Odd changes in readings during insulation resistance testing

Still waiting for account to be passed so can’t reply directly. 

have updated the post with regards to questions. 


During insulation resistance testing I was getting very low readings whilst the cores were left terminated, removing them from the terminals caused the  readings to increase dramatically. I have never come across this before, does anyone have any suggestions of possible causes? 


Split load metal consumer unit - British General

Flat PVC cabling, modern cabling. masonry and stud walls

All MCBs off, 

RCD off. 

RCDs do not trip when installation is energised. 


Firstly, with Line and neutral bridged testing against earth. All cores left in the terminals. This gave a very low reading, which I tried to further narrow down by separating line and neutral and testing each down to earth. 


The cores left terminated in the top of the MCB, neutral bar and MET. MCB open, RCD open. 

@500V

Between line and earth ~0.3 MΩ

Between neutral and earth ~0.3 MΩ


@250

Between line and neutral ~0.07 MΩ (USB sockets still connected) 

This was done just as a control. 


Cores removed from the terminals. Including CPC removed from the MET. 

@500V

Between line and CPC >999 MΩ

Between neutral and CPC  >999 MΩ


@250

Between line and neutral ~0.07 MΩ (USB sockets still connected)

Proves there is an issue within the CU with earthing. 


I have not tested from the RCD’s terminals to the MET. 

However, I did test from the top of each MCB (under the same RCD) to MET which all gave the same low reading. But if they are open there should be no path back to the RCD. 


On the other RCD a lighting circuit gave the same results, extremely low when terminated then rose dramatically when removed from terminals including earth core from MET. 


From the results I can only assume something is wrong in the CU, but what could cause such readings? 

Regards