Siglent SSA3000X Plus Harmonic Measurement

[eloso]

Hi

Is there anyone who has got their head around the automatic harmonic measurement function of the SSA3021X Plus ?  There is something I can't figure out and the manual/google is not getting me to where I want to be.

I have used the Spectrum Analyser with measurement modes switched off to look at the fundamental of a carrier at 10.115mhz. Looking at its second harmonic, I get a result, using a delta marker, of the second harmonic being -55db down on the fundamental. So far so good.

But the Siglent seems to have a super-duper auto harmonic measuring fuunction built in. If I select that function (off the "Meas" burron)  I get a number of options down the right hand of the screen - Fundamental Auto/Manual, Freq Step Auto/Man, Harmonic Num etc etc.   

1st problem. Whenever I select the harmonic measurement function the frequency span goes barmy and jumps from 10Mhz-40Mhz up to over 1 Ghz. This is in auto mode. I cannot get auto mode to start a measurement at a fundamental of the signal I am presenting to the machine i.e. at 10Mhz. I have to select Manual and key in the frequency. The auto seems to work for frequency step and in fact everything seems to fall into place and I get a set of readings for 2nd through to 10th Harmonic once I have manually keyed in the frequency of the fundamental. But I should have the machine tell me the frequency of the fundamental and not have to key it in ?

2nd problem, The second harmonic is given as around -65dbBc. That result is not consistent with my having looked at the signal in Spectrum Analyser mode where it was -55dB. 

I am obviously missing a very big trick with the so-called auto harmonic measurement function. Can anyone help me to understand what trick it is that I am missing ?




Thanks !

Eloso

Modified with added  screenshots and numbers adjusted to match screenshots

[rf-loop]

Hi

Is there anyone who has got their head around the automatic harmonic measurement function of the SSA3021X Plus ?  There is something I can't figure out and the manual/google is not getting me to where I want to be.

I have used the Spectrum Analyser with measurement modes switched off to look at the fundamental of a carrier at 10.115mhz. Looking at its second harmonic, I get a result, using a delta marker, of the second harmonic being -55db down on the fundamental. So far so good.

But the Siglent seems to have a super-duper auto harmonic measuring fuunction built in. If I select that function (off the "Meas" burron)  I get a number of options down the right hand of the screen - Fundamental Auto/Manual, Freq Step Auto/Man, Harmonic Num etc etc.   

1st problem. Whenever I select the harmonic measurement function the frequency span goes barmy and jumps from 10Mhz-40Mhz up to over 1 Ghz. This is in auto mode. I cannot get auto mode to start a measurement at a fundamental of the signal I am presenting to the machine i.e. at 10Mhz. I have to select Manual and key in the frequency. The auto seems to work for frequency step and in fact everything seems to fall into place and I get a set of readings for 2nd through to 10th Harmonic once I have manually keyed in the frequency of the fundamental. But I should have the machine tell me the frequency of the fundamental and not have to key it in ?

2nd problem, The second harmonic is given as around -65dbBc. That result is not consistent with my having looked at the signal in Spectrum Analyser mode where it was -55dB. 

I am obviously missing a very big trick with the so-called auto harmonic measurement function. Can anyone help me to understand what trick it is that I am missing ?

(Attachment Link)
(Attachment Link)

Thanks !

Eloso

Modified with added  screenshots and numbers adjusted to match screenshots

There is also other things what need comment but at this time I will comment only what are bolded by me. (at this time I do not have this model for test so better to leave these other things to others) But..

First you tell "Looking at its second harmonic, I get a result, using a delta marker, of the second harmonic being  -55db down on the fundamental."

But in image you have measured 3. harmonic, not second harmonic (aka 2th or 2.)!  Fundamental is signal first harmonic, here, there and everywhere.


If look list of harmonics in "attach 2"  there is clearly fundamental what is 1. harmonic. Then 2. -65dBc (looks more like noise*) and 3. harmonic  -55dBc.
* I do not know what parameters (example RBW etc etc ) this automatic harmonics measurement use.

In your image "attach 1"  you have measured 1. harmonic ( = fundamental) and then 3. harmonic just same -55dBc. (2. harmonic not visible)

At this time I do not have this model for test anything. But these harmonics in your images is just right.

[eloso]

Thanks - at least the main part of the query is now cleared up.   On reading your comments I was about to disagree (doh!) when I realised that when I used the Delta Marker function for the difference between Marker 2 and Marker 1 it reads the difference not just for the Y axis but for the X axis too. So I read 20.22998Mhz and took it at face value, not realising that really it is the Delta on the x axis too  - i.e. the Marker 1 frequency of 10.115Mhz + 20.22998 Mhz which is of course the third harmonic.  So both forms of measurement work out the same.

Well spotted and thank you very much.

I still don't know why the auto function doesn't seem to do anything - for a measurement mode that is meant ot be fairly rapid and automatic it seems wrong to have to type in the frequency, but that is an annoyance, whereas when I thought I was getting bad readings , well that would have been pretty awful.

