[evlatests] Antenna 5 Bandpass (cont).
Dan Mertely
dmertely at nrao.edu
Fri Nov 4 13:02:27 EDT 2011
Hi Jim. Would you try to organize a definitive
test of the L-band bandpass of ea05, searching for
the ripple written up in C127766? As you can see
in the message at the top of this email chain, the
W8 monitor tap was removed back on Friday, 07 Oct,
2011, so a test at this time should determine
whether the ripple problem (only seen in
correlated observations--see Urvashi's 27 Oct,
2010 email at the bottom of this email chain) is
in fact due to the W8 monitor tap.
The W8 monitor is quite a useful tool, used by not
only the VLA IP group, but also non-NRAO astronomers
and observatories. I would like to get it back on-
line. Since all the ea05 ripple reports varied as
to polarization and occurrence, and since the W8mon
tap hardware has changed a number of times (without
a correlated change in ea05 ripple reports), I think
we need to determine whether it was the cause, now
that it has been totally disconnected for nearly a month.
Thanks.
-Mert
> -------- Original Message --------
> Subject: Changes to ea05 L-band RCP delays?
> Date: Fri, 07 Oct 2011 12:22:46 -0600
> From: Dan Mertely <dmertely at aoc.nrao.edu>
> I removed the W8 Monitor tap from the RCP output
> of the L-band FE on ea05 this morning. The tap
> had a 3 inch long 90 deg hybrid, plus a number of
> coax adapters on it. I would recommend checking
> L-band delays as a result of the shortening of the
> RCP signal path.
>
> I removed it in order to debug the wide-band noise
> hump seen in some observations since Oct 2010. It
> would be useful to re-check the correlated bandpass
> in the same way the bump was discovered, in order
> to determine whether it has now gone away.
>
> Ref: The L-band noise bump described in C127766.
>
> -Mert
> -------- Original Message --------
> Subject: Re: [evlatests] Antenna 5 Bandpass
> Date: Mon, 10 Oct 2011 12:27:52 -0600
> From: Dan Mertely <dmertely at nrao.edu>
> Organization: NRAO
> To: Ken Sowinski <ksowinsk at nrao.edu>
Certainly about the right length (~100 ft for
1/2 wavelength). There are a lot of other
splices and connectors along the way, however.
In addition, the tap was a -10 dB hybrid as a
splitter, going into 2 of the old blue Miteq
L-band post amps (in series) via a isolator,
before hitting the long stretches of 1/2 inch
heliax. The main (EVLA) signal path shouldn't
have seen any back reflections through all that!
-Mert
> -------- Original Message --------
> Subject: Re: [evlatests] Antenna 5 Bandpass
> Date: Mon, 10 Oct 2011 12:11:10 -0600 (MDT)
> From: Ken Sowinski <ksowinsk at nrao.edu>
>
> About the length of a cable from the vertex room to
> a connector at the gound?
> Rick Perley wrote:
> In reviewing my L-band 'Herc A' data, I fine the fine-scale ripple
> present in the data taken last March, but apparently absent in the data
> taken last December 17. The only caveat (and the reason for
> 'apparently') is that the data taken on 17 December had a coarser
> resolution -- 2 MHz -- rather than the 1 MHz of the other data sets.
> The frequency period of the ripple is about 5 MHz, so it is just
> possible that the 2 MHz resolution smoothed the ripple in the 17
> December data. I regard this as unlikely.
> _______________________________________________
> evlatests mailing list
> evlatests at listmgr.cv.nrao.edu
> http://listmgr.cv.nrao.edu/mailman/listinfo/evlatests
> -------- Original Message --------
> Subject: Fwd: may be i stand corrected
> Date: Tue, 4 Jan 2011 22:02:58 -0700
> From: Joseph P. McMullin <jmcmulli at nrao.edu>
>
> Sorry, not quite done perhaps. The original plots did not use continuous
> lines as we'd done
> before; when you do that you see ea05 is spikier (at around a 10% level)
> in the R feed in both spws...
>
> --
> Dr. Joseph P. McMullin
> EVLA
> Ph: +1 575 835 7315
> E-mail: jmcmulli at nrao.edu <mailto:jmcmulli at nrao.edu>
> http://science.nrao.edu/evla/
> -------- Original Message --------
> Subject: C127766
> Date: Thu, 30 Dec 2010 12:47:53 -0700
> From: Jim Ogle <jogle at aoc.nrao.edu>
> Organization: NRAO
> To: <nrao-rfi at aoc.nrao.edu>, <echavez at nrao.edu>, <jjackson at nrao.edu>,
> > <rlong at aoc.nrao.edu>
>
> Work status on work order number C127766 is CLOSED on 12/30/2010
>
> Work Requested: Failure - from JIRA, all baselines with ant05 at L-band
> showing severe bandpass ripples. Seen during RSRO.
