Links to sections/plots:
 

sections	plots
Intro
lbw	10mar02 focus curve
lbn	11mar02 focus curve
sbn	10may06: sbTx after new horn installed J2253+161 28ap06: sbTx before horn change J2253+161 25apr06 J0745+101 28feb03 3C138 (after dome move). 18feb03 CTA21 (5 strips) 18feb03 3C138 16dec02 focus curve. 09mar02 focus curve 20jan02  focus curve after final reflector adjustment. 07feb01  focus curve after first dish adjustment.
sbw	07mar02 focus curve.
sbh	02apr02 focus curve.
cb	10jan03  focus curve after horn moved. 23jan02  focus curve after final reflector adjustment. 09feb01  focus curve after first dish adjustment.
xb	06feb02  focus curve after shimming horn .44 inches. 24dec01 focus curve after receiver installed.

Comparing focus curves (2001 and 2002)
Correcting for kildal horn phase center error.
Comparing focus curves dec02-feb03 before/after dome move
Comparing focus curves 28feb03 to 2007

18feb03:focus curves after dec02 to feb03 before/after dome move
10jan03:focus curves for the various receivers 2001 and 2002(before dome moved)
13may02:focus curves correcting for kildal phase center.

 

---

Intro:

Adjusting the telescope optics for maximum gain includes adjusting the pitch, roll, and focus.  The focus includes:

1. The distance of the secondary from the main reflector. This can be adjusted with the tiedowns moving the entire platform up or down.
2. The distance of the tertiary from the secondary. This can be changed by moving the tertiary. As of feb01 this is not yet moveable.
3. The distance from the horn to the tertiary. This can be changed with the tertiary and by moving the feed.
Number 2 and 3 above have been surveyed with a theodolite and are left fixed.  It is assumed that these distances do not change much as we track an object (If you've ever sat in the feed tower as the azimuth accelerates you'll know that this is not quite true!!!). As the dome tracks a source its distance to the primary changes because of dynamic weight imbalances,  imperfections  in the azimuth/elevation rails, and from temperature affecting the main cables. These distances (weight and rails) are on the order of 2 inches (as of feb00).  Temperature change is about .2 inches platform height/deg F.

    When the platform is moved vertically a distance D the zenith angle pointing and the focal length changes..  After moving the platform up by D, the dome must be moved downhill along the za rails to continue pointing at the same position. The radial distance that the dome ends up moving is D*cos(za) (for small motions).  Since this motion does not change the relative distance between the secondary, tertiary, and horn, this focal length change does not include any multipliers (e.g.. when the tertiary is moved, it moves relative to the secondary and the horn so a multiplier enters into the distance calculation).  This motion does make a small change in the pitch of the dome since we end up sitting at a slightly different za along the elevation rails.

    The focus curves at high zenith angle are biased. The plotting assumes that the change in gain is from the change in focus. Above 15 degrees the spillover also reduces the gain. The bias direction depends on the direction the tiedowns are moving (up or down) and whether the source is rising or setting.

    The secondary, tertiary, and horn were surveyed into position from within the dome.  Reflector paddles on the mouth of the dome were then surveyed from within the dome and from the monument AO9 at the center of the reflector. As the dome moves, the motion of the dome relative to the paraxial surface is then measured. (just as a note: we need to remeasure the position of the AO9 monument relative to the average curvature of the dish during the dish photogrammetry measurements). There are a set of six distomats that also measure the average height of the platform (not the dome) every few minutes to an accuracy of about 1 mm.

Focus curve  technique by moving the platform:

This method cycles the tiedowns between 20" and 6" (the ratio of tiedown motion to platform motion is about 1.7 to 1 .. the cables and booms stretch and bend a bit). It take 14 minutes to go one direction. Since this motion changes the pointing, we also move the gregorian and turret to search for the optimum gain. This take two minutes and results in 7   gain measurements for each 14" tiedown motion. While this is happening, the distomats are measuring the average platform height every minute. We fit gaussians to these 7 points getting the peak, width, and offset. The peak gives the max gain, the width in minutes can be converted into the focal distance, and the offset (time) is used to find the platform position that gave the maximum gain for this 14 minute measurement.

