Summary

• Transit inclinometry is affected by the elevation brake setting.
• However, there is no sign of the broadening of the 'transit effect' seen in recent transit tracking datsets.
• We attempted to configure the elevation brake so as to produce constant additional drag on the antenna during tracking, but it didn't help clarify the nature of the broadening.

Recent broad effects seen in elevation residuals during transit tracking have lead to speculation that an additional force is occasionally at work. One proposed force is drag caused by the elevation brake, whose contact with the elevation arc is not uniform at all elevations. The transit inclinometry experiments described here were designed to compare transits with and without additional drag.

Transit inclinometry data were first taken in the usual way, with the antenna transitting at elevation=10^o. (See here for a description of the original data format and the data processing that generates the plots below).

The transit step is very small (~1") and not the largest feature in the data. Features ('lumps') of this size are common, and short (~a few seconds long) excursions of twice that amplitude ('spikes') occur at several azimuths : 178.6, 179.0, 180.2, and 180.7. A short range of azimuth (179.5 to 179.9) is also affected by noise as commonly seen in previous datasets

Then the elevation brake was disengaged. The normal response to this is for the failsafe application of the brake, but Nash K. then manually overrode the clamp. A mark on the release nut shows that 4 full turns are necessary to 'fully' release the brake. A second transit inclinometry dataset was taken at the 3 full turns mark, where "some" drag was obvious:

Some of the 'lumps' and 'spikes' are diminished, but overall there is little difference.

The clamp was set to the 2 turn mark : ie more clamped than before, but it seemed obvious that the antenna was firmly held. At the 2-and-a-half mark some motion of the clamp disk was obvious, but we were unable to command the antenna to move to a new elevation - effectively the clamp was still on. Similarly at the 2 & 5/6th mark, but the antenna was moveable again at 2 & 11/12ths ! - which is probably the limit of the resolution of this activity. Another run was performed :

The change is amazing ! The transit step itself is clearly defined and remains v.small (~ 1" in amplitude), but the 'stuff' before and after is noise, twice as much as seen above, but comparable with that seen in older data. Although at first I thought that perhaps this noise had damped out the other 'lumps' and 'spikes', but the 'lumps' are actually still visible in the envelope of the noise, and it is impossible to tell if the 'spikes' are still there, too.

18-second averages of this data still show the lumps :

Although the pointing *may* be varying on a 1-second timescale on a +-2" scale the average position over a jiggle point, for instance, is <1" rms. It seemed that there may be benefits from having the elevation brake partially engaged (dragging) during observations, so this was tried on the sky . . .

On-sky transit tracking
Leaving the elevation brake in the condition described above, and setting SCURVE.AMP to zero (which is essentially justified from the transit inclinometry above), we tracked 3c273 from (az,el) = (153,70) at HST 21:36 to (205,71) at HST 22:48 :

The azimuth tracking performance is excellent : rms = 0.6" !!
The elevation residuals unfortunately do not suggest that the application of partial drag (if that's what's happening) has any beneficial effects upon tracking. Indeed, as we then set up to track another source (CenA, a v.low dec object), the antenna failed to move down beyond 63.5deg in elevation : the elevation arc apparently does not move uniformly between the plates of the elevation brake, and it seemed we hit a patch of increased drag sufficient to bind the antenna completely. We reset the manual override of the brake to its fully engaged position and re-connected it (which then disengaged it during normal operations). The antenna then moved as freely as usual, and we tracked CenA for 30mins through transit:

Azimuth performance is again excellent : rms = 0.7", but the elevation residuals again defy the conventional understanding and show only a monotonic increase across this azimuth range (178-184). It may be that a 'broad' transit curve is in play and we see only the central part of it - but it must then have a half-width of more than 4 degrees. Any abrupt narrow transit step is lost in the noise, and the small (1") suggested step size from the transit inclinometry is confirmed. At this level I'd prefer the s-correction be disabled.