*SOLVED* Power supply grounding problem?

[Aethelstan]

** EDIT I have tried a different PSU and the problem persists **
** EDIT 2 Solved by tying the PGM pin to gnd. Funny how the problem manifested itself **

Hi, I have a dual rail power supply made using a centre tapped transformer a pair of LM317/337 that I have built myself. I was trying to use it to supply a PIC18F on a bread board and flash an LED, which would work until I disconnected my PICKit3. After a bit of trial and error, I discovered the circuit will work correctly if I had the ground connected to the PICkit or to my oscilloscope. The power supply does power the breadboard, but the program doesn't run. If I connect the oscilloscope probe to my power supply with the ground part of the probe connected to the grounding tab on the scope, I see a +-8V wave at mains frequency, and adjusting the power supply does not change the magnitude of the wave. If I connect the ground part of the probe to the power supply ground, then I see a nice flat linear voltage that seems well regulated. I have checked continuity of the transformer, and also between the ground terminal and the centre tap.

To my mind, the ground reference of the PSU is not staying at 0V, but I have no idea why.

I have attached a schematic, the 12V supplies represent the centre tapped transformer. J2 and J3 are connectors for the potentiometers to control the voltage. J1 and J4 are connected to the LED voltmeter modules.

Thanks for any advice.