---

History

• 25feb03: shimming of elevation rails completed.
• 27feb03: gregorian dome raised about 1.65 inches.

---

Measurements:

---

Lband narrow focus curves: (top...)

11mar02 lband narrow focus curve.
On 11mar02 a focus curve was done with lbn (1405 Mhz) during the day moving the tiedowns from 20" to 6" and back (11 strips were done). The source was   J2253+161 (3C454.3). The motion of the platform did not always go over the peak focus position (it was too low). The figures show:

1. Fig 1 plots the peak value versus sample, azimuth, and za. The vertical strips are the start of each 14 minute motion of the platform.
2. Fig 2,3 (polA,polB) plot the gain peak versus sample of strip for the 11 strips. The red lines are the gauss fits. You can see that strips: 2,3,4,11 did not move far enough to go over the gain curve. Some of the curves look like they started down the other side but then the last point ended up high (strips 8,9,10).
3. Fig 4 plots the platform motion and the position of the gain peak for each strip followed by the focus height versus azimuth and zenith angle. For strips 2 and 11 i moved the focus position to the edge of strip (it was outside the strip).
4. Fig 5 plots the distance needed (in lambdas) to reduce the gain by 3db. This is close to the value of 1 lambda that were measured by other receivers.

The fits to the data for lbn are hampered by our ability to move the platform low enough with the tiedowns to go over the entire gain curve. The data gives an upper limit of the focus. This receiver is focusing below all of the higher frequency receivers.

processing: x101/020311/foccur.pro

---

Lband wide focus curves: (top...)

10mar02 lband wide focus curve.
On 10mar02 a focus curve was done with lbw (1405 Mhz) during the day moving the tiedowns from 19" to 5" and back (11 strips were done). The source was   J2253+161 (3C454.3). The motion of the platform did not always go over the peak focus position (it was too low). The figures show:

1. Fig 1 plots the peak value versus sample, azimuth, and za. The vertical strips are the start of each 14 minute motion of the platform.
2. Fig 2,3 (polA,polB) plot the gain peak versus sample of strip for the 11 strips. The red lines are the gauss fits. You can see that strips: 2,3,4,5,8,9,11 did not move far enough to go over the gain curve. Some of the curves look like they started down the other side but then the last point ended up high (strips 2,4,5).
3. Fig 4 plots the platform motion and the position of the gain peak for each strip followed by the focus height versus azimuth and zenith angle. For strips 2,5,8, and 9 I  moved the focus position to the edge of strip (it was outside the strip).
4. Fig 5 plots the distance needed (in lambdas) to reduce the gain by 3db. Most receivers give a value of 1 lambda. lbw is probably low because the fits did not have enough of the gaussian to fit.

The fits to the data for lbw are hampered by our ability to move the platform low enough with the tiedowns to go over the entire gain curve. The data gives an upper limit of the focus. This receiver is focusing below all of the higher frequency receivers.

---

Sband narrow focus curves: (top...)

10may06 sbn Tx focus curve using new horn. Src J2253+161

    A room temperature receiver was mounted on one channel of the tx Horn and then piped back into the sbn post amp (polB). The tsys for this receiver was about 125 K. A focus curve was done on J2253+161 starting at 6:50 till 9:30.  A total of 10 passes were done with the tiedowns (moving up/down). This should be compared with the focus curve done on 28apr06 with the old tx horn.
    The plots show the measured focus of the telescope (.ps) (.pdf) :

page 3 : The  blue lines are the measured focus position for this source. The black * in the top plot show the platform position. The 11th pass showw the tiedowns losing tensions (so i only used the first 10 passes). The black lines in the middle and bottom plots are the model focus position for the path that this source tracks on the dish. This model was made before the dome was moved up 1.65 " on 27feb03. There is a dc offset because of this , but it still shows how much the focus error changes for this source as we track across this part of the dish. The bottom plot has a dashed red line at 1256.22 .This is the position we are currently using for the platform height (it is also the height that was used when the last model was made).

