• 6 stations * 2.0Gbps/station
• Convert output from CVN to Mk4 format
• Fringe fitting with fourfit
• Comparison with DiFX output using same observation sets

(k14349, cn1502, apsg38)

– Multi band residual delay (rate), SNR – Multi band total delay (rate), SNR

8

Difference between DIFX and SCORR:  S band MBD < 10 ps  X band MBD < 3 ps  SNR < 0.5%

K14349 cn1502 apsg38 Data sources IVS CVN IVS Use same delay model Y N N S band MBD Difference < 10 ps Y Y Y X band MBD Difference < 3 ps Y Y Y SNR Difference < 0.5% Y Y Y

Difference between DIFX & CVN software correlator

9

• IVS observation k14349
• Ny, Sh, Ts, Wz
• 16IFs,8MHz/IF
• 30scan

Compare with K5:k14349

Comparison Results：

S-band X-band Delay <80 ps <13 ps Delay rate <1 ps/s <0.5 ps/s

Delay compari riso son，X X band Confirms the two correlator results are consistent.

10

Outline

• 1. Background
• 2. The progress of CVN software correlator
• 3. The applications of CVN software correlator
• 4. Future work

Phase-reference VLBI results of CE-3 Lander

• Target: CE-3 Lander, 2016-04-28
• Calibrator: 1936-155, 0.619JY, <2.5°
• Stations: Kv (SEJONG), Kb(KASHIM34), Ho(HOBART26),

Cd(CEDUNA), Hh(HARTRAO)

• Valid Session Length： ~2h

12

 Results difference Phase Reference – NASA LROC: ~122m (0.0040, -0.0003) Deg Phase Reference – CE-3 mission: ~147m (0.0048, 0.0005) Deg  Results keep consistent with each other well  Need more observations to verify

RA (deg) Dec (deg) Phase Reference

• 19.5076

44.1211 NASA LROC

• 19.5116

44.1214 CVN Lunar mission

• 19.5124

44.1206

CE-3 lander position results by 3 methods

13

Observation of GPS – 126b

• Cd-Ho, Ceduna30m &

Hobart26m

• Residual time delay ~2ns ,

~7m position accuracy

• Session length ~2h
• Bw=16MHz
• 2016-05-05

Residual time delay VS. time

14

Observation of GPS-PG28

• SHANGHAI & URUMQI , Session length ~1.6h, Bw=8MHz
• 2018-05-11

Auto-Correlation Spectral of SH & UR Cross-Correlation Spectral

15

• 4. Future work
• Lunar and deep space exploration mission
• Geodesy and astronomy data processing

CVN, IVS(VGOS), EVAN, AOV VGOS broadband ability

16

Thanks！