Comparison Rigol DSA815-TG with Siglent SSA3021x spectrum analyzer

[rf-loop]

I expected it to be worst than that by at least 10 dB due to the Phase Noise of the TG PLL in the Siglent SA at 1 kHz offset.  That is why I still wonder if Averaging was used in the measurement, and if not, where did the 'A AVG 20' on the right hand of the Display come from in the Siglent picture.

I'm interested about why you think averaging is bad? Can you explain how it affect result reliability?

Because  sometimes I ref too often  HP/Agilent papers, here is R&S about measuring pn figure using spectrum analyzer.

http://www.rohde-schwarz.at/file_6189/1epan16e.pdf

Step 3:
Activate signal averaging:

Set video filter to noise weighting:

For noise averaging, FSE sets the video
bandwidth to 1/10 of the resolution
bandwidth

Activate trace averaging:

FSE performs a sliding average over
10 sweeps to obtain stable results. The
sample detector is activated at the same
time for correct noise weighting.

Somehow I think Rohde and Schwarz do not try fool readers, I think they know what they are talking.
 

[pascal_sweden]

What about the noise floor level on the more high-end Rigol spectrum analyzers, such as DSA832-TG, DSA875-TG, or DSA1030A-TG3?

Are these noise floor levels all the same as in the base model Rigol spectrum analyzer, DSA815-TG?

Bringing this question up again.

How does the noise floor level in the higher end models of the Rigol DSA800 series match up with the Siglent SSA3000 series?

[D3f1ant]

And how do the 6ghz variants compare with Keysight or R&S machines?

[videobruce]

Attached are two Spectrum Analyzer comparison charts I have compiled over the past three or so years. Updated several times, the 1st is in price order and includes the three new Rigol & Siglent models. The specs were taken from the manufactures spec sheets and/or user manuals. Both attachments will print out on one page each (I split theses purposely).
The problem doing this are the inconsistent ways manufactures spec these. DNAL is the most troublesome since there are more than one way to make it look better. Same goes for phase noise, though not as bad. Those two are probably the most important specs from my research over the years.

If you see errors please report and name the source.

Regarding the specs, Rigol states this:

Typical value and nominal value are defined as follows.
      
Typical value:  defined as the specifications when the product is under specified conditions.
Nominal value:  defined as the approximate quantity in the application of the product.

Siglent states this:

Typical: 80 percent of the measurement result will meet at room temperate (approximately 25 ? ). It has 95th percentile reliability. This date is not warranted and does not include the measurement uncertainly.
Nominal: The expected mean or average performance or a designed attribute such as the 50 ? connecter. This date is not warranted and does not include the
measurement uncertainly.

Both are quotes including any typos.

[tautech]

Attached are two Spectrum Analyzer comparison charts I have compiled over the past three or so years. Updated several times, the 1st is in price order and includes the three new Rigol & Siglent models. The specs were taken from the manufactures spec sheets and/or user manuals. Both attachments will print out on one page each (I split theses purposely).

The problem doing this are the inconsistent ways manufactures spec these. DNAL is the most troublesome since there are more than one way to make it look better. Same goes for phase noise, though not as bad. Those two are probably the most important specs from my research over the years.

If you see errors please report and name the source.

Thanks, the last basic one I got was posted here:
https://www.eevblog.com/forum/testgear/siglent-ssa3000x-spectrum-analyzers/msg827624/#msg827624

[pascal_sweden]

I read in another forum post that Rigol has further improved the noise floor level in the DSA815-TG model.

So if you have a newer revision, you most likely have a lower noise floor level, and it is now coming closer to the Siglent one. Still of course the screen resolution is lower, and there is still a delta in the noise floor. Moreover the DSA815-TG can not get 3 GHz super powers

[ErikTheNorwegian]

Your chart..
The Rigol DSA 1030A was presented in august 2009 .
Ref http://assets.tequipment.net/assets/1/26/Full_Line_Catalog1.pdf
See under Rigol milestones..

I bought a DSA 1030A TG3 for 1415 Euro just tree weeks ago from RIGOL Clearenance.
Very happy about it.

Erik

[tautech]

As I get time I'll update with some of the specs that are missing or contradictory with the 2 Siglent SSA3000X models.

My SSA3032X boot time is ~20 seconds.
When I unbox a new SSA3021X next I'll confirm boot times for it too.

[blackfin76]

I bought a DSA815-TG in 2013 for €1300,- and unlocked all the features. Yes it's a 'noisy' analyzer but that's not a real issue for most uses. You can actually do some EMC pre-compliance measurements with it (unlocked QP detectors ) and if you hook up a directional coupler you have a nice antenna analyzer as well (unlocked VSWR). For looking at the smallest of signals a pre-amplifier combined with filters may be used. The 10Hz resolution bandwidth is even very useful for frequency measurement, the build in frequence reference is almost spot on. I'm still very happy with my purchase, thanks Rigol for supplying such cool equipment for such a low price 

[videobruce]

the Siglent only allows to save BMP, there is no JPG or PNG option (yet?).

The manual stated all three are available, but initially only .bmp & .jpg were available. After talking with Siglent yesterday, the next firmware version will add .png.
Why the Chinese seem to think .bmp is a viable option is beyond me. That and VGA are grossly outdated and neither have any place in todays world AFAIC. .jpg would be ok if they didn't use such a high compression rate, or at least allow some choices in compression levels.

[bozidarms]

There is one  interesting video from National Instruments:
- How to Measure the Noise Floor of Your Signal Analyzer -
         ,
Presentation is performed on extremely pricey laboratory grade instrument.
One can repeat every single step on Siglent SSA analyzer with astonishing precision and similarity.
Only slightly difference is in change of noise flor with lowering of reference level (but this depends of manufacturer philosophy how to build the instrument) and sometimes in speed.
At the end, more or less  the same result(within a few dB) is achieved with Siglent SSA analyzer.
Finally, we have become a competitive and affordable instrument, almost without shortcomings (if any)
- a real game changer!