28apr06 sbn Tx focus curve using old horn. Src J2253+161

    A room temperature receiver was mounted on one channel of the tx Horn and then piped back into the sbn post amp (polB). The tsys for this receiver was about 125 K. A focus curve was done on J2253+161 starting at 8:00 till 10:15.  A total of 8 passes were done with the tiedowns (moving up/down). We missed the first 30 minutes of the source rising. During the 9th pass around 10:15 we started to loose tension in the tiedowns so it was not used. This  focus curve used the old tx horn. It should be compared with the focus curve that will be done next week when the new horn is installed. The plots show the measured focus of the telescope (.ps) (.pdf) :

page 3 : The  blue lines are the measured focus position for this source. The black * in the top plot show the platform position. The last pass show the tiedowns losing tensions (so the platform stopped moving up). The black lines in the middle and bottom plots are the model focus position for the path that this source tracks on the dish. This model was made before the dome was moved up 1.65 " on 27feb03. There is a dc offset because of this , but it still shows how much the focus error changes for this source as we track across this part of the dish. The bottom plot has a dashed red line at 1256.22 .This is the position we are currently using for the platform height (it is also the height that was used when the last model was made).

25apr06 sbn focus curve using J0745+101

    A focus curve was done with on J0745+101 using sbn. This is the first focus curve in 3 years. It was done before  the new rcv horn was installed. The data taking started about 17:30. There were a total of 8 passes. The plots show the measured focus of the telescope (.ps) (.pdf) :

• page 4 : This shows the focus height. The green and blue lines are the measured focus position for this source. The black lines in the middle and bottom plots are the focus position for the path that this source tracks on the dish. This model was made before the dome was moved up 1.65 " on 27feb03. There is a dc offset because of this , but it still shows how much the focus error changes for this source as we track across this part of the dish. The bottom plot has a dashed red line at 1256.22 .This is the position we are currently using for the platform height (it is also the height that was used when the last model was made). The source is focusing a little below 1256.22. This needs to be compared to how this sources focus errors compare to an average track across the dish.

28feb03 sband narrow focus curve (after dome move) (.ps)  (.pdf)

     Focus curves using sband narrow were done on 3C138 (J0521+166) in the evening after the dome had been repositioned.  We tried to move the dome up 1.5 inches. This should have moved the platform focus position down 1.5 inches.The model focus postion is lynn bakers using the 17feb03 survey with the dome in the 'old' position (before the move). There are also plots of the comparison of the focus curves before and after the dome move.

processing: x101/030228/doitfoc_0521.pro

 

18feb03 sband narrow focus curve.

     Focus curves using sband narrow were done on CTA21 (5 strips) and 3C138 (11 strips) starting in th evening. This data was taken after the shimming and before the dome move. The model focus postion is lynn bakers using the reflector as the center and the 19feb03 2dmodel.

processing: x101/030218/doitfoc_0318.pro,doitfoc_0521.pro

 

16dec02 sband narrow focus curve.

    A focus curve using sband narrow was done on 3C454.3 on the risin half only. The highest point (at transit) should be ignored since the data didn't go far enough over the peak to get a good fit.
The values came out close to what was measured in jan02. The plot of most recent focus curves shows the comparison.
processing: x101/021216/foccur.pro
 

09mar02 sband narrow focus curve.

    The focus curves for cband, sband wide, and lband wide were coming out much lower than the xband,sbn focus curves. I redid the sband narrow focus curve to make sure that the tertiary hadn't moved since jan02. The first strip was bad because levels had to be readjusted.
The values came out close to what was measured in jan02. The plot of most recent focus curves shows the comparison.

processing: x101/020309/foccur.pro

20jan02 focus curve after final reflector adjustment.

