Finding the best horn/turret position by tracking sources.

apr2002

Links to plots:

lbw gain and turret position
lbn gain and turret position
sbh gain and turret position
cband gain and turret position
xband gain and turret position

Sections:

lbw
lbn
sbh
cband
xband
summary

Finding the best turret position.

The horns are surveyed into position using a theodolite and the secondary as the reference frame. This should give the optimum signal for perfect optics. Collimation errors in the azimuth direction (roll errors) can cause the maximum gain to be at a different turret position.
The turret position was checked by tracking a source rise to set and doing turret scans at 5 different turret positions. At each of the 5 turret positions two turret scans were done (the dome going uphill and then downhill). It took about 25 minutes to step through all 5 turret positions. This was repeated rise to set to give 5 or 6 measurements at each turret position. The source deflection (as a percentage of Tsys) for polA and polB was averaged together and then plotted versus zenith angle with different colors for each turret position. The measurements were also interpolated in azimuth to the average position for each set of 5 turret position to compare the curves. If the central turret position was at the optimum position then there should be 3 turret positions of about the same gain (since the +/- 1*turstep should not change the gain very much) and then two others that should drop off equally. The bottom plot shows the roll collimation error versus za for this source. The plots below show the results of the measurements.

lbw: 11apr02 J0020+156 with turret position = 286.52 (top)

The lbw beam width at 1400 Mhz is about 3.3 arc minutes. A 2 deg turret step (2*.75= 1.5 arc minute on the sky) was used (about .5 beam widths). The lbw gain and turret position is plotted versus za and azimuth. The central position of 286.52 centers the offsets pretty well. The surveyed optimum position for this horn is 284.48 so the current position of 286.52 is 2.4 turret degrees from the surveyed position (after the survey the offset was added in backwards). It just so happens that this incorrect position gives the maximum gain. The current position is the green line while the surveyed position is the bright red line.
These plots also give you an idea of what will happen with the multibeam array. The 4 degree turret step is (4*.75)/3.3=.9 beams on the sky. The gain decreased by about 8%. I think the beams on the multibeam array will be offset from the center by 2*beamwidths.

processing:x101/020411/doitlbw.pro

lbn: 22apr02 J0665+146 using turpos 310.14. (top)

The lbn beam width at 1400 Mhz is about 3.3 arcminutes. A 2 deg turret step (2*.75= 1.5 arc minute on the sky) was used (about .5 beam widths). The lbn gain and turret position is plotted versus za and azimuth. The -2 degree offset step is lower in gain than the +2 degree offset. It looks like the turpos position of 310.14 is about 1 degree low from the optimum power positino.
processing:x101/020422/doitlbn.pro

sbh: 01apr02 J2253+161 (3C454.3) with turret position =125.827 (top)

The sbh beam with at 3500 Mhz is 80 asecs. A .75 degree turret step is 33.75 asecs or .42 beams on the sky. The sbh gain and turret position is plotted versus za and azimuth. The turret steps were centered about .2 degrees below the surveyed position (the data was taken before the survey). The green, blue, and dark red lines are all at about the same gain so the measured best position is at least .57 degrees above the surveyed position (it probably wouldn't hurt to redo this since we didn't go over the other side of the gain vs turret position curve).

processing:x101/020401/doitsbh.pro

cband:11apr02 J0318+164 (CTA21) with turret position = 206.65 (top)

The cband beam width at 5 Ghz is about 60 asecs. A .75 turret degree step was used. This is 45 Asecs or .75 beams. The cband gain and turret position is plotted versus za and azimuth. Turret steps were centered on the surveyed optimum position of 206.65. The best measured position is probably between the blue and dark red positions about 1 turret degree above the surveyed position.

processing:x101/020411/doitcb.pro

xband:01apr02 J0238|+166 with surveyed turret position = 161.50 (top)

The xband beam at 8800 Mhz is 33 asecs. A .75 turret degree step was used. This is 45 asecs or 1.4 beams. The xband gain and turret position is plotted versus za. The central turret position used was 162.34 deg or .84 degrees above the surveyed position (the data was taken before the survey was done). There are not very many points and they are at low za where the gain changes rapidly do to the collimation errors there. It does look like the best turret position is larger than the surveyed position by a few degrees. This source receiver should also be done over with a stronger source.

processing:x101/020401/doitxb.pro

Summary (top)

The results for the various feeds is summarized in the table below.

rcvr/freq	step Used turDeg/beams	turret offset from survey for max gain (turret degrees)	turret offset from Curpos for max gain	comments
lbw/1400	2.0 Turdeg/.5beams	+2.4 deg from (284.48)	0 deg from 286.52
lbn/1400	2.0 Turdeg/.5beams	+1.0 deg from 310.14	same
sbh/3500	.75 Turdeg/.42beams	>+.57 deg from (125.827)	same	redo so we can go over the other size of the gain curve
cb/5000	.75 turDeg/.75beams	+1 deg from (206.65)	same
xb/8800	.75 turDeg/1.4beams	+ 1deg? from 161.50	same	needs to be redone, not enough points.

The technique probably gives accurate results if you start centered on the optimum position. Then you can see the falloff on both sides with turret position. This occurred with lbw. With the other receivers all you can say is that it looks like you need to move in a certain direction to make it better. In all cases the maximum gain is occurring at a turret position that is larger (closer to the stairwell) than the surveyed one. This may be caused by the negative roll error that we have for most za's.

 home_~phil