> 10/28/10 - only RCP, both IFs.
>
> Corrected Action: Adjusted or Completed. Ripple was previously found to
> move from RCP to LCP with move of W8 monitor tap. Today (12/30) I
> replaced the -10 dB directional coupler and added an isolator to the
> tapped output (ahead of the 2 W8 monitor Miteq amps). The entire tap
> circuit was moved back to RCP. RCP should be checked again for the
> ripple. The IFA & IFB bandpasses look "normal" afterward, using the
> bandpass plotter to verify correct installation. -Mert
> -------- Original Message --------
> Subject: Re: L band
> Date: Mon, 29 Nov 2010 15:25:34 -0700
> From: Rob Long <rlong at aoc.nrao.edu>
> Organization: NRAO
> To: Dan Mertely <dmertely at aoc.nrao.edu>
>
> Do you have an amp between the coupler and the building?
>
> Rob
> Dan Mertely wrote:
> Hi Guys. I appreciate the ideas, but at this time I
> think there is no need to get too fancy. In the next
> day or so I'll get up there & replace the short RF
> cables used, and clean-up all the connectors. If that
> doesn't eliminate the ripple, I can change out the old
> directional coupler for another--perhaps one with better
> isolation. It's a simple system, I'd hate to complicate
> it.
>
> -Mert
>
>
> Rob Long wrote:
>> I just checked with ALMA after lunch to see if they have an older
>> optical TX/RX pair that we could try out with Dan's setup. I have a
>> box of parts to look through and I'll let you know what I find.
>>
>> Rob
>>
>> Jim Jackson wrote:
>>> Dan,
>>>
>>> Is this something that may be easy to fix by adding filters, DC
>>> blocks, isolators, etc to the W8 monitor cable?
>>> The other option is to move the W8 monitor to fiber using a pair of
>>> 2 GHz analog optical transmitter/receivers. This might be nice since
>>> it could be moved to any antenna location, not just W8.
>>>
>>> Cheers,
>>> Jim
>>>
>>> At 05:27 PM 11/24/2010, Joseph P. McMullin wrote:
>>>> Hi,
>>>>
>>>> Sorry for the long delay but we finally got our homework done on L
>>>> band (I ran a 30 minute
>>>> L_osro yesterday to make sure I had adequate time on source to
>>>> review the pattern).
>>>> Both Urvashi and I looked at the data and the quick summary (from
>>>> her; I did a longer integration cycle and still
>>>> see a dramatic difference with the LCP):
>>>> ---
>>>> Now, the 'L' polarization shows a ripple, and 'R' does not. I think
>>>> this is consistent with what you described was done.
>>>>
>>>> I looked only at scan 30, spw 0 and did 5-second calibration
>>>> solutions. I plotted amp(gain) vs channel . Each plot consists of
>>>> an overlay of several 5-sec gain solutions - they overlap => no
>>>> time-variation within the scan.
>>>>
>>>> Attached plots :
>>>> EA05 : Orange = R pol, Blue = L pol. Ripple is only in L pol.
>>>> EA10 (just for comparison) : same as above - no ripples.
>>>> ---
>>>> Plot is attached.
>>>>
>>>> Joe
>>>> --
>>>> Dr. Joseph P. McMullin EVLA Ph: +1
>>>> 575 835 7315
> -------- Original Message --------
> Subject: Re: EA05 L-band ripple
> Date: Wed, 27 Oct 2010 14:20:35 -0600
> From: Urvashi R.V. <rurvashi at aoc.nrao.edu>
>
> Hi Dan,
>
> - Both IFs, "R polarization" only (seen after making gain solutions).
> - Vertical scale is linear in plot_bandpass_ea11_ea05_ripple.png
>
> Two more plots are attached:
> - Amp (gain solutions) vs channel-number for EA05, spw_04 (AC) and
> spw_13 (BD), (leaving out first 3 and last 3 channels in each spw).
> - Linear scale for the y-axis.
> - Ripples are in the R polarization only.
> - Plot panels are 10 seconds each (left-to-right, top-to-bottom) - to
> show that these don't change much with time.
> - Phase vs Frequency shows a similar ripple, about +/- 20deg (from
> gain solutions)
> - Calibration corrects for this, although residuals on EA05 baselines
> (after imaging) are about 40% higher than the others.
>
> Please let me know if you need more info on this.
>
> Thanks.
> Regards
> Urvashi
> (उर्वशी)
More information about the evlatests
mailing list