Trying to verify a YIG oscillator output amplitude

[jlmoon]

I have been evaluating a case of trouble with a Hp 8563a 10khz to 26Ghz spectrum analyzer and beginning to scratch my head a little.
Could someone give me some clues to why my YIG output would be varying in amplitude when performing some basic span / LO functional test per the Hp service documents.  I have measured both on the 1st LO output (front panel) and directly to the YIG/LO SMA connector.  What I am encountering is this:  I can dial up some frequencies setting the span to zero and when I set center frequency to some areas the YIG LO goes down in amplitude until there is no output.  When working correctly the YIG output is around +6 to +7 dbm.  I have a video that actually shows the LO behaving correctly until it reaches an area of the programmed span and then poof the LO amplitude drops into the noise.  Turn the analyzer center frequency back down a little and the LO fires back up and all is good, approach the edge of trouble and the LO output drops 30 - 40db in level until it disappears.  Currently I am not seeing any signal from 0 to ~2.7 Ghz , 2.8 to approx 3.1 Ghz all is good, go above ~3.2Ghz to the top end of 26Ghz  and the output drops off again.  Isn't it correct to assume the YIG output should not vary in amplitude but only frequency based on the current imposed on the Loop coil.  After all this is the LO input to the mixer which derives your frequency of interest.
 
Any advice would be greatly appreciated.

Thank you,

JLM

[acbern]

The amplitude should be stable within the specs of the manual, what you see is certainly beyond. What you should do is to measure the currents in the coils from the regions that are ok transitioning into the bad regions. if the current jumps (both coils), the driving circuitry (DACs...) seems to have a problem. If everything is ok, what may have happened (and I have not seen that myself, but read about it) the YIG balls within the case may have moved, causing such problems.

[jlmoon]

acbern,

That is my next step exactly, I am going to put my A6302's on those coil lines and look at the currents as this is going on.  It is quite strange that the amplitude is present and at the correct level but only in some areas of the span range.  I hear alot of folks say the YIG's / L.O.  in this particular series of analyzers are subject to high failure rate like around 99% of the time.  Before I spend $1000 on a rebuilt one, I do want to rule any other issues out.   
Thank you for the insight