Procedures Automatic Fitting <Prev Next>

Other Automatic Fitting Notes

(Draft)

Nick: Data_180711_8X.ovr

Plot Units: Mhz
Baseline: WIndow 0, Noise 1 is OK
Also try: (slow!) Window 5000, Noise 1
For peak finding, try Noise Mult 2.0
Use Delete and measure at each end

AgI

First Step

Unit MHz, Microwave
B = 1342 (float)
Zoom in: 13160-13540
T=2
I = 5/2
eQq0 = 500  (just to see width) (float)
Click on 5 most intense lines
Select first 2
Window = 100
Accept = 0.1 (Default)
AgI1.pgo

Search: 729 trials; First one OK (others rubbish)

Fitting All Lines

Fit - look good
Transitions Window: "Replace"
Discard, "Low Intensity"
Discard, "Outside line list range"
Select all
Window 0.1
Search
Fit, floating D as well
All, Update
Assigns additional 8
Use lots of zoom, Show and Edit probably spurious assignment
- remove by clearing "Std Dev"
AgI2.pgo
Clear unset and Save to file AgI1.obs in (say) simple format.

Other Isotope

Zoom 7750-8300
Float B only
Click on three most intense lines
Select first 1
Window = 200
AgI3.pgo

Search - two good fits corresponding to two isotopes
Cycle as above
AgI4.pgo, AgI2.obs

CF3I

First Step: K=0

Symmetric top, Microwave, Units MHz
T=2
B = 1500 MHz
Need C sensible, so try C=5000 (Must be small enough to give K=1 population; exact value not important.)
Jadjustsym off
I nucleus I=5/2
Zoom 8700-9500
eQq = 500  (just to see width) (float)
To speed up, set IntensityUnits = Arbitrary, Jmax=10
Right click on (say) 5 most intense transitions, picking K = 0 only.
Select first 2
Window = 400
CF3I1.pgo
90,000 searches, so speed up with second linelist, generated as above with multiplier=200
To use this, hide inactive window
Not only is this faster, but it fails to find the true match
Gives a few plausible matches - take one with eQq = -2144 and fit
CF3I2.pgo (Could fit some more K=0 lines, but better to confirm model first)

K=1

Need C sensible, so try C=15000 (Must be small enough to give K=1 population; exact value not important.)
Now pick a few strong K=1 lines; Use trial and error or transitions window.
Now select all the assigned lines so far, and one K=1 line
Fix eQq, float DJK
CF3I3.pgo
One good fit;

K=2

Fit again including eQqComplete fit using transitions window, but using "Discard, Outside plot range"
Narrow window to 0.1
CF3I4.pgo

J=3

Clear unset; shift window to 11800-12500
Click on strongest line
Float DJ
CF3I5.pgo
Search - may need to widen window
Add lines for this J and fit
CF3I6.pgo

All J

Repeat for 15100-15450 region

This reveals peak finding issue for peak at 15253.6; use Show and edit from residuals window, and manually measure peak from original overlay.
Also float Cperp

For J=5 region at 18300 need to use original linelist