Checking out the performance of the Siglent SSA3000X Tracking Generator with "Keypress Switchable Attenuator 0-90dB"
Hi,
As a follow-up to my last two posts, featuring the new error map compensation for SMA/NWT/D6/LTDZ devices, I decided to create the same error map with my Siglent SSA3000X.
Using the built-in TG, I connected the TG SOURCE to the INPUT of the switchable attenuator and from its OUTPUT back into the RF INPUT of the Siglent.
The attenuator costs around 60 Euro at eBay and looks like this:
Using the "VMA Simple Spectrum Analyser for Siglent SSA3000X", I recorded each trace for 0dB, 5dB, 10 dB, ..., 60dB (in 5dB steps).
Recording the first trace at 0dBm - no attenuation.
Then I united each exported trace into an error map file using the "VMA Simple Spectrum Analyser for SMA/NWT/D6/LTDZ". This is possible because all versions of my software produce compatible files when exporting traces.
Using the new Calibration tab of the version for SMA/NWT/D6/LTDZ devices, I generated the Error Map file, which basically joins all exported traces into one single map.
The result is a CSV file that can be read into Excel and which contains each trace in a separte column, with the first column showing the respective frequencies (9kHz up to 3.2GHz).
Error Map file opened directly in Excel
The TG function of the Siglent SSA3000X can use Normalization. What this does is to compensate the TG signal, so that it matches the set TG Level (in this case I set that to 0dBm).
I produced two series: one with Normalization switched off and one with Normalization switched on.
Here are the results:
Normalization switched OFF
Normalization switched ON
As a comparison, here are the equivalent results produced by the D6, using the latest version of my software:
Normalization switched OFF
Normalization switched ON
Note that "Normalization" is called "Error Map Compensation" in my software!
It amazes me how much performance can be extracted from cheap devices as the NWT and D6!
My opinions:
- SMA with only one ADF4350/ADF4350 is not recommended in this context due to lack of TG functionality. It does not allow to create an error map with the built-in signal generator and results using a cheap noise generator are poor.
- LTDZ has even worse peformance than the D6. I know the price is very tempting, but it is not very sensitive and has a noise floor of around -70dBm. You cannot use it on signals with less than -40dBm.
- The really good device is the NWT4000 by BG7TBL.
- The NWT and D6 can, in some cases, be a much cheaper alternative to a professional spectrum analyzer.
Regards,
Vitor
Hi Vitor,
ReplyDeleteas i can see you try to use kT2.5-90/1S-2S adjustable attenuator which keep parameters up to 2.5Ghz.
https://ae01.alicdn.com/kf/HTB1sZlsXizxK1RjSspjq6AS.pXaA.jpg
but you use it to correct devices which is 4.4Ghz or 6Ghz working range. From my point into your correction results you can correct not only errors of T.G. and input, but frequency related errors of attenuator too.
also it looks strange that oscillation (deviations?) which you can see on Siglent SSA3000X at 1600, 1800 Mhz also present for D6 (at least around 1600 for small attenuation and both for bigger attenuations). Are you sure that it's related to T.G. and input and not related to self-resonance of this attenuator and(or) cables you used? In case if this deviation result of calibration then real calibration curve can be worst then original for some frequencies.
ReplyDeleteHi,
DeleteThanks for your feedback.
I am aware that the attenuator is not rated above 2.5GHz, but the truth is that I don't have anything better!
Also, the comparison is meant to be "amateur/hobby" level and I wanted to show that the SMA/NWT/D6 does not perform that bad, when compared to the Siglent SSA3000X range of spectrum analysers.
The TG of the SSA3021X is not perfect, mind you! But I agree, part of the issues shown are certainly due to the attenuator, too.
Again, this is the best I can do with my available tools.
Regards,
Vitor
Vitor, Any interest to expand your software app so it can work on the Siglent SVA1032X? The spectrum analyzer part should be identical to the SSA3000x. Would you be willing to make sure source code public?
DeleteHello,
DeleteThe software is not open source and I will not release its source code.
The SVA1032X will probably work, though the supported frequency range will be certainly wrong. Also, it will for sure not support any of the VNA functionalities of this device.