AnsweredAssumed Answered

PNA-X and IMD issues

Question asked by mirek on Dec 10, 2012
All comment apply to Firmware  A.09.80.03
Issues with IMD class in frequency segmented sweep
1.     Problematic IFBW setting.
The system use IFBW for IM measurement from “Sweep Settings IMD tone IFBW”. The problem with that is that the operator does not have any control of IFBW over measured frequency in segment sweep.  In production environment you try to optimize the IFBW to the measured IMD value so you will set segmented sweep with different IFBW as necessary to improve the speed of measurement or improve the dynamic range.  That is not passible right now. 
I will suggest disabling the IFBW settings from Frequency Sweep Settings form (live the main tone IFBW settings as is) when segment sweep is enable. 
2.     Incorrect power settings IMD during calibration.
The system performs 3 powers leveling:
-     Arbitrary power leveling (0 dBm)
-     F1 Source 1 leveling
-     F2 Source 2 leveling
   During IMD calibration, a specifically source leveling, the system relay on power setting from F1 (couple) or F1 & F2 (uncouple). In Vector Calibration it uses different power. It use setting forom power from segment sweep. I have problem with that.
First of all the power for leveling of F1 & F2 should be arbitrary (for example 0dBm). The reason is that power for DUT during the test might be very different than the calibrated power, for example with IMD that is measure with fix output power like amplifier.  The gain of the measured amplifier will change that power. Let say you like to perform measurement at +35dBm output then there is no way to level sources at that power.
Power in vector calibration.
One the segment sweep is enable vector calibration is using power set in segment sweep, however it is not calibrated level so the system do not know if something change in input configuration.
That a problem as I experience that I experience that in my setup.
I discover that because in my case I set 5dBm in segment sweep and my setup contain on Source 1 13dB attenuator that the system could not see.  In such case it will be better if the system use the leveled arbitrary power (0dbm). 
The system should stick to the convention that is implemented in other classes: if segmented sweep is enable use IFBW and power from that segmented sweep for measurements.  
Leveling and Calibration should be perform with some arbitrary values (specially power settings) 
I will suggest to re thinking the calibration methodology for IMD in the following way:
-     Switch off the internal combiner
-     Level  Source 1 at arbitrary power (0dBm) with power meter for all frequencies (F1, F2, IM3 …) with calibration of receiver R1 in the same time 
-     Perform Vector calibration at arbitrary power (0dBm)  
-     Switch On Internal combiner 
-     Enable Rx  leveling for R1 receiver (for input IMD) or B (for  output IMD)   
You do not need to level both sources with power meter. One source is sufficient. Any way if you perform output IMD all source leveling is invalid as B receiver RX leveling takes over.  With fix input you just enable R1 Rx leveling.
Such approach gives you fare more robust, faster and simpler calibration process.