I cannot see the 'vicious circle' in the derivation of prices. We need
prices to determine prices in any case. The price of bread depends on the
price of flour, labour, etc. In the case of fixed capital, if you use the
Sraffa model and treat fixed capital as a joint product, the "lifetime" of
each vintage of a machine will be one year. If you take a depreciation
model, you can start with a series of lifetimes, from one year to its
physically maximum lifetime, and calculate prices and the rate of profit on
the basis of each. One will be the economically optimum lifetime and you
can consider that competitive pressures select for that in the same way as
they would select the most profitable of any range of technologies. (You
could do the same exercise, considering other inputs fixed and varying the
amount of labour input, say: the economic labour input will be the input
which gives the maximum rate of profit)