Today I measured the balancing of the OMC DCPDs, using an excitation at 103Hz into the OMC length servo, and measuring the height of the second harmonic in the nullstream channel.  The DCPDs are only, as far as I can tell, 1% misbalanced.  In the attached plot data are in red, the blue is a by-eye fit.  I have set the balance to be 50.5% PDA, 49.5% PDB (1% towards PDA on the slider).