Calibration

[MVK]

Hello andy,

In case of RF signal generation how to add correction factor and where can add?

I am using RFSG NI PXI 5670 to generate RF Signals,when generated signal  is –60dBm using NI RFSG 5670, it is measured using standalone agilent instrument (which has been calibrated recently) ,but measured value is around –63.5dBm Difference is around 3.5dB. I want to adjust my RFSG so that I should able to generate –60dBm signal.since this is not a linear scale, dB is logarithmic scale right… so what I have to do? why the difference,is the difference will remain constant for any value generated or else keep changing.

[Andy_Hinde]

Hello MVK,
When using the NI-RFSG driver with the PXI-5670 RF Vector Signal Generator, you don't need to add any extra function calls to get best possible amplitude accuracy performance. The NI-RFSG driver handles all of this automatically. If it has been more than one year since you had your PXI-5670 calibrated, then you should return the unit to National Instruments for a recalibration. Also, this measurement is dependent on the amplitude accuracy performance of the Agilent instrument which you used, as well as that of the 5670. The 5670 has an amplitude accuracy specification of +/- 0.8 dB for signals between 10 MHz and 2.7 GHz in frequency and -30 to -80 dBm. This would have to be combined with the accuracy of the Agilent instrument.

Thanks,
Andy

[MVK]

Dear andy,

What you are telling is absolutely correct andy. Eventhen I am having some cable loss. I mean what ever I have generated is passing over 1)SMA to SMA cable 2) SMA to N-type adopter 3) N-type to N-type cable 4) attenuator and 5) N-type to N-type cable then it reaches the UUT(unit under test). So all over around 3.5dB signal loss is there. That is what I have mentioned in the first message. I want to compensate this loss. As per your statement from “The 5670 has an amplitude accuracy specification of +/- 0.8 dB for signals between 10 MHz and 2.7 GHz” even I want to compensate this 0.8db also. Because my application is very complicate. 0.1dB to 0.3dB will me acceptable range. So what should I do in this situation?

[JSflow]

MVK,

Think about it. If you know there are cable losses from various paths and you'll be using same cabling and connectors

throughout the whole measurement, you can just simply add whatever value had been lost due to the cable loss. Or, if

it's been more than a year, send it back to NIC and get the calibration done.  Also, for the inaccuracy between Agilents and

ours, measure those out and also add the off value to it.

[Andy_Hinde]

Hello MVK,
Typically, this is accomplished by measuring the cable loss X and (a) generating a signal that is X higher than what you actually need or (b) adding X back to the acquired measurement. If you have 3.5 dB of signal loss, than add 3.5 dB back to the acquired measurement, or generate a signal 3.5 dB higher than what you are currently requesting.

Thanks,
Andy