The fundamental difficulty in the design of permanent balances is that they are called upon to measure forces and moments with a high accuracy in the whole range up to the maximum load without appreciable distortions in the structure and down to a small fraction of it. For
example, an experiment on a wing with laminar profile can be followed by tests on a bluff body (such as a radar antenna), and the balance itself should be able to measure the drag of both bodies with sufficient precision. For this reason, permanent balances are designed to have a sensitivity equal to 10-4 of the full scale: in this way also in measured quantities hundred times smaller than the maximum range the accuracy does not fall below 1%. (The same problem occurs with the majority of measuring instruments, but while it is quite normal to provide a laboratory with many pressure gages for different pressure ranges, a wind tunnel is equipped with a single null reading balance, which is therefore “permanent.”)
Usually the balance is mounted outside the test chamber and the connections are made with the model suspended with wires or mounted on rigid struts. In practice, there are not two balances which adopt the same system and a full discussion of all the sophisticated configurations would occupy too much space. Most connections can be classified into two types: parallel links and virtual center.