    These figures were taken after completion of the reflector adjustments. They also use the model focus curve from aug01 theodolite survey. I've included polA and polB in the plots.
    processing: x101/020120/foccur.pro
 

07feb01 measuring sband  narrow focus curves after first dish adjustment.

On 7feb01 focus curves were done at sband (2380 Mhz) during the day moving the tiedowns from 20" to 6" and back (9 strips were done). The source was   J2253+161 (3C454.3) which has  a flatten spectrum. The flux at 2380Mhz is (about) 13.8 Jy.

1. Figure 1. This is the gain (% of Tsys)  as a function of sample, azimuth, and zenith angle. The vertical dashed lines are the start of each motion up or down of the tiedowns. The peak of 4.4 Tsys gives an SEFD of 3.1 Jy/Tsys.
2. Figure 2 shows the gain as a function of time for each strip. The red curves are the gaussian fits. The vertical red line is the fit center. This value is used to compute the position of the platform for the peak gain.
3. Figure 3 top plots the platform height in feet (from the distomats) versus minute of the day. The green line connects the platform heights where the peak gain was measured on each strip. The center and bottom plots show the platform height for max gain versus azimuth and zenith angle. The red line is the computed focal height coming from the theodolite measurements using a zero focus error height of 1256.35 feet. Above 15 degrees za the measured and computed values diverge. This may be caused by the large pitch error at high za or that the theodolite measurements at high za were known to have some problems (not enough signal to noise to measure all the  targets at the dome opening).
4. Figure 4 is the focal width. It is the distance (in lambda's) for the gain to fall 3db. This comes from the gaussian fit widths.

processing: x101/010207/doit.pro

---

Sband wide focus curves: (top...)

07mar02 sband wide focus curve.
On 07mar02 a focus curve was done on the source J2253+161 using sband wide at 2680 Mhz. The tiedowns were moved from 20 to 6 inches in 14 minutes. A total of 11 of these strips were done tracking the source from rise to set. This was done ast 11:00 to ast 13:45.

1. Fig 1 plots the source strength by sample, azimuth, and za.
2. Fig 2,3 show the data by strip and the gauss fits for polA and polB.
3. Fig 4 top plots the platform position versus time with the position of the source peak overplotted in color.
4. Fig 4 center, bottom plot the focus height versus azimuth and za. The green, blue curves are polA,B while the black curve is the expected focus position from the model using a radius of 869.781 feet.
5. Fig 5 plots the fraction of a wavelength motion needed to change the gain by 3db. It is consitant with the other receiver with 3db motion equals 1 lambda.

The focus for sbw is about 1inch below cband and 3 inches below sbn. I need to redo the sband narrow focus curve to make sure that something has not changed in the optical path (tertiary position).

processing: x101/020307/foccur.pro

---

Sband high focus curves: (top...)

02apr02 sband high focus curve.
On 02apr02 a focus curve was done on the source J2253+161 using sband high at 3500 Mhz. The tiedowns were moved from 19 to 5  inches in 14 minutes. A total of 11 of these strips were done tracking the source from rise to set. This was done ast 9:00 to 11:45 ast. On 29mar02 the sbh horn was surveyed into position. The horn was found to be .703 inches low in dome centerline z coordinate. This focus curve was done without correcting for this error.
1. Fig1 plots the source strength by sample, azimuth, and za.
2. Fig 2,3 show the data by strip and the gauss fits for polA and polB.
3. Fig 4 top plots the platform position versus time with the position of the source peak overplotted in color.
4. Fig 4 center, bottom plot the focus height versus azimuth and za. The green, blue curves are polA,B while the black curve is the expected focus position from the model using a radius of 869.781 feet.
5. Fig 5 plots the fraction of a wavelength motion needed to change the gain by 3db. It is consitant with the other receiver with 3db motion equals 1 lambda.
The focus for sbh falls close to cband. The Ying/kildal horns look to have a phase center error that is proportional to lambda. If we had lowered the sbh horn by .7 inches (as the survey said to do) the focus height would have fallen between cband and sbw.
processing: x101/020402/foccur.pro

---

cband focus curves: (top...)