[pascal_sweden]

That's what's so nice about Siglent and Rigol. Their pricing is fair!
Finally students and enthusiasts can have stellar-grade equipment at an affordable cost.
We all can become skilled RF engineers from our home or garage

[videobruce]

I bought a DSA 1030A TG3 for 1415 Euro just tree weeks ago from RIGOL Clearenance.

Great price. I don't understand why they continue to seel the 1030 & the 1030A other than they have a battery option.
I had a chance to use the 1030a two years ago and found the performance similar to the 815. The older 1030 was worse.

[videobruce]

Attached are screen shots of the respective DANL specs from each manufacture (including the 832E). With Rigol, it's pretty straight forward. I used the typical figure since it's close enough to the other spec. For Rigol's Phase Noise, that was on another page and it was -80dBc for the 815 and -90dBc for the 832E.

With Siglent it's another story! I used the 1st spec (normal I assume) for the table in my previous post. I added a red outline around figures for 2.1 & 3.2GHz.
The DANL spread on Siglent's chart varies from -97dBm to -161dBm depending on what criteria is used. If you saw the u-tube video from Siglent, the stated DANL was -161dBm (using the inflated 1Hz criteria).

DANL long story short: the 3021 blows the 815 out of the water,
BUT,
the 832E blows the 3031X out of the water at 3GHz!

I updated the charts in my previous post.

[ErikTheNorwegian]

I  bought this bok from Amazone..very good read and info about RF measurements.

Communications System Laboratory offers an integrated approach to communications system teaching. Inspired by his students’ expressed desire to read background theory explained in simple terms and to obtain practical computer training, Dr. Kumar has crafted this textbook, ideal for a first course in communication systems.

The book merges theory with practical software and hardware applications. Each chapter includes the following components: a brief theory that describes the underlying mathematics and principles, a problem-solving section with a set of typical problems, a computer laboratory with programming examples and exercises in MATLAB® and Simulink®, and finally, in applicable chapters, a hardware laboratory with exercises using test and measurement equipment.

Covering fundamental topics such as frequency and bandwidth, as well as different generations of modulation including current 4G long-term evolution (LTE) techniques and future technologies like ultra wideband (UWB) systems, Communications System Laboratory provides engineering students with a deeper understanding of how electronic communications link the world.

https://www.amazon.co.uk/gp/product/1482245442/ref=oh_aui_detailpage_o03_s00?ie=UTF8&psc=1

[ErikTheNorwegian]

I bought a DSA 1030A TG3 for 1415 Euro just tree weeks ago from RIGOL Clearenance.

Great price. I don't understand why they continue to seel the 1030 & the 1030A other than they have a battery option.
I had a chance to use the 1030a two years ago and found the performance similar to the 815. The older 1030 was worse.

Yes, cant argue the price.
Im only now, a hobby tech, so its just to have it to explore more of the RF side of electronics since ive have not worked with that in my working days. So a great instruments to a great price.

[tautech]

As I get time I'll update with some of the specs that are missing or contradictory with the 2 Siglent SSA3000X models.

My SSA3032X boot time is ~20 seconds.
When I unbox a new SSA3021X next I'll confirm boot times for it too.

Almost the same, call both ~22 seconds.

[rf-loop]

Attached are screen shots of the respective DANL specs from each manufacture (including the 832E). With Rigol, it's pretty straight forward. I used the typical figure since it's close enough to the other spec. For Rigol's Phase Noise, that was on another page and it was -80dBc for the 815 and -90dBc for the 832E.

With Siglent it's another story! I used the 1st spec (normal I assume) for the table in my previous post. I added a red outline around figures for 2.1 & 3.2GHz.
The DANL spread on Siglent's chart varies from -97dBm to -161dBm depending on what criteria is used. If you saw the u-tube video from Siglent, the stated DANL was -161dBm (using the inflated 1Hz criteria).

DANL long story short: the 3021 blows the 815 out of the water,
BUT,
the 832E blows the 3031X out of the water at 3GHz!

I updated the charts in my previous post.

Attached is screen shot of the real Displayed RBW 10Hz Average Noise Level (DANL 10Hz) from Spectrum real screen, not from specifications sheets. Just naked truth as it is IRL.

[nctnico]

But that is your unit! When selecting equipment you need to look at the guaranteed specs, not what a single unit does because a different unit can be much worse. Also temperature can have a huge influence. The specifications take that into account as well.

[rf-loop]

But that is your unit! When selecting equipment you need to look at the guaranteed specs, not what a single unit does because a different unit can be much worse. Also temperature can have a huge influence. The specifications take that into account as well.

Of course, is it unclear to someone? 

If I show my random single unit screen shot do you think someone can think this image is from 1000 equipment random sample group without note that image is fom single unit.
Environment temperature +23 celsius. SSA power on 2 hours before this. Input terminated using Narda 50ohm N.

Perhaps some other people can do same so we get more variations. Perhaps my unit is most bad? Who knows.

[g8fek]

Place an external VLNA (Very Low Noise Amplifier) infront of a spectrum analyser and reduce the DANL <-170dBm/Hz.

Amongst other features, the SAT1 Spectrum Analyser ToolKit at g8fek.com provides -170dBm/Hz DANL 30MHz ... 2GHz from almost any spectrum analyser.

Remove the 50 ohm load from the VLNA input (introducing ambient noise temperature power, as specified in standardised DANL testing) and you can see something approaching the magic -174dBm/Hz !!

Lots of useful references and links on the web site.

Hope that's useful.