search

ska-sa / katpoint

Coordinate library for the MeerKAT project
BSD 3-Clause "New" or "Revised" License
1 stars 3 forks source link

MKAIV-1555 Fix sign of NIAO term in delay model #75

Closed ludwigschwardt closed 3 years ago

ludwigschwardt commented 3 years ago

The non-intersecting axis offset is typically +1.0 m for MeerKAT and brings the dish closer to the source at low elevations. It should therefore behave like a shorter cable and reduce the calculated delay. Achieve this by adding a minus sign to the cos(el) scale factor.

This mirrors SPAZA-288 which updated katpoint v1 in a similar way. The change to katpoint v0 behaviour should be minimal, since all dishes share the same array-level elevation and they typically also have the same NIAO value for MeerKAT. The relative delay per baseline pair is therefore identical, and only absolute delay is affected.

ludwigschwardt commented 3 years ago

Whaddaya mean? We have 100% test coverage of that module! 😜 

...
katpoint/delay.py                    118      0   100%
...
----------------------------------------------------------------
TOTAL                               3198    151    95%
----------------------------------------------------------------------
Ran 87 tests in 1.049s

OK

(Man, I miss the good old days of a 1-second test suite. It's now 20 seconds with sluggish old Astropy.)

We test the method, but only with delay models having NIAO = 0, mostly because no-one cared about it yet... The new katpoint verifies it against Calc, so things have improved 😄