9feb01 measuring cband  focus curves after first dish adjustment.

On 9feb01 focus curves were done with the same source J2253+161 at 5000 Mhz using the cband receiver. The flux used at 5000 Mhz was 15.3 Jy.
1. Figure 1. This is the gain (% of Tsys) as a function of sample, za, and azimuth. At transit and high za the gain dropped. This could be do to the pitch and roll.
2. Figure 2 are the fits to each strip. You probably should take the last fit with a grain of salt...
3. Figure 3 plots the platform position versus time. The green line is the platform position for each peak gain. The missing platform positions are where the distomats dropped out. The bottom two plots show the platform height at the peaks versus azimuth and zenith. The red line is the computed focal position using the model. There is about a one inch offset between the measured and computed values (and relative to the sband measurement). Since it is the same source, the difference must be from the position of the horns. The divergence of the model and measured focus error at high zenith matches that of sband.
4. Figure 4 shows the focal width in lambda's for the gain to drop 3 db. The value at high za should probably be ignored since the fit wasn't so great.
processing: x101/010209/doit.pro
23jan02 cband focus curve after final reflector adjustment.
    These figures were taken after the completion of the reflector adjustments. They include the model focus curve from the aug01 theodolite survey. Pol A is green and pol B is blue.
    processing: x101/020123/foccur.pro

10jan03 cband focus curve (after horn moved).
    These figures were taken after the cband horn was moved down 1.08 inches to put it at the correct phase center (28aug02).  The two points at low za should be ignored since we did not move the platform high enough to go over the other side of the gain curve.
At 10 degrees za the platform height focus went from 1256.26 to 1256.38 or 1.4 inches. It should have moved 1.08/cos(10) or 1.1 inches. About 2500 lbs have been added to the dome  (trolley replacement) since the 23jan03 measurement. This has caused the dome to tilt down which may have caused the extra .3 inches platform motion.
    processing: x101/030110/doitfoc.pro

 

---

Xband focus curves: (top...)

24dec01 measuring xband focus curves after receiver installed.

    On 24dec01 focus curves were done on the source J2253+161 at 8.8 Ghz using the xband receiver. The receiver had not been surveyed into position with the theodolite. The  heating resistor to prevent condensation was not installed until later so the system temperature was probably 100K.

    The focus curve routine was modified to move the tiedowns from 16 to 6 tiedown inches in 10 minutes instead of the normal 20 to 6 inches. Moving at 1 tdInch/minute gave 5 points for each direction to fit gaussians to. The dome moved +/- 2.5 beamwidths in 2 minutes so it traversed the beam in about .4 mintues. During this .4 minutes the platform moved .4*1./1.74=.23 focus inches (ignoring sinza). This is .15 lambda in focus so there was not too much smearing. It probably wouldn't hurt to slow the tiedowns back to 14 minutes for 10 inch motion to give 7 points per strip. This would improve the gauss fits for each strip plus cut down on the smearing in focus.

