Hi all,
Because the cov19, I finally have time for playing with my MS2665C SA.
After looking around for the service manual, I finally come to this post here.
My unit look pretty much the same symptoms like your guys said and the RF board look very similar to jimon posted.
So may be I can share what I had found, may be someone someday need these information.
Lets share my story.
I got my unit from the china online second hand shop.
The display level is over 0 dbm over the whole frequency range. The level cal cant help, but the displayed level will change after each level cal.
The interesting thing is that if I connect a external signal, say 5GHz 0dbm to the input while doing level cal, the level seems more reasonably normal, at least I can see the spectrum peak response on the correct frequency. I suspecting the cal signal cant switch in to the RF path, but confirm I am wrong after a simple check.
I study the block diagram, and check if the signal can reach before the IF output, when operating at zero span. Luckily, IF level follows the SG level change.
I check all LO that I can reach on the RF assembly, they are
a. 3.5-7.5GHz 1st L.O,
b. 100MHz Ref,
c. 94-106MHz for SAMPLER. They all seems fine and in lock. When I look closer to the spectrum line, the peak seems swinging, and a pulse like signal shown when the C.F below 3GHZ.
I wondering may be the level cal process referenced to a wrong level because some LO drifted at the sampling instant?
One LO I can not check at the beginning is the 2nd LO, which operating at 4GHz inside the 3GHz converter. I don't want to touch every single screw I see, coz experience tell me this ack may end up to a tragedy.
The missing 4GHz was found when I figuring out which screws holding the 3GHz assembly, it is moving like sweeping. This signal should be a CW signal at exactly 4GHz for 2nd mixing, moving around definitely abnormal.
The 4GHz can be check on the SMA connector shown in Signal Flow.png.
Since no schematic is available, the only way to fix is trial and error. After probing around the pcb trace, I confirm the 4GHz VCO is controlled by "N5" - SOP8 IC marked as 812. The -11V is controlling the VCO freq. on N5's pin 4. May be adjust the R22 and R40 can make it lock on again.
So I started my iteration, I adjust R22 a little bit and install it back and check the 2nd LO. I am so lucky that the signal now not swinging any more. Then I adjust more to find if there is another sweet spot and check the test point LND15 (-11V), which is accessible when the module installed back, with oscilloscope A.C couple mode to find a spot the amplitude of the "noise like" signal, which is the error voltage of VCO, is minimized.
Finally, the problem is fixed, and the cal status are all zero which means every thing is good. The minor issue is when using a very small span, 20kHz, a FM like signal is observed but actually a CW signal is feeding. May be this is the phase noise of 4GHz or other L.O, but its good enough for me and I don't want to create new problem while hunting for the minor problem.
Forgive my long article, I just finish the repair and so happy to share the information with all you guys.