fixing scale error in efex taus, and updating the counts

Since the counts changed, and I did not have trig validation in my sparse checkout, this is a new MR (from the closed !46713 (closed) )

As pointed out by @asonay , there were no Tau TOBs passing to L1Topo. This was found to be due to a recent change to use the internal firmware energy scale for calculations in the algorithms. This was not factored into the Taus, meaning their et values were 25x too small. With this merge, this has been fixed. Also tagging @nicholas

Additionally, as suggested by @jburr , the scaling was moved into the formTOBs classes, where the tob and xtob scales are stored in the headers.

This is an urgent fix, as the taus are needed for ongoing developments of the combined taus in L1Topo. @dzanzi .