#16 fit parameters do not get errors when fixing others

[sebastian_busch]

I fit the function

ELSEIF (ifc.eq.76) THEN ! Diffusion (x) internal
pi = twopi / 2.
DO i = 1, n
Y(i) = dabs(P(1)) * dabs(P(2)) * (1./pi) *
* dquot0(dabs(P(3)),(X(i)-P(6))**2+P(3)**2) +
* dabs(P(1)) * (1. - dabs(P(2))) * (1./pi) *
* dquot0((dabs(P(3))+dabs(P(4))),(X(i)-P(6))**2+(dabs(P(3))+dabs(P(4)))**2) +
* dabs(P(5))
ENDDO

to my data, convoluting with vanadium, weighing with errors. both, van and sample, have errors:

# name unit s value
1 a 1.0000
2 A0 0.50000
3 w1 1.0000
4 w2 ! 10.000
5 bg ! 0.00000
6 c ! 0.00000

Modify parameters [-] ?
stored 16 spectra as file 3
Range of fit (* = full range) [*] ?
Weighting with x-stepwidth [n] ?
Weighting with y-error [n] ? y
Logarithmic weighting of [n] ?
Convolute with resolution [n] ? y
stored 16 spectra as file 3

cf
ci 3 (w1)
--> has no error bars

cp release 4 (w2)
cf
cp set 4 (w2) back to 10
fix 4 (w2)
cf

ci 3 (w1)
--> has error bars

[sebastian_busch]

Logged In: YES
user_id=1528569
Originator: YES

sometimes even releasing 'w2' does not create error bars. in that case, releasing and re-fixing 'c' helped.