osy January 28, 2021 at 1:53 am
- This topic was modified 1 year, 5 months ago by osy.
I’m attaching 3 screen shots while implementing Scalar Network Analysis.
1. Before calibration, I see fluctuations with period of 200kHz.
Any specific reason for this?
2. After Calibration I see flat line. Is the calibration set some specific reference value to each 200KHz band? What is the mechanism of calibration?
3. When I test with 10uF capacitor splitting 3 sweep ~30KHz, 30KHz~250KHz, 250KHz~, I always see level discontinuity at 30kHz and 250KHz of 2~3dB. This causes concerns on measurement result from our customer. Can I get resolution to remove discontinuity and the origin of the discontinuity? Is it from calibration characteristic? I’m really in need to resolve this ASAP.
Attachments:You must be logged in to view attached files.
Justin CrooksModeratorJustin Crooks January 28, 2021 at 10:07 am
This is likely due to impedance steps in the TG and SA. Although our systems are 50 ohms nominal, the SWR is not great in spots. This can be largely mitigated with a 6 dB fixed SMA attenuator on the TG and SA. Their attenuation will calibrate out, and should make any discontinuities all but disappear as your return loss will improve by 12 dB.
osyParticipantosy January 28, 2021 at 3:30 pm
Thansk for your comment.
And can you give me any valuable option try in API or configurtion/setting to smoothnout? Even 6dB attenuator shows similar result even though it align a little bit.
(image rejection, reference level,…)
End customer push me to investiagte any resolutuion withoout sacrificing dymanic range.
Justin CrooksModeratorJustin Crooks January 29, 2021 at 5:17 pm
Did you put a 6 dB attenuator on both the TG and the SA? It should align fairly well.
If you want more dynamic range, you could amplify the TG output. A good amplifier that covers your frequency range would eliminate the effect of SWR on the TG side. A gain of 10-20 dB, followed by the device under test, followed by a 10-20 dB attenuator to the SA should preserve dynamic range and give you excellent return loss on both the input and output for your device under test.
You must be logged in to reply to this topic.