After testing 4P1L with the 126C OPT, Paul Leclercq suggested a different operating point to see if distortion above 8Vrms on the OPT version was due to grid current. So firstly did a quick test with the CCS and found that -15V and 170V was a good operating point for 15mA. The 4P1L performs brilliantly with CCS and THD is only 0.05% when driven to 10Vrms.
Unfortunately distortion with 126C grows exponentially above 9Vrms. OPT is tested floating and not sure if is the Pete Millett interface a limitation with it.
EDIT
Andy Evans suggested that secondary green should be negative so phase is 180. Aparently there is a thread meantioning the distortion introduced by the way I was using the OPT.
Anyway, problem sorted! Look at now how nicely the 126C can perform at 10Vrms:
THD=0.12% at Vo=75Vpp. Can’t test higher without an additional stage due to my external oscillator can only provide nearly 3Vrms.