Dear Sir:

I used a 30 cm airline to verify a calibration.

After shorting the airline, I saw about 1dB of peak to peak ripple.

I wanted to know if there is an algorithm that has been developed which can tell the NA user that his calibration can be improved by changing the cal kit inputs. In other words, substitute different cal kit inputs into the NA, determine the resulting ripple pattern, and make a decision as to whether changing cal kit inputs are appropriate.

Also, can you tell me the worst case phase and magnitude error I can expect with a 1 dB peak to peak ripple pattern (assume I have a perfect airline)?

Thank you for your time and I look forward to reading your reply.

I used a 30 cm airline to verify a calibration.

After shorting the airline, I saw about 1dB of peak to peak ripple.

I wanted to know if there is an algorithm that has been developed which can tell the NA user that his calibration can be improved by changing the cal kit inputs. In other words, substitute different cal kit inputs into the NA, determine the resulting ripple pattern, and make a decision as to whether changing cal kit inputs are appropriate.

Also, can you tell me the worst case phase and magnitude error I can expect with a 1 dB peak to peak ripple pattern (assume I have a perfect airline)?

Thank you for your time and I look forward to reading your reply.

If you used a load for calibration, put the calibration load on the airline and read the peak value; it will be 6 dB above the return-loss of the load (assuming perfect airline). Record this value. If it is near -24 dB, then you know the ripple is due to poor load. If not, convert the load return loss to linear, subtract from 0.057 and the remainder is due to source match.