1. Figure 1. This is the gain (% of Tsys) as a function of sample, za, and azimuth. At transit and high za the gain dropped. This could be do to the pitch and roll.
2. Figure 2 are the fits to each strip. You probably should take the last fits with a grain of salt...
3. Figure 3 plots the platform position versus time. The green line is the platform position for each peak gain. The bottom two plots show the platform height at the peaks versus azimuth and zenith. The red line is the computed focal position using the model from aug01. There is 1.5 to 2 inch offset between the measured and computed values (and relative to the sband measurement).
4. Figure 4 shows the focal width in lambda's for the gain to drop 3 db. It is close to 1 lambda between 5 and 10 degrees za. Above and below this it rises. This is the same region where the pitch and roll goes bad. This may be degradation of the fits, or the relative gain with focus is changing slower do to the collimation error.
processing: x101/011224/foccur.pro

06feb02 xband focus curve after shimming horn .44 inches.
    A focus curve was done on J2253+161 after the horn was moved radially upward .44 inches via shimming. The heating resistor was installed and Tsys polA was about 45K. model 13A was used for the pointing.  5 strips were done (after transit) with 7 points per strip (instead of the 5 pnts per strip on 24dec01). On the second strip the distomat update rate was switched from 2 to 1 minutes (I forgot to switch it when I started). The bottom plot of figure 4 shows that the platform height for the focus moved -. 7 inches.  This is probably correct within the accuracy of the measurements. The fits for the  24dec01 data  had larger errors because of the high Tsys. Comparing figure 1 top plot from both dates shows the gain as a % of Tsys going from  30% to 80% . This was do to the decrease in Tsys.
processing:x101/020206/foccur.pro

---

Comparing focus curves 2001 and 2002: (top...)

1. Fig 1 Top.   The solid lines are near jan02 while the dashed lines are around jan01. The sband and cband heights have moved up (it wants a longer focal path) by about .1 feet (about 1") between jan01 and jan02. This could be from the adjustments on the dish or more likely when the tertiary motors were installed in aug01 and the tertiary was resurveyed into position. The green xband with the +'s was taken before the horn was moved .44 inches up. The horn motion made the path length longer so the platform now focuses at a lower position. The platform height change is in pretty good agreement with the horn motion.
2. Fig 1 bottom. This plots the most recent focus curve positions for all the receivers with the dates of the measurement.  The sband narrow curve was repeated on 10mar02,16dec02  to make sure that the tertiary hadn't moved. The results came out the same as the 20jan02 data. The lband value jumps around a lot. I could not get the platform low enough to go over the other side of the gain peak for all of the lband positions. The cband data was remeausured 10jan03 after the horn focus had been moved.

    When the ao9 survey  was performed on aug01, the platform was fixed at 1256.35 feet. If the analyzed data then showed a focus error of say + 2 inches (the path length is too long), then you would expect to see the focus position of the platform to be 2 inches (divided by the cos(za)) below 1256.35.

    The sband narrow and xband receivers use a horn built by lynn baker. The lbn,lbw,sbw,cb use a ying/kildal horn. The ying/kildal horn focus position changes as a function of frequency. The table below shows the platform height in feet above sea level for the focus of each feed at za equals 9 degrees. It also plots how much the ying/kildal horns differ from the sbn,xb horns in inches and lambdas. The model position is also included in the table. For the reference position of sbn,xb i used 1256.355 (the average of the two).
 

rcvr	freq Mhz	focHght(ft) za=9	inches below sbn,xb focus	lambda below sbn,xb focus
lbn	1405	1256.0	-4.3	.51
lbw	1405	1256.0	-4.3	.51
sbw	2680	1256.11	-3.0	.67
sbh	3500	1256.28	-.9 (-.7=-1.6)	.27 (.47)  **
cband	5000	1256.27	-1.1	.44
sbn	2380	1256.35	-.06	0
xband	8800	1256.36	+.06	0
model		1256.15	-2.5

** The sbh focus curve results were taken with the horn .7 inches low. This caused the platform to focus higher than if the horn had been positioned the same as the other kildal horns. The values in () include the .7 inch offset.
If the sbn,xb horns are positioned correctly , then the ying/kildal horns are positioned incorrectly by about lambda/2.

---

Correcting for kildal horn phase center errors. (top...)

---

Comparing focus curves after dec02 to feb03: (top...)

---

Comparing focus curves feb03 to 2007: (top...)