3 ply thermal sandwich with copper plate in middle on a Delta-Q 21A charger fail

[Labrat101]

Hi I had a  chargers to repair .
 Delta-Q 913-3600 . 21Amp @ 36volts.  Date 2008 .
The Thermal isolation between the power transistors to the main heat sink consisted of
some what a strange setup . There are 3 layers of insulation that is not like the normal Mylar insulation material but more like coppered heat tape . with a thickness of 0.05mm   . The problem is they put a copper sheet between the sheets and over time the the last diode after the transistor arced though the sheet to the copper insert . ( BanG)
the copper sheet is connected via 10 resistor to the emitters of the 3 power transistors .
I can not think of any reason that this scenario has . Over the normal single isolation sheet mylar or similar .
I have put up some photo's if some has any idea's why or what this odd array set up has to offer .
The other 4 transistors have the same setup with the thermal sandwich . also with the copper plate connected to the
Emitters via a smd 10 resistor .
The whole metal casing is mains grounded .
 
Picture 1 is the 2 plastic sheets they are soft and flexible . and the copper sheet.
picture 2 is the placement order without the paste
 & main heat-sink with another thermal sheet also 0.05mm thick and the locking plate.

I want to just replace this setup for a single thermal pad without the copper sandwich .
 Any idea's for this . or is it a Chinese patent

[Labrat101]

Basic I just need to know the reason for this topology . I have never come across this method on heat sink
cooling before . . I don't think its to do with interference or RF emission .
So I was hoping that someone  else has seen this before and knows the reasons behind this type setup .

Thanks

[Labrat101]

Thanks .
The only trouble was when I reassemble I used high grade thermal paste . Not metalized. The sheets they used are only 0.05mm thick.
And we use 240v/250v here so the voltages on the primary is around 375v after the large caps totaling 440uf . . The end transistor is a diode from the HV . So there sheet failed.  I'm not sure how it ever managed to last as long as it did.
What you said sounds probably correct.  Built for failures.  Some how I will have to find a better solution.
If you have any idea on solving this . I am open for ideas.
All the other power units I have the heat sinks just one layer and the heatsinks are to neutral/ground .
 This type of unit is normally fitted inside an electric car . The mobility type .