AnsweredAssumed Answered

Question regarding N9020A Signal Analyser

Question asked by tomarmitage on May 7, 2013
Latest reply on May 8, 2013 by tabbott
We are having an issue with a N9020A spectrum analyser we are using in a test fixture and can't explain what we're seeing.

We have a test that performs a low-level signal measurement at 5MHz.  The test is failing and it seems to be due to a low-level peak above the noise floor in the measured spectral output at 5MHz that is not coming from the unit under test. 

Further testing seems indicate that this 5MHz peak is an artefact of the N9020A and is not due to our test set up or environment.

We've used a second, and different spectrum analyser (Agilent E4401B) to establish that this 5MHz signal is not present in the environment nor in the AC power supply.  We have eliminated other near-by equipment as the source of this 5MHz  peak by systematically turning them off one by one until only the N9020A remains.

Further, we have found that by changing the centre frequency on the N9020A to something other than 5MHz we can get the peak at 5MHz to disappear.

We can easily replicate the response by using the default settings of the N9020A and making only the following changes :

1) terminate the RF input on the N9020A with 50 Ohms
2) disconnect all other cables (except for power) from the N9020A
3) set the RF input to DC coupling
4) set centre frequency to 5.0MHz
5) set span to 100Hz
6) RBW and VBW should automatically set to 1Hz
7) set reference level to -50dBm
8) set mechanical attenuation to 40dB
9) enable trace averaging

At this point we see a noise floor at about -115dBm, and a peak at 5MHz of around -92dBm.

Turning on averaging smooths out the noise floor but does not reduce the peak at 5MHz so the peak does not appear to be random.

However if we then change the centre frequency by only a fraction away from 5MHz (eg 5MHz + 2Hz) the peak at 5MHz disappears, and other peaks appear at other nearby frequencies.

Can you tell us if this is normal behaviour for the N9020A or does this indicate that there is a problem with our unit?

Is the response we are seeing an artefact of the signal processing in the N9020A?

If so can you give us any guidance on how to avoid this?

Our N9020A is using SW version A.10.04 - will updating the software to the latest revision fix the problem?

Is this something we will just need to work around?

Attached are screen shots showing the response we are seeing at 5MHz with centre frequency of 5.0MHz +0Hz, +2Hz and +7Hz showing what I was trying to describe above.  Also attached is the saved state of the N9020A to help in replicating the problem.

We'd appreciate any insight into the cause of this problem.

Tom Armitage