Floating HP 6113A still measures voltage between positive terminal and GND

[Tim T]

I picked up a HP 6113A power supply with known problems in order to learn a bit about trouble shooting.   A new 3000 uF electrolytic cap did a lot to improve power output beyond the 10 watts it was previously capable of putting out ( half of what was expected), and to clean up a 120 Hz spike measured off the output terminals.

All was looking pretty good until I removed the strap across the rear GND and negative terminals in order to float the power supply, and I am now measuring negative voltage across the positive terminal and GND. It starts out as only a few mV but gets to -0.5 V ( and continues to go more negative) after about 5 minutes of warm up.

I have three other vintage power supplies that can be floated ( two HP 6216As and a Power Designs  2020B) and don't see this behavior on any of them.

This Keysight article talks of placing resistors and/or capacitors between the output terminals and ground in order to deal with stray ESD and RFI http://powersupply.blogs.keysight.com/2014/11/why-do-i-measure-voltage-to-earth.html  and from what I can tell such resistors/capacitors are lacking in the 6113A, so on the surface this doesn't appear to apply.

So my question boils down to might there be another problem to solve in this 6113A or is this normal behavior and if so why?

Thanks,

tim