I think I have found the reason why the gigafil used on the CMU200, in contact with the microwave absorber, stop working properly!
I have a CMU200 and a few days ago, after turning it on after a long period of being turned off, I noticed that the spectrum analyzer was displaying spurious signals with nothing connected.
After running the selftest I discovered that it gave me an error on the TXLO3LEVD voltage.
After a quick search on the net, I discovered that this is a common error on the CMU200 caused by the microwave absorber sponge which tends to degrade and corrode some Murata Gigafil contained on the analyzer's RXTX board.
So I unsoldered the two filters in question and checked them at the VNA, noting that, in both, the bandwidth was much wider than normal (120 MHz instead of 25 MHz).
It has been written that the sponge corrodes the filters as it degrades but, at least for those of my analyzer, not in such a way as to completely destroy the external copper shell of the filters.
So, I asked myself, what is the reason that generates the malfunction?
Then I looked at them under the microscope, after cleaning them with isopropyl alcohol, to get rid of the sponge residue.
And I made a discovery: as the microwave absorber degrades, it creeps into the side holes of the filter, closing them.
Both filters used consist of three cells. And laterally to the filters there are three large holes on one side and three small holes on the opposite side. In the 1842 MHz filter, for example, small holes have a diameter of about 0.4 mm in diameter. If the conductive sponge closes at least one of these small holes, it behaves like a short circuit of the corresponding cell, distorting the behavior of the filter.
It was enough to unclog the small hole blocked with a copper wire to restore the correct behavior of the filter! Also, to be safe, I kept the filter wet for a few hours in isopropyl alcohol, to remove all sponge residue from the holes.
I hope I have been of some help to those who have similar problems in the future.