Hi,
here are some plots:
http://www.slac.stanford.edu/~kerstin/vubfit/spec_nonnormalized.eps
http://www.slac.stanford.edu/~kerstin/vubfit/mcspec1_rescaled.eps
http://www.slac.stanford.edu/~kerstin/vubfit/mcspec2_rescaled.eps
>From the unfolded spectrum:
*****************************
the calculation I was interested into is
N(b->c lnu)=88 Mi (BB~)*0.216=19 Mi (B->clnu)
[I did not put the factor 2 in the number of Bs because I suspect
you do not have it in your efficiencies - can you confirm?]
which means that the fraction of 'rho'lnu/b->c lnu would be, from
the raw spectrum 42995.5/19,000,000=2.3 10^3
which is reasonably close to the exclusive average between rho+ and
rho^0 (2.2 10^-3). The bias correction would already bring it down to 1.7
10^-3 which is already ackward.
>From the MC:
*************
The numbers for the third bins are
79699.7 +- 49406 for the first and
152941 +- 56278.3 for the second, dividing by 19 Mio we get
(4.2+-2.6)10^-3 and (8.0+-3.0)10^-3, respectively.
>From the BFs in generator hybrid signal MC I find 1.8 10^-3
(with B->Xlnu of 10.8%). So the first one has this within one sigma, the
second one is about 2 sigma away (but is larger than what is in the MC).
Here are the numbers I used for the scalings:
Nsl(data) = 33677.8
Nsl(MC1) = 58106.1
Nsl(MC2) = 58372.4
Br(b->u)(data) = 0.00227
Br(b->u)(MC) = 0.00135
Kerstin
|