Regards and thanks again

Eloso

[hendorog]

The Auto function simply searches for the peak of the fundamental.
You can type in an approx fundamental freq, and within some limit it will find the exact peak.
Makes sense as the dBc harmonic measurements rely on the peak amplitude being known accurately.

I did a test with a 10MHz fundamental.
Turn on Harmonic measurement and set freq to 10MHz.
All good, harmonics shown and freq detected as 9.9999MHz
Then I changed the fundamental to 9.95 MHz on the Sig Gen.
Change Fundamental setting on SSA from Auto to Manual and back to Auto.
Now the SSA detects that the 9.95 MHz tone is the fundamental automatically.
No manual change of freq was done on the SSA.

I had mixed success with larger freq changes. 10MHz - 9MHz didn't work, but 10MHz - 11MHz did for example.

[eloso]

Hi,

Thanks for taking the time to do that test. I am getting different results and have tried this several ways that are slighly different too.

In a nutshell, pressing the "Auto" softkey on the menu will always make the thing jump to a fundamental of over 1Ghz. No matter how close the manual setting was or wherever in the sequence of measuring operations I do this.

I have tried many combinations including fully autouning the analyser with the blue autotune button, then turning on harmonic measurement, then setting frequency manual on the panel frequency button, different spans, doing this with the harmonic measurement already on, and already off and whatever combination I try it will not behave anything differently. I have set it to frequencies very close to the actual fundamental i,e, withing a few khz and never once has it jumped to, nor identified, the precise fundamental.

I am guessing that either there is some other parameter either deliberately or erroneously linked to this measurement that is set differently on your machine, or when I installed the SVA firmware as I did quite some time ago, it has broken something.  My gut feel is that I haven't broken anything with the cross flash and there is some bug/quirk that means I am overlookinging some setting somewhere that has an impact, or there is some secret, precise sequence of button pressing that is eluding me.

In one sense it isn't exactly a showstopper but having to note the precise fundemental frequency on one screen and then enter it on another - well heck it sure is  annoying me !! 

Thanks again,

Eloso

[hendorog]

One thing i did which i didnt mention was to set the reference level within a few dB of the fundamental amplitude.
I will have another play to see if i can narrow down when it works and when it doesnt.

[eloso]

Thanks,

I have tried it with varying reference levels around the same level as the carrier. No good. 

I tried it with manually setting a frequency well below, close to, above and well above the actual fundamental - no joy.

I fiddled with the resolution bandwidth in the hope that this had not been taken into account by Siglent but still no change.

Nope, still head scratching. I am guessing these additional measurement modes are typically about volume production testing rather than quick point and click test rig setups.  Regardless, I'd still love to see this "Auto/Manual Fundamental" working.

Cheers

Eloso

[eloso]

OK, think I have it sorted

Since following your suggestions I have been trying this whole process on a fundamental signal generated at 7Mhz and as reported, I got no-where with it. I didn't think it was too relevant that I had changed to 7Mhz from 10Mhz.

Actually it is very relevant. The auto function does not work on 7Mhz but it works on frequencies above 10Mhz.  When I changed back to 10Mhz it worked.

So to do an auto harmonic measurement :

1. Set up amplitude parameters to suit the signal.  Whether it is needed or not I ensure that the ref level is somewhere close to the fundamental peak.
2. RBW set to auto (doesn't seem strictly necessary but does affect speed of measurement)
3. Select "Meas" and navigate to "Harmonic"
4. Tap "Fundamental" and manually enter a frequency known to be below that of the fundamental. How far below doesn't seem to matter. A frequency of 1Mhz always seems to work for me. After pressing the Mhz soft button the "fundamental" dialogue box disappears.
5. Tap "Fundamental a second time. Now press Auto.  A scan is made and the analyser sets the frequency of the fundamental accurately and displays the wanted harmonic measurements.  It doesn't really matter, but the dialogue box for "Funamental" remains open, obscuring the trace, but the trace is only showing a notional repesentation of the signal which is why this is unimportant. Just a bit ugly.

If you change frequency, you need to enter manually a frequency again that is below the new fundamental.  Again 1Mhz seems to do fine.  Then select Auto again and it works as before.

Its not the way I would implment a UI - not by a million miles.  A single button marked "Auto" could be pressed and it will seek a new fundamental based on the signal present at the time of pressing Auto.  Or use the "Restart" to make  a new measurement including the first scan and identifying the new fundamental.   Anything other than the half-baked way it works at the moment 

Its a lovely machine so lets keep this in perspective but I would close by saying this aspect of the UI really should be classed as a bug requiring a fix.  I also wonder if there is a good reason that the measurement is limited to signals of 10Mhz and above.  Us radio amateurs have some fairly low frequency needs.  I have a feeling that the predecessor of this SA only went down to 10Mhz ? Maybe someone can confirm.  If so then it seems likely that this bug originates from  code for the earlier model that sets the pre-scan to start at 10Mhz - something that should have been changed but was overlooked when the frequency range of the scope was improved.

Kind Regards

Eloso

[hendorog]

Awesome, thanks that is something I can submit as an improvement and bug fix.

I wonder if there should just be a peak track on off option, same as a marker?
Also i thought it would be nice to be able to jump back to SA mode to view a selected harmonic.