Ken Scharf wrote:

I added a diode in series with the VPP output from the emitter or Q1
(with the anode toward Q1's emitter) and a second shut diode
from the MCLR pin to VSS (anode toward VSS). Now the VPP voltage
at MCLR swings from -0.05v to +12.6 volts. I hope that's enough
VPP. I may have to change out the 8.2v zener, I measured the voltage
across it and it appears on the low side. I could also just connect
a 1n4001 or two in series with it to raise the vpp supply.

Anyway seems the programmer is safe now. Still need to put a scope
to the PGC and PGD lines and check they are in spec.

When I modified my PicStart Plus Microchip specified the voltage should
be 12.75-13.25. Microchip had a mod to lower the voltage from nominal
13.75 to nominal 13.0 for the newer chips (new about 7 years ago).