FT-857D filter plaque

I noticed that my FT-857D suddenly had a lack of sensitivity in FM. Looking a bit closer i experienced that only strong FM signals could be received on all bands. The other modes were ok.
Reading a little bit gives an impression what happens. The most common fault seems to be that one of the Toko ceramic filters (one of 3) fails. Usually this is CF1005 which results in low sensitivity and/or crackling noise in AM and FM. For me it was different. No problems with AM. Therefore i expected CF1003 (narrow band FM filter) failed. There is a very good description of the reasons written down by SV8YM.
external link: http://sv8ym.blogspot.com/2010/07/myster…ng-filters.html
Today i disassembled the transceiver and measured the resistance of the filter ports to ground. CF1005 was around 14kOhms both ports, CF1002 around 55kOhm both ports and CF1003 several tenth of kOhms one port and the other 150Ohms only. This causes the pindiode switch to fail which results in very bad sensitivity.
When i tried to disassemble the filter from the board for accident i removed the cap of the filter. Thus i decided to take a close look without removing the filter from the PCB. Find some photos below.
What is visible are the small pieces of material present on the edges of the ceramic plates. This was descibed by SV8YM. One of the plates showed a very big amount of material that forms the short that was measured before. Therefore i removed one plate after each other, cleaned them with a scalpell knife and placed it back to the filter. Special care needs to be taken on the correct position of the small metal plates that contact the ceramic elements with the filters mechanical structure.
After mounting the filter cap again and measuring the resistance it can be seen that it is in normal range again. After assembling the transceiver FM is ok again.
Overall i assume that this repair will not help for a long time because the filter is already damaged. But at least it helps until i collected all necessary spare parts. The problem in general is that the voltage for switching the pin diodes for the RF paths is not decoupled from the filters. DC voltages destroy ceramic filters. This was not taken into account for all filters by the designers of Yaesu. Most probably because of an missing hint in the datasheet of the filter.
I will change the circuit later when replacing the filters.

CF1003 filter cap removed

CF1003 filter cap removed

metalized surface of one of the filter plates

metalized surface of one of the filter plates

bad crystal on the edge of one of the plates

bad crystal on the edge of one of the plates