AnsweredAssumed Answered

E4404B align problem

Question asked by vents on Dec 2, 2013
Latest reply on Oct 18, 2015 by aleksandr

I recently got an E4404B spectrum analyzer. According to "ESA SA Getting Started Guide" I started with running internal alignments - I connected AMPTD REF OUT to INPUT and after some warmup time set it to "Align Now, All".

It finished aligning with no errors, but when I turned Amptd Ref out ON to view the signal, it didn't show -20dBm @ 50 MHz, but -28dBm instead. I tried to connect external generator (also -20dBm @ 50 MHz) and it still showed -28 dBm. Then while looking through control options I found an option to change coupling mode. When I set it to DC,  it actually showed correctly ~ -20dBm, but when changed back to AC -> -28dBm. Then I left it like that and after a while I noticed it was showing +10 dBm instead. Then I tried again to change it to DC coupling and it showed +5dBm. When changed back to AC coupling = -14 dBm. Then I set it to align again and after it was done it was jumping between ~ -10dBm and ~+10dBm while AC coupled and reading constantly ~ -19dBm while DC coupled (still with the same -20dBm 50MHz ref signal).

I have tried to run aligned a few more times and got different results again, for example, one time after alignment it showed -30 dBm while AC coupled and +5dBm while DC coupled. After a minute or so it jumped to -16dBm (AC) and -20dBm (DC). 
Also, when I try to run Align Now, RF it sometimes says that no reference signal was found, but sometimes completes the procedure successfully (but again I get different results, e.g. 3dBm (AC coupling) and -1dBm (DC coupling)).

I have never before worked with a spectrum analyzer and generally I'm pretty new to RF stuff, so any suggestions/ideas will be appreciated.

Thanks in advance

P.S. After reading some threads on this forum it seems that a mixer could be damaged. I read that in the case of a damaged mixer the noise floor would be high. When I set the frequency span to full range, indeed the noise floor in 0 - 3GHz range is ~ -30dBm compared to ~ -55 in >3GHz range. If that is the case, could someone please provide some information about repair procedure?

Edited by: Vents on Dec 2, 2013 3:29 AM