| Products > Test Equipment |
| Siglent Network Analyzer usage |
| << < (2/2) |
| kladit:
Rob, thanks for you quick answer. I did a SOLT calibration and tryed to put in the parameters for Load in USER1, but it is still grayed out so no input possible. I think it should go the other way round, but may be I am wrong. First I buy a calibration kit. With it I get a sheet of paper with its correct data, describing its differences from an ideal calibration kit, measured by highest precision analyzers. These differences I have to put in in USER1(or2) to tell the analyzers math, which always calculates with the (not existing) ideal calibration kit the differences to it. And then do the calibration. The efforts to manufacture a calibration kit near the unreacheable perfection one and to determine the small differences to this is resposible for the enormous price differences in calibration kits. I started the exploration of my network analyzer with the attempt to measure the quality of my Miny-Ciruits HAT BNC attenuators. These are specified for DC to 2GHz, so I assumed to get a return loss of better 20dB at 2 GHz. I used a male BNC calibration kit from sdr-kits, set the anlyzer to measure S11 and did 1 port SOL calibration. Then I measured a 30 dB attenuator and got a return loss of 35 db (100kHz) and only 10 db (2 GHz). Next I used a Spinner BN533863 calibration kit at the same cable but necessarily with and additional BNC double-female and an BNC to N connector. After calibration I used the port extensions to move the calibration plane back to the end of the BNC cable (-274ps) and then measured the same 30 dB attenuator. This time I got a return loss of 35 db (100kHz) and 24 dB (2 GHz), not changing with and without 50 Ohm termination at the end oft the attenuator (as expected). May be other owners of an SVA will try to reproduce a similiar procdure and post their results, to enlighten us all about the right way to get resonable results when using these analyzers. Many users have reported to be happy to have unlocked their SSA3021X+ to an SVA but I have not read much of their experiences in using these as SVA up to now. The next procdure will be to build a simple fixuture for smd elements, measure the fixtures data and save it in the analyzer for correction of the measurement plane. This is called de-embedding. After that the real usage of the network analyzer can start. If anybody has done this already and is willing to share its results this will helpfull for all users here, the manual is not helpful in this respect, to express it friendly. Rob, please ask Siglent for a detailed description of how to add the data for a calibration kit to USER1 USER2 and why it is not possible to enter data for Load. |
| tautech:
--- Quote from: kladit on May 20, 2024, 05:00:11 pm ---......... The next procedure will be to build a simple fixture for smd elements, measure the fixtures data and save it in the analyzer for correction of the measurement plane. This is called de-embedding. After that the real usage of the network analyzer can start. If anybody has done this already and is willing to share its results this will helpful for all users here, the manual is not helpful in this respect, to express it friendly. --- End quote --- A simple fixture for TH components I cobbled together here: https://www.eevblog.com/forum/testgear/test-equipment-anonymous-(tea)-group-therapy-thread/msg4345558/#msg4345558 Inspired by Chris Grossman's work: |
| tautech:
For completeness, a pic of the 2 SMA TH fixtures I've cobbled together. They are across and through incarnations with the through with a wire shorting link fitted for Through calibration. ST42 SMD tweezers case for scale. |
| kladit:
Here again, this times I hope I will get the pictures up to eevblog. First S11 calibraition with sdr-kits cal-kits. The 50 Ohm load is 49.85 Ohm. Then measure of mini_circuits HAT30+ attenuator Smith and return-loss. Second calibraition with Spinner BN533863R000. The 50 Ohm load is 50.05 Ohm. Then measure of (the same) mini_circuits HAT30+ attenuator Smith and return-loss. The results of these simple S11 measures with different calibration kits are very different. So to make a statement about return loss is impossible with this instrument (SSA3032X-R). Or is there a failure in my approach to measure? The question is will it be possible to improve the accurracy of Siglent-VNAs by input the correct load resistance of the cal-kit in use into the definition of a USER cal-kit? With current firmware is it not possible, the input-fields are grayed out. I have read the high accuracy of an DG8SAQ-VNA(sdr-kits) or the Funkamateur VN-VA5 depends only on the acurracy of the correction data, delivered with an calibration kit. To measure and get this correction data a very high precision network analyzer is needed, unpayable for hobbyist/amateurs. But by simply transfer the correction data of an cal-kit ascertained by a high precision analyzer the accurray of this is transferred to a relative cheap analyzer with very sophisticated firmware capable to use it. Can Siglent for their network analyzers provide a firmware which will give them the same possiblilities and so accuracy as the much cheaper ones metioned above? |
| ozkarah:
Could you please first calibrate with the Spinner kit, then measure the SDR Kit components. This should give a brief info about how they differ. |
| Navigation |
| Message Index |
| Previous page |