[evlatests] switched power issues

[Paul Demorest]

On 2016-05-25 12:28, Paul Demorest wrote:
> On 2016-05-25 11:36, Paul Harden wrote:
>> On 5/20/2016 3:49 PM, Paul Demorest wrote:
>>> the short story is that I think the practice of leaving all receiver 
>>> cals firing at all times is maybe not so good; this may be causing a 
>>> small gain modulation at the cal switching frequency; we should 
>>> change this default behavior, and test what effect this has on the 
>>> Pdif compression issue.
>> 
>> The results are interesting and fairly convincing.  It seems the next
>> thing to do would be to cycle CAL on and off through the different
>> receivers to see if a particular band receiver is the culprit, or a
>> global effect.
>> 
>> More to the point (thinking out loud), I have always wondered about
>> the low band receiver.  The Tcal injection on the microwave receivers
>> is in the order of 2-3K and confined within the RF components (i.e.,
>> the cryo dewar serves as a nice shield); the low band receiver Tcal is
>> 20-30K, or 10dB higher.  This power is injected into the noise coupler
>> *before* the LNAs, thus little isolation between the LBR receiver
>> inputs and the MJPs (74 MHz) or the P-band dipoles.  This 20-30K Tcal
>> power may well be imposed back to the dipoles, which would be a fairly
>> efficient radiator located just underneath the subreflector and a
>> radiator above the feed horns (in fact, could be a "double whammy")
>> 
>> I have never figured out how to measure if the low band CAL switching
>> power is being radiated by the dipoles into the microwave feeds.  You
>> can see Tcal switching (about 1-2 dB jumps) on the 74 MHz low band
>> outputs and about 0.5dB jumps on P-band.  Therefore, I would recommend
>> repeating your test with the CAL switching to the low band receiver
>> turned off first.  Note that the LBR CAL switching is turned off via
>> the new F318 module, not the F317s like the microwave receivers.  The
>> operator knows how to turn off the low band CAL.  I know from my
>> visits to antennas for low band work, the CAL switching to the LBR is
>> almost always on.  I'm certainly curious myself if the low band
>> dipoles are radiating Tcal into other receivers and if you can detect
>> this.  If so, you have discovered a nasty lingering problem, and best
>> of all ... a very easy fix!
> 
> Thanks for bringing up the low-band cals, I neglected to describe this
> in the last message.. when I did this last week, I did try turning
> them off and I do see the effect you mention.  The S-band Pdif values
> drop by about ~5% (or even more in some cases) when the LBR cals are
> off.  In constrast with the other effect I described, this does _not_
> appear to have any effect on gains.. ie, it's not detectable in the
> cross-correlations, only in Pdif and autocorrelations.
> 
> Doing this test is a little tricky because of the way the executor
> controls the cal devices.  The cal setting selected for the main
> receiver in use is always also sent to the F318, so it's not easy to
> control the state of the LBR cals separately.  Sending a low-level
> command to the F318 works temporarily (for the current scan) but it
> gets reverted on the next scan when the config commands are re-sent.
> Does anyone know a way around this?

Here is a plot from this test observation showing Pdif values vs time 
for ea22 (same antenna I showed before).  These numbers are straight out 
of the SysPower table.  The dashed red lines show times when I sent 
commands to turn cals off or on, and the dotted line show scan 
boundaries.  The S-band cals were left on the entire time.  At the start 
and end of the plot all cals are on (usual mode).  This illustrates a 
few things:

1. When the cals are turned off the measured S-band Pdif drops by ~5%.

2. The LBR cals turn themselves back on at the next scan boundary, but 
all others stay off.  In this state the Pdif is a few percent higher 
than when all cals are on.  This seems consistent with the gain 
modulation and resulting apparent Pdif compression that I described 
before.

3. When all cals are turned back on (third red line) things go back to 
the previous state.

In this test, there was a single subband in each IF, with identical 
tuning.  So we expect IFs A and B (also C/D) to look the same.  This 
does seem to be true aside from a small difference in the overall power 
level.

-Paul



-------------- next part --------------
A non-text attachment was scrubbed...
Name: ea22_pdif.png
Type: image/png
Size: 88513 bytes
Desc: not available
URL: <http://listmgr.nrao.edu/pipermail/evlatests/attachments/20160526/e5964fb9/attachment-0001.png>