[mmaimcal]forwarded message from Stefano Stanghellini
Jeff Mangum
jmangum at nrao.edu
Tue Sep 30 01:32:49 EDT 2003
Hi Bryan,
"Bryan" == Bryan Butler <bbutler at aoc.nrao.edu> writes:
Bryan> the repeatibility should be << the spec on antenna location determination,
Bryan> if you don't allow for OTT and hence can't measure it well, IMHO.
Bryan> this means it should be of order, say, 10 microns, and over much
Bryan> longer periods than weeks (again, because you can't measure it
Bryan> well). if you *can* measure it via going OTT, then it can be relaxed
Bryan> significantly - to what you proposed before (i.e., of order half
Bryan> the antenna location error [30-ish microns] over weeks-long periods).
So, it seems to me that the discussion now hinges upon how much better
125 degrees is over 100(ish) degrees for K-correction determination.
Cheers,
Jeff
Bryan> On 2003.09.29 07:34, Al Wootten wrote:
>> Hi Jeff
>>
>> The OVRO antennas do not go over the top. I understand that the
>> inability therefore to separate the axis non-intersection from the 'up'
>> term when antennas are moved is one of the primary problems with data
>> from OVRO--there is always a phase drift across the bandpass. If the
>> antenna has an axis non-intersection which is very small, such that it
>> does not affect the baselines significantly, then the phase drift will
>> be minimized. But we move antennas every four days and approving a
>> specification which we know will complicate baseline determination, and
>> which will introduce phase errors, should only be done if the cost of
>> implementing over the top is large. Repeatability matters but only if
>> repeatable within the baseline errors. What should the stability be? I
>> proposed a change less than the baseline error over the two weeks
>> between antenna moves in the baseline plan. I haven't heard much
>> discussion of this.
>>
>> Clear skies,
>> Al
>>
>> Jeff Mangum wrote:
>> > Hi Al,
>> >
>> > Stefano makes a very good point. Isn't it repeatability that matters?
>> > If he can deliver an antenna with the axis non-intersection spec he
>> > describes, do we really need such a large OTT?
>> >
>> > Note that I of the belief that OTT is not necessary, or even desired,
>> > for pointing.
>> >
>> > Cheers,
>> >
>> > Jeff
>> >
>> >
>> > "Al" == Al Wootten <awootten at nrao.edu> writes:
>> >
>> > Al> Hi Bryan and Jeff, and everyone
>> > Al> Stefano has a different take on this--specify the axis non intersection
>> > Al> to be very small and don't go over the top. His perspective is different
>> > Al> from those discussed so far so I circulate it here. I'll have to go to the
>> > Al> ASAC telecon for the rest of the morning (I did dig up my leader passcode
>> > Al> Bryan).
>> >
>> > Al> Clear skies,
>> > Al> Al
>> > Al> MIME-Version: 1.0
>> > Al> Content-Length: 6689
>> > Al> Content-Type: text/plain; charset=us-ascii
>> > Al> Return-Path: <sstanghe at eso.org>
>> > Al> Received: from cv3.cv.nrao.edu (cv3.cv.nrao.edu [192.33.115.2])
>> > Al> by polaris.cv.nrao.edu (8.11.6/8.11.6) with ESMTP id h8OCXIg24686;
>> > Al> Wed, 24 Sep 2003 08:33:18 -0400
>> > Al> Received: from eso-wall-ext.hq.eso.org (firewall-user at eso-wall-ext.hq.eso.org [134.171.69.199])
>> > Al> by cv3.cv.nrao.edu (8.11.6/8.11.6) with ESMTP id h8OCXD512077;
>> > Al> Wed, 24 Sep 2003 08:33:13 -0400
>> > Al> Received: (from uucp at localhost)
>> > Al> by eso-wall-ext.hq.eso.org (8.10.2+Sun/8.10.2) id h8OCXBh06502;
>> > Al> Wed, 24 Sep 2003 14:33:11 +0200 (MEST)
>> > Al> Received: from mercury.hq.eso.org(134.171.7.20) by eso-wall-ext.hq.eso.org via csmap (V6.0)
>> > Al> id srcAAAY7aOSm; Wed, 24 Sep 03 14:33:09 +0200
>> > Al> Received: from serapis.hq.eso.org (serapis.hq.eso.org [134.171.7.10])
>> > Al> by mercury.hq.eso.org (8.12.6/8.12.6) with ESMTP id h8OCWrKm003885;
>> > Al> Wed, 24 Sep 2003 14:32:53 +0200 (MEST)
>> > Al> Received: from eso.org (pc003513.hq.eso.org [134.171.24.104])
>> > Al> by serapis.hq.eso.org (8.11.7+Sun/8.11.7) with ESMTP id h8OCWrw03131;
>> > Al> Wed, 24 Sep 2003 14:32:53 +0200 (MEST)
>> > Al> Message-ID: <3F718EF1.A2ABB716 at eso.org>
>> > Al> X-Mailer: Mozilla 4.79 [en] (WinNT; U)
>> > Al> X-Accept-Language: en
