Vertex Antenna Metrology System Issues and Tracking
Last Update: JeffMangum - 25 June 2011
Metrology Modes Lookup Table
- "66 0 0 0" = Tiltmeter and Linear Sensor ON
- "64 0 0 0" = Linear Sensor ON, Tiltmeter OFF
- "2 0 0 0" = Tiltmeter ON, Linear Sensor OFF
- "98 0 0 0" = Tiltmeter ON, Linear Sensor ON, Hexapod ON
- "0 0 0 0" = All metrology OFF
The following links lead to information and issue tracking for each of the ALMA Vertex antennas. This information includes:
- Metrology system calibration "constants", including AN0 , AW0 , TiltXOffset , TiltYOffset , Tiltmeter Filter Time , Tiltmeter Blanking Time , SA2 , SA3 , phi2 , phi3 , Linear sensor Offset 1A , Linear sensor Offset 1B , Linear sensor Offset 2A , Linear sensor Offset 2B , and the absolute value of the mechanically-measured non-perpendicularity (NPAE).
- Issues (both resolved and unresolved) related to metrology system refurbishment, repair, etc.
Metrology System Investigation and Calibration Routines
- dvMetroIndexMeasurement.py: Measures AN0 and AW0 by monitoring tiltmeter output from antenna positioning at four cardinal directions. Syntax is "dvMetroIndexMeasurement.py -a DV03" to run the measurement on DV03, for example.
- dvTiltmeterCal.py: Modified version of dvMetroIndexMeasurement.py which determines the tiltmeter X and Y offsets in addition to AN0 and AW0. Syntax is "dvTiltmeterCal.py -a DV04" to run the measurement on DV04, for example.
Common Issues to All Antennas
Temperature Imbalance Left/Right Fork Arm
On 2009-11-08 NickEmerson
noted the following:
On average the left side of the yoke of every antenna we've tested so far is approximately 2 degrees warmer than the right side of every yoke. Do you remember, we also saw something like this on the prototype? I think what we found was that there was warm air coming up from the pedestal through two holes in the floor of the yoke, one hole on each side. On the Left side is where ALMA cables are run, and on the right side are Vertex cables. The hole is maybe 6 inches in diameter. Vertex has enough cables to practically plug this hole (on the right), while we only have a handful so the left side hole is mostly open. So, my guess is (and I think we saw this on the prototype as well), the warm air is rising up into the left side of the yoke, but not the right. Do you think this is important? It would be easy to fix, we can just plug up the hole with some foam. If this works and the gradient disappears, I calculate that it will change NPAE, on average, by a couple of arcseconds (but this should be compensated by the linear sensors so we wouldn't notice). JeffMangum
suggested that he plug the left-side cable access hole to minimize any relative temperature gradients in the fork arms.
Tiltmeters Fail to Work at Ambient Temperatures Less Than 0 C
On 2009-11-07 noticed that on DV01 tiltmeters stop working when ambient temperature less than 0 C. Tiltmeters on DV01 are unique, but from same manufacturer as on other Vertex antennas. Not clear at this time if problem is just with tiltmeters on DV01 or common to all.
- 2009-11-08: NickEmerson believes that this is a problem in the a/d converter (software error) and not a problem with the tiltmeter itself.
- 2009-11-09: Apparently Vertex neglected to read the sign of the tiltmeter temperature sensor output. Problem has been fixed.
Residual 2*Az in Tiltmeter Y Measurements Producing Excess dS at High E
While analyzing first several all-sky pointing runs from DV06 noted that there was an excess of dS at low ZA (high El):
suggested that this is due to local deformations of the transverse near the tiltmeter which are included in the tiltmeter metrology correction but are not structural deformations that produce pointing offsets. To try to verify this theory we tried several things:
- Analyze Metrology OFF All-Sky Measurements: There was a single metrology OFF run made on the evening of 2009-11-05. Starting with the metrology ON pointing model, then realizing (with some help!) that the azimuth encoder shaft was realigned on 2009-11-07 (thus leading to a +250-ish arcsec change in IA), and allowing most terms to float (but not all at the same time), I made a best-fit to this run. Final RMS was about 1.7 arcsec, which isn't that great, but the reason for this test was to look for excess dS at high E (low ZA) with the metrology (specifically, the tiltmeter) turned off. Plot of dS versus ZA on the same scale as the all-metrology-ON plot shown above shows that there is no excess residual in dS at high E (low ZA) when the metrology systems are OFF: . This supports the theory that the residual dS at high E is due to deformations detected by the tiltmeter but not deformations which lead to pointing degradation.
- Residual 2*Az in Tiltmeter Y Response: The tiltmeter X and Y response values (strictly speaking, for those metrology measurements made when the antenna is not slewing) should in theory be well-fit with functions of the following forms:
XTilt = -AN0*cos(Az) + AW0*sin(Az) + TiltXOffset and
YTilt = -AN0*sin(Az) - AW0*cos(Az) + TiltYOffset. In practice there is a 2*Az-like response in the residuals. We are specifically concerned about the Y-residuals as those affect azimuth. Looking at these residuals for one all-sky run on DV06 we get the following result: . In fact, this residual pattern was apparent in DV05: , DV03: , DV02: , and DV01 (figure not shown...you get the idea). These residuals can be reasonably well-fit with terms like HVSA2 and HVCA2.
What is left unanswered in this analysis is the question: Why does DV06 show this excess dS at high E when the other antennas did not?
. Several possible parts to this answer:
- DV06's all-sky pointing performance is much better than the others, so this excess dS is showing up more clearly.
- DV06's excess deformation is slightly worse than the others (though only slightly, based on the residual tilt response plots shown above).