[BioC] Loess normalization at the probe set level ... how to?

[Emmanuel Levy]

Hello,
I just read a paper that I actually strongly recommend:
"Preferred analysis methods for Affymetrix GeneChips revealed by a
wholly defined control dataset" PMID 15693945

In this paper they advice to do 2 normalization steps, one after the
background correction and one on the expression summary (using the
loess method).
I am not sure about how to do the latter (loess norm on the exp. summary). 
I have 2 conditions (say A and B) and triplicates (say A1, A2, A3 and
B1, B2, B3).

Once I get the expression summary:
eset.mas = mas5(affybatch)

How should I perform the loess normalization step?

Thank you very much in advance for your help,

   Emmanuel