>> > Al> References: <16240.49949.949604.622363 at polaris.cv.nrao.edu>
>> > Al> X-Virus-Scanned: by amavisd-milter (http://amavis.org/)
>> > Al> X-MailScanner-Information: Please contact postmaster at cv.nrao.edu for more information
>> > Al> X-MailScanner: Found to be clean
>> > Al> X-MailScanner-SpamCheck: not spam, SpamAssassin (score=-6.3, required 7,
>> > Al> BAYES_01 -5.40, EMAIL_ATTRIBUTION -0.50, QUOTED_EMAIL_TEXT -0.38,
>> > Al> REFERENCES -0.00, REPLY_WITH_QUOTES 0.00,
>> > Al> USER_AGENT_MOZILLA_XM 0.00)
>> > Al> From: Stefano Stanghellini <sstanghe at eso.org>
>> > Al> To: Al Wootten <awootten at nrao.edu>, jbaars at nrao.edu, jbaars at eso.org
>> > Al> CC: demerson at polaris.cv.nrao.edu,
>> > Al> Ewine van Dishoeck <ewine at strw.leidenuniv.nl>, dsramek at nrao.edu,
>> > Al> rkurz at eso.org, mrafal at eso.org, Kraus Maximilian <mkraus at eso.org>,
>> > Al> Massimo Tarenghi <mtarengh at eso.org>
>> > Al> Subject: Re: Antenna RFP
>> > Al> Date: Wed, 24 Sep 2003 14:32:49 +0200
>> >
>> > Al> Dear Al,
>> >
>> > Al> Thank you for your prompt answer.
>> > Al> Still, I do not agree yet to the various reasoning, neither I found in the
>> > Al> Calibration paper sufficient information to convince me.
>> >
>> > Al> The point is that an engineering specification shall specify what we really
>> > Al> want and not what we want as a safety basket.
>> > Al> This is the more true as every specification comes with a price tag.
>> >
>> > Al> My understanding of the issue is that we want long term stability of the offset
>> > Al> because this is equivalent to the change in distance between the antennas.
>> > Al> Today, apart from the three millimeters maximum offset, there is in the
>> > Al> Calibration paper no value of stability, neither in the prototype specs.
>> > Al> In addition the paper refers to slow variations (16 months interval measurement
>> > Al> at the VLA). You refers to measuring at every repositioning of the antenna on a station.
>> >
>> > Al> This is contrary to my understanding of the issue: A good manufactured
>> > Al> antenna will have fast variations due to run-out and these variations are
>> > Al> changing permanently. Is this what you want to measure? Or it is only the
>> > Al> non-repeatable slow variations which you want to measure?
>> >
>> > Al> Now some side information: The AEC antenna has an axis offset in the
>> > Al> range of 0.16 mm. The one of vertex (I need to check) should be
>> > Al> around 0.5 mm, but measured with a lower level of accuracy)
>> > Al> As I have understood the issue, this axis offset does not necessary
>> > Al> be so small, if it is well known.
>> > Al> This offset can be measured very precisely during acceptance testing
>> > Al> and I intend to specify this measurement. (Typical precision lower than
>> > Al> 100 micrometers).
>> >
>> > Al> The variation of this offset has to be kept small by design. The reason
>> > Al> for this offset variation are mainly:
>> >
>> > Al> - thermal gradient (but the antenna will be insulated) Here you have
>> > Al> no control on it and it will change at least twice daily, so you are not
>> > Al> solving the problem measuring every week or so;
>> >
>> > Al> - Elastic deformation. This will be very small, and it needs to be specified
>> > Al> in order to maintain it small. Imagine you have an elevation structure which
>> > Al> is not balanced: you will have a moment and this moment will bend the yoke
>> > Al> arm. You will have a variation of the offset of say, x micrometer, depending
>> > Al> on the elevation angle. This is however small and repeatable. Furthermore it is
>> > Al> independent from the interval within your measurements. It can be computed
>> > Al> rather exactly, and possibly measured.
>> >
>> > Al> - Wind induced elastic deformation. This cannot be seriously measured
>> > Al> by going over the top. It can be only computed.
>> >
>> > Al> - bearing run-out. I expect bearing run out to be in the order of 30 microns
>> > Al> for azimuth axis and << 10 microns on the elevation. This run out however
>> > Al> has a repeatable fast moving component (twice /over 360 degrees azimuth,
>> > Al> 0.5 times over 90 degrees elevation) and a slow varying non repeatable effect.
