Re: [Ifeffit] Ifeffit Digest, Vol 162, Issue 17
Topic 1: We have a routine /edit within ifeffit (our modified version) which propagates and fits uncertainty.
Working on a couple of minor details before passing it on to Matt and Bruce for general use.
Topic 3: Monochromator glitches should be eliminated in absorption spectra under normal circumstances if the dark current estimation is well defined. I recommend careful measurement and recording of this for all spectra.
------------------------------------------------------------
Christopher Chantler, Professor, FAIP
Editor-in-Chief, Radiation Physics and Chemistry
Chair, International IUCr Commission on XAFS
President, International Radiation Physics Society
School of Physics, University of Melbourne
Parkville Victoria 3010 Australia
+61-3-83445437 FAX +61-3-93474783
chantler@unimelb.edu.au chantler@me.com
http://optics.ph.unimelb.edu.au/~chantler/xrayopt/xrayopt.html
http://optics.ph.unimelb.edu.au/~chantler/home.html
________________________________________
From: Ifeffit [ifeffit-bounces@millenia.cars.aps.anl.gov] on behalf of ifeffit-request@millenia.cars.aps.anl.gov [ifeffit-request@millenia.cars.aps.anl.gov]
Sent: Friday, 12 August 2016 11:22 PM
To: ifeffit@millenia.cars.aps.anl.gov
Subject: Ifeffit Digest, Vol 162, Issue 17
Send Ifeffit mailing list submissions to
ifeffit@millenia.cars.aps.anl.gov
To subscribe or unsubscribe via the World Wide Web, visit
http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
or, via email, send a message with subject or body 'help' to
ifeffit-request@millenia.cars.aps.anl.gov
You can reach the person managing the list at
ifeffit-owner@millenia.cars.aps.anl.gov
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Ifeffit digest..."
Today's Topics:
1. Statistical errors in linear combination fits (Joshua Kas)
2. Re: Statistical errors in linear combination fits (Bruce Ravel)
3. Athena interpolation when removing mono glitches
(Michael Gaultois)
4. Re: Athena interpolation when removing mono glitches (Bruce Ravel)
----------------------------------------------------------------------
Message: 1
Date: Thu, 11 Aug 2016 13:31:45 -0700
From: Joshua Kas
Topic 3: Incorrect dark current (offset) is just one of many ways that glitches can escape being normalized out, and one of the few you can correct in software. Others include but are not limited to 1. Any detection-system nonlinearity in I0 or detected channel (F or T). This includes deadtime in fluorescence detectors. 2. Inadequate detection speed in QXAS so that the recording of I0 lags behind that of fluorescence (or vice versa if using a Lytle detector or similar). 3. Harmonic/stray beam content varies while going through the glitch. 4. Beam position or shape varies while going through the glitch. If the crystals are in a diverging beam, then each part of the crystal sees a slightly different incidence angle, so the exact energy at which the two reflections contributing to the glitch both satisfy the Bragg condition varies across the crystal surface, resulting in non-uniform reflection. There are undoubtely many more. mam On 8/12/2016 2:36 PM, Christopher Thomas Chantler wrote:
Topic 1: We have a routine /edit within ifeffit (our modified version) which propagates and fits uncertainty. Working on a couple of minor details before passing it on to Matt and Bruce for general use.
Topic 3: Monochromator glitches should be eliminated in absorption spectra under normal circumstances if the dark current estimation is well defined. I recommend careful measurement and recording of this for all spectra.
------------------------------------------------------------ Christopher Chantler, Professor, FAIP Editor-in-Chief, Radiation Physics and Chemistry Chair, International IUCr Commission on XAFS President, International Radiation Physics Society School of Physics, University of Melbourne Parkville Victoria 3010 Australia +61-3-83445437 FAX +61-3-93474783 chantler@unimelb.edu.au chantler@me.com http://optics.ph.unimelb.edu.au/~chantler/xrayopt/xrayopt.html http://optics.ph.unimelb.edu.au/~chantler/home.html
________________________________________ From: Ifeffit [ifeffit-bounces@millenia.cars.aps.anl.gov] on behalf of ifeffit-request@millenia.cars.aps.anl.gov [ifeffit-request@millenia.cars.aps.anl.gov] Sent: Friday, 12 August 2016 11:22 PM To: ifeffit@millenia.cars.aps.anl.gov Subject: Ifeffit Digest, Vol 162, Issue 17
Send Ifeffit mailing list submissions to ifeffit@millenia.cars.aps.anl.gov
To subscribe or unsubscribe via the World Wide Web, visit http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit or, via email, send a message with subject or body 'help' to ifeffit-request@millenia.cars.aps.anl.gov
You can reach the person managing the list at ifeffit-owner@millenia.cars.aps.anl.gov
When replying, please edit your Subject line so it is more specific than "Re: Contents of Ifeffit digest..."
Today's Topics:
1. Statistical errors in linear combination fits (Joshua Kas) 2. Re: Statistical errors in linear combination fits (Bruce Ravel) 3. Athena interpolation when removing mono glitches (Michael Gaultois) 4. Re: Athena interpolation when removing mono glitches (Bruce Ravel)
----------------------------------------------------------------------
Message: 1 Date: Thu, 11 Aug 2016 13:31:45 -0700 From: Joshua Kas
To: ifeffit@millenia.cars.aps.anl.gov Subject: [Ifeffit] Statistical errors in linear combination fits Message-ID: Content-Type: text/plain; charset="utf-8" Hi All, I was wondering if it is possible to pass a value for the uncertainty in the data to Athena when doing linear combination fits. We have calculated chi^2 values using a simple estimate of the statistical uncertainty, and find that our values differ by several factors of 10 when compared to the reported values from Athena. I assume that this has to do with value of the uncertainty that Athena is using, but I certainly could be mistaken. In any case, the reduced chi^2 reported by Athena is much smaller than 1, while the fit is off by fairly large amounts compared to any reasonable estimate of the statistical error. Thanks in advance for any help and sorry if this info is already available in the archives, Josh Kas
Hi Chris, On Fri, Aug 12, 2016 at 4:36 PM, Christopher Thomas Chantler < chantler@unimelb.edu.au> wrote:
Topic 1: We have a routine /edit within ifeffit (our modified version) which propagates and fits uncertainty. Working on a couple of minor details before passing it on to Matt and Bruce for general use.
Thanks. At this point, Larch *should* be able to do this too. That is to say, the feffit() function in Larch can propagate uncertainties in chi(k) through the EXAFS analysis steps. It definitely needs more testing and I'm not at all certain this is done the same way you're doing it. Currently, Larch does not currently propagate uncertainties in measured intensities to uncertainties in mu(E) in any automated way, and also doesn't propagate uncertainties in mu(E) to uncertainties in chi(k). The basic tools to do these are available, and I think it would not be hard to add these features. A simple test case of full error propagation would be really helpful. --Matt
participants (3)
-
Christopher Thomas Chantler
-
Matt Newville
-
Matthew Marcus