[mmaimcal] Meeting tomorrow

[Al Wootten]

Folks,

I should have noted the wording of the proposed change:
As an indication, we propose a base level of 8dB image gain, with 7dB allowed 
in 10% of the frequency span of any IF band.

For discussion (see all the words below):
Should the base level remain 10 dB image gain, with 7dB allowed in 5% (10%?) of
the band?  Or should it be lowered overall as proposed, which affects system 
noise by letting more sky noise in through the other sideband?  Lazareff
provides an illustration whereby the fiddling required to meet the SB
spec increased system noise by 20K.

If I did my numbers right, if a receiver had 8dB rejection rather than 10 dB
I'd expect it to have about a 2% increase in Tsys (about 4K) and a little over
3% for 7dB.

Clear skies,
Al

Al Wootten writes:
 > Folks,
 > 
 > Let's have an imcal meeting tomorrow.  We need to finish discussion on
 > the B7 change requests.  These are 170, 171 and 173.  CRE 172 is on the
 > prolongation of ATF activity, also of interest. We should also discuss 
 > the progress at the ATF.
 > 
 > CRE 171: Sideband rejection.
 > At the Board meeting, Vigroux and van Dishoeck inquired about the B7
 > CREs.  Ewine worried about the effect on sensitivity.  Here is how I
 > responded to Ewine, for discussion.  Comments please?
 > 
 > As you may recall, no one from the science ipt was invited to the B7
 > CDR.  We did, however, have contact and some discussions with 
 > the review committee members.  It seems to be a characteristic of the
 > B7 cartridges that toward the upper end of the USB side there is
 > decreased performance.  A close look at the plots Lazareff supplied with
 > the CRE illustrates this well.  He illustrates a non-compliant assembly,
 > No. 9, which shows excellent noise temperature at many places in the band,
 > but suffers at the high end.  He notes that the noise spec is relaxed
 > for 370-373 GHz, that it applies to the integrated spec over 4-8 GHz,
 > and that there is a relaxed spec over 20% of the band.  There is no
 > 'relaxed spec' for any percentage of the band in image rejection.
 > Thus, a cartridge Lazareff illustrates shows rejection out of spec in only
 > 2.5% of points.  In one case, 'improvement' of the cartridge image rejection, so
 > that all points met the spec, degraded noise performance by about 20K.
 > Since that spec is easily met, and is met by the retuned cartridge, this
 > cartridge is in spec but in a way which might not be preferred by an astronomer.
 > 
 > We worried about the sideband rejection ratio spec enough to remove it
 > from the document approved by the Board, and put it at a lower level in
 > the hierarchy.  When Stephane wrote Memo 505, on band calibration, the
 > spec was (not yet agreed by all) 14 dB.  This was lowered to 10 dB in one
 > of our negotiations--I think Tony Kerr suggested the number.  It wasn't
 > really scientifically driven so much as negotiated.  Should it have
 > specified 10dB over 95% of the band?  In retrospect, I think it probably
 > should have.  
 > 
 > For B6 we had this discussion also, and I expect it to arise shortly for
 > B3.  For B6, the builders chose to build a cartridge with a much wider
 > IF than necessary--in the end, they just unilaterally chose to provide
 > ALMA with a narrower IF, promising to meet specs over 5.5 GHz rther than
 > 8 GHz (the receiver will have a few points with 9 dB rejection over the
 > whole 8 GHz).  There was some negotiation here--at first the offered
 > 4 GHz but we countered with 5.5, suggesting that being able to do all the
 > CO isotopes at once to full spec was an important science driver.  We
 > discussed using waivers with Webber, but he did not like the paperwork
 > involved.  He chose a solution which required neither waivers nor
 > change request in the end, but one which can really only apply to B6.
 > 
 > How should one approach errant receivers--with CRE or with waivers?
 > We had a discussion with Webber on this for the SB gain ration for B6.
 > We had suggested waivers but he disliked the paperwork.  Darrel and I prefer
 > waivers as then errant receivers are identified and flagged, for possible
 > upgrade to spec.  If we have a CRE, the spec is changed and nothing
 > gets improved until the next generation.  But if a large fraction of
 > receivers might be waivered, then a CRE becomes better, particularly from
 > the supplier's point of view.
 > 
 > I think that perhaps for the B7 sideband gain ratio CRE, a modified spec
 > over some percentage of the band might be allowable without deterioration
 > in the ALMA science goals.  I think that this could meet everyone's goals,
 > from scientist to engineer.
 > 
 > Clear skies,
 > Al