>> > Al> (I would judge this to be less than 15% of the other) This is probably the
>> > Al> only one of relevance, and it looks like being very, very small compared to
>> > Al> the overall accuracy requirement of 65 micrometer.
>> > Al> Note that I will specify the max acceptable bearing runout and get it measured
>> > Al> in the frame of the acceptance testing, so we will have all the antenna around
>> > Al> the value I mention above.
>> >
>> > Al> In conclusion:
>> >
>> > Al> - There is a lot that can be done at engineering level, both in terms
>> > Al> of specification and acceptance testing;
>> >
>> > Al> - I still have not fully understood the "real" requirements and I do not agree
>> > Al> to put in the spec "safety basket" unless I am convinced there is not better
>> > Al> method of clearly specifying what ALMA needs. (On this I would kindly ask our
>> > Al> colleague Jaap, who is my safety basket, to give me some help on the matter);
>> >
>> > Al> - specification comes with a price tag and I feel authorized to challenge
>> > Al> them in this phase;
>> >
>> > Al> - last but not least (?) Going over the top is a considerable source of hysteresis
>> > Al> in the system, (releasing stresses, closing gaps, changing the direction of friction)
>> > Al> and I am against it in such a precise machine unless it is absolute necessary.
>> > Al> (ESO avoids to do this in telescopes like the VLT.)
>> >
>> > Al> Issue to be followed. regards,
>> >
>> >
>> > Al> Stefano
>> >
>> >
>> >
>> >
>> > Al> Al Wootten wrote:
>> >
>> >
>> >>>Current prototype spec on non-intersection of axes is ~3mm; current draft
>> >>>RFP spec is ~2mm and some would argue to make it <1mm.
>> >>>
>> >>>Current spec for 'over the top' motion is 125 degrees and we were asked
>> >>>if this could be made smaller.
>> >>>
>> >>>Currently, non-intersection of axes is <1mm at OVRO and BIMA.
>> >>>The preponderance of opinion
>> >>>within the science IPT is that as long as it is <3mm the absolute value
>> >>>is not critical but it must be stable.
>> >>>
>> >>>What we need to know is how accurately we can measure it
>> >>>and how stable the axis non-intersection remains over the measurement
>> >>>interval.
>> >>>
>> >>>We need to measure the axis non-intersection to the same precision as the
>> >>>baseline accuracy, which is 65 microns in the current calibration document
>> >>>Calibration of ALMA (see http://almaedm.tuc.nrao.edu/forums/alma/dispatch.cgi/docapproval/docProfile/100684/d20030917184242/No/t100684.htm
>> >>>
>> >>>Each antenna will move about every two weeks, after which the measurement
>> >>>of axis non-intersection may be made with the antenna on a new pad. It
>> >>>should be stable to 65 microns over at least that time period.
>> >>>
>> >>>It is possible to measure the axis non-intersection with antennas which
>> >>>do not go 'over the top' but it can be tedious. With ALMA the tedium is
>> >>>relieved by the abundance of calibrators revealed by its sensitivity,
>> >>>the excellent weather on the site, and the very high frequencies to which
>> >>>it will operate. However, there are also many antennas!
>> >>>
>> >>>Many observatories with antennas which do go 'over the top' have found the
>> >>>ability to do that useful. OVRO, BIMA and the VLA do; IRAM does not.
>> >>>Most believe that the ALMA antenna should provide the capability to go
>> >>>over the top to secure the measurement of axis non-intersection.
>> >>>100 degrees is not enough owing to the trigonometric dependencies of the
>> >>>equations involved; 125 degrees gives a much better measurement.
>> >>>
>> >>>Some cited the scarcity of clear zenith weather as supporting over the top
>> >>>motion, to relieve long antenna slews while tracking a source. However
>> >>>the preponderance of opinion is that pointing is seldom accurate 'over the top'
>> >>>and this was not a driving reason for the spec.
>> >>>
>> >>>In summary, the Science IPT thinks that
>> >>>
>> >>>*the absolute accuracy of the non-intersection is not as important as its
>> >>>stability; it must be measured to the same accuracy as baselines, or
>> >>>65 microns and should remain stable over periods of many weeks to months.
>> >>>
>> >>>*it is much easier to measure the axis non-intersection if the antennas
>> >>>go over the top by about 125 degrees.
>> >>>
>> >>>Clear skies,
>> >>>Al
>> >>
>> >
>>
>>
>> _______________________________________________
>> mmaimcal mailing list
>> mmaimcal at listmgr.cv.nrao.edu
>> http://listmgr.cv.nrao.edu/mailman/listinfo/mmaimcal
>>
More information about the mmaimcal
mailing list