[daip] ACCOR

[Leonia Kogan]

Hi Eric,

>As you predicted, there are corresponding differences. 

This fact explains the SN table output of ACCOR.
Now there is the question: why this difference has a place

>I can send you some plots, if you want. 

Yes, I want

>ACCOR solutions are inversely proportional to the amplitudes of the 
>auto-correlation.

Yes, This is what ACCOR should produce

>But I am not so clear about how to get the number in SN table. It is said 
>that the
>mean values of the AC should be close to the unit. So, I guess this is a 
>relative
>number. If so, which one is the reference?

This number is relative to the ideal sampler at the digital converter of the 
given antenna and If at the given time. There is no reference. I repeat 
this number is absolute number which shows the distinction of the given 
digital converter from the ideal one. If the number is equal 1, then 
the digital converter is ideal.

>The next question is what is the cause for such a difference we see in the 
>ACCOR
>solutions. Or, why does the AC amplitude change at MK?

This is a question we have to answer here having analized your data

>What is the difference between setting
>digicor=1 in FITLD and running ACCOR?

You have to put DIGICOR=1 for future running of ACCOR.
Having DIGICOR=1, FITLD provides the digital correction including 
multiplication by the coefficient which makes the mean value of the spectra 
equaled 1 for VLBA correlator and for ideal digital converters.

Leonia



----- Begin Included Message -----

>From zshen at center.shao.ac.cn Thu Dec 11 23:43 MST 2003
Date: Fri, 12 Dec 2003 14:43:12 +0800
From: "Z.-Q. Shen" <zshen at center.shao.ac.cn>
Subject: Re: [daip] ACCOR
In-reply-to: <200312111642.JAA03295 at bonito.aoc.nrao.edu>
To: Leonia Kogan <lkogan at zia.aoc.nrao.edu>
Cc: daip at bonito.aoc.nrao.edu
MIME-version: 1.0
X-Mailer: QUALCOMM Windows Eudora Version 6.0.0.22
Content-transfer-encoding: 7BIT
References: <200312111642.JAA03295 at bonito.aoc.nrao.edu>
X-MailScanner-Information: Please contact postmaster at aoc.nrao.edu for more information
X-MailScanner: Found to be clean
X-MailScanner-SpamCheck: not spam, SpamAssassin (score=-6.3, required 7,
	BAYES_01 -5.40, EMAIL_ATTRIBUTION -0.50, IN_REP_TO -0.37,
	REFERENCES -0.00, REPLY_WITH_QUOTES 0.00)
X-Lines: 60
Status: RO

Hi Leonia,

I tried the POSSM for both calibrators scans and SgrA scans. As you predicted,
there are corresponding differences. I can send you some plots, if you 
want. The
ACCOR solutions are inversely proportional to the amplitudes of the 
auto-correlation.
But I am not so clear about how to get the number in SN table. It is said 
that the
mean values of the AC should be close to the unit. So, I guess this is a 
relative
number. If so, which one is the reference?

Though there are differences in the amplitude of the AC data, almost all 
the AC
data have the same bandpass shape, suggesting the bandpass response is stable.

The next question is what is the cause for such a difference we see in the 
ACCOR
solutions. Or, why does the AC amplitude change at MK? Is this related to 
the data
recording, or the correlation? I saw this happened in another 3mm 
experiment too.

Furthermore, I see there is a parameter in FITLD, digcorr, to control the 
application
of the VLBA correlator's digital correction. What is the difference between 
setting
digicor=1 in FITLD and running ACCOR?

Cheers,
Eric


At 00:42 2003-12-12, Leonia Kogan wrote:
>Hi Zhiqiang,
>
>ACCOR is rather simple program. So I do not think that there is any bug
>in the code (although everything can happen).
>
>ACCOR measure the mean value of the autocorrelation spectra.
>
>So it would be good to check the data itself.
>
>Please, look the autocorrelation spectra at  calibrators times and the SgrA
>time. To do this you need running POSSM
>with APARM(8)= 1 => plot total power
>Use TIMERANG specifiing the given times
>
>I guess you'll see a difference between short and long scans and between
>IF=1,7 and IF=2-6, 8
>
>Let me know the result of this test
>
>Thanks
>
>Leonia
>
>



----- End Included Message -----