breaking the loop with diffstbprobe

Hello, I have a linear regulator with 2 loops, one within the other. For stability analysis if I break a wire with iprobe which breaks all loops I get a certain answer of unity gain freq, phase margin and gain margin. So that works. If i then break the loops at a differential point of the same inner loop such that all loops are broken again but use the diffstbprobe component my equations fail because there is no unity gain crossover point so I suppose the complicated circuit suddely seems first order! Which I think is unlikely. So as an experiment if I break the loop again at the single wire point but put in 2 vcvs components so the single wire break looks differential then use the diffstbprobe component again I get answers for UGF,PM and GM. (Perhaps its wrong to use vcvs for this purpose as it forces the loop gain sense to be voltage instead of voltage and current as middlebroke requires and also the bilateral nature of the break may artificially become unilateral but it still shows the diffstbprob can work). Anybody had success breaking differential loops?

Cadence version is 6.1.5-64b.500.8

Rgds

Lance