Author Topic: FFT @ high precision mode with HAMEG HMO vs Rigol DS2000 scopes.  (Read 26858 times)

0 Members and 1 Guest are viewing this topic.

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: FFT @ high precision mode with HAMEG HMO vs Rigol DS2000 scopes.
« Reply #50 on: August 18, 2012, 05:36:22 pm »
Hi PA4TIM

Well let me say that my intention never have been that scope FFT can "replace" proper SA.

My intention was to find out how good or one may say how bad are FFT's at new modern scopes with HR or High Precision modes, like Hameg HMO's or Rigol's DS2000/DS4000. As not everybody have or really needs, nice SA lying around on his bench.

Simple question: What can be seen on modern HR mode scope FFT's ?

No judgement or comments if it is crap or not, just few simple screens, to show what is possible and where are the limits.

Well actually after few pages of posts, discussion, and attempts that is still open question.

Seems that guys that have the right gear for the job do not find that it is worth to use their time for few simple tests and others seems do not have much but their first scope on their bench and good will.

Anyway, so...what actually You mean by Your HMO being light Years from proper SA...??
Any screens possible..??

Kind regards
Rosendorfer
 

Offline PA4TIM

  • Super Contributor
  • ***
  • Posts: 1125
  • Country: nl
  • instruments are like rabbits, they multiply fast
    • PA4TIMs shelter for orphan measurement stuff
Re: FFT @ high precision mode with HAMEG HMO vs Rigol DS2000 scopes.
« Reply #51 on: August 18, 2012, 07:50:46 pm »
It was not my intention to make you think FFT is crap or the scopes.

I think, from your answers, you have no or not much experience with spectrum analysers.

The big difference is a scope uses the time domain and from that sampled info that makes forinstance a sinewave on your screen it uses FFT to transform this info to the frequency domain.

If we take a sinewave, the DAC meaures the whole signal, it determes amplitude and frequency from it. It is only 8 bit or so ( digital knowledge is not my thing but I have build and restored serveral SAs, digital and anlog) so the info of he harmonics, phsase noise, pm, AM, FM ect is all in the amplitude at one point at time x. All those points form a sinewave trace.

It is like standing in a bar and all people shouting at the same time, or a radio that has no tunig, you receive all stations at the same time. Bij filtering ( ears or software) we can extract one station. But tall that data is decoded by a 8 bit ADC so there is not much info in there.

A spectrum analyser can be seen as a normal receiver, you turn it across the frequencybands and you here station after station but never two at the same time. M
So if you want good details you can tune in on a station, switch to narrow filter and only hear the parts of the music that are 1KHz in tone.
The SA has a sweepgenerator that does the tuning for you. You can choose the sweepwidth, you can choose the filter width. You can look at the phasenoise from a -30 dBm 12,5 MHz carrier while that carrier fills your screen. See picture below.


Now the sinewave again. It sweeps and listens at the first harmonic, uses the full capability of the DAD only for this signal, then it sweep on and reaching the second harmonic it uses the full capability of the DAC to measure the second harmonic.

You see the difference ?  8 bit for the whole spectrum or 8 bits only for one tiny bit of spectrum.

There is a " but"  In this. Take two pulses, one at 10 MHz, one at 50 MHz. The happen exact at the same time. They last 10 nS. If they occure just within the sample time of the scope ( not in the dead time) both signals are part of the whole spectrum ( for instance 100MHz (and the attenuated signals outside the bandwidth that still are received) the scope listens too. The software filters transform the signal to frequency and you see the poles growing out of the " grass"  ( noisefloor)
But a SA that sweeps at low speed ( like some digital SAs do) are not at teo frequencys at the same time. You can make such a setup by making the pulse repeatble that the SA sweep speed such that it sees the 10 MHz but every time just misses the 50 MHz.
But this is a very extreme situation that would never happen. All SA's were very expensive precision products with very high quality filtering, ultra lownoise amps and very well engineered so bndwidt, sweeptime, filtering was optimised to be sure these things did not happen.

Since a short time there now are cheaper SA's from Rigol. It would be nice to compare one to a real SA to check the performance.


And that is what you want here. You need FFT screenshots from a known signal ( amplitude, phaseniose, harmonics ect) and then compared to the screenshots from a Agilent, Tek, HP or R&S spectrum analyser. Then you can make conclusions.
I can hook up a generator to my Hameg but the screen shot will tell you nothing. You need a reference. I have seen screenshots from the Rigol with intermodulation and strange bumps in th trace but the owner does not have enough knowledge to see if it is a measurement mistake, a bad genertor, overdriving the input amp or just a not to good SA. So the pictures are worthless.
A friend has the Hameg SA and a Tek SA. We compared those and that the Hameg performed very well. My friend is a ex-metrologist who used to work with expensive SAs for a living.

I am now on a short holyday but next week i will make some screenshot compares between FFT and SA for you.

Fred
« Last Edit: August 18, 2012, 07:55:09 pm by PA4TIM »
www.pa4tim.nl my collection measurement gear and experiments Also lots of info about network analyse
www.schneiderelectronicsrepair.nl  repair of test and calibration equipment
https://www.youtube.com/user/pa4tim my youtube channel
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: FFT @ high precision mode with HAMEG HMO vs Rigol DS2000 scopes.
« Reply #52 on: August 20, 2012, 05:12:43 pm »
Hi PA4TIM

That would be more than welcome....

I'm really looking forward to see Your tests...

Kind regards
Rosendorfer
 

Offline PA4TIM

  • Super Contributor
  • ***
  • Posts: 1125
  • Country: nl
  • instruments are like rabbits, they multiply fast
    • PA4TIMs shelter for orphan measurement stuff
Re: FFT @ high precision mode with HAMEG HMO vs Rigol DS2000 scopes.
« Reply #53 on: August 21, 2012, 03:00:48 pm »
http://www.pa4tim.nl/?p=2815 Here the results of FFT tests in normal and hires mode compared to a HP 5885B SA.


Look at the amplitude and low noise signal, no averaging, tricks or filters, just pushed the autobutton and used the 50 ohm input.-40dBm at 10MHz.


No difference between Hires and normal.
(see site)

Normal mode, hanning, 32Kpts

Hires mode, hanning, 32Kpts
But, who to trust ? I noticed this later and the HP could not show all harmonics at the same time. So FFT on the hameg has a plus in this field (but see the comment on my site, FFT can in most cases not replace a SA)

A-10dBm signal so the SA can show all at once

Normal mode, 32Kpts, hanning
Harmonics:  -70.26, -60, -84, -71, -78, -89, -85
deviation:             0,    +9    -8    +5   +2  -7    -1


HiRes mode, 32Kpts, hanning
Harmonics: -66.4,  -60,  -75, – 75, -82, -81, -78
deviation:      +3.6   +9     +1    +1    -2    +1   +6
Looks better at the higher harmonics


No difference between modes

Looks the same but instead of around 4MHz/div it is here 5KHz/div. The SA rules  ;)

The Agilent trick, (but they used larger amplitude difference, but could not get both signals synced because the pulser had no inputsync.
 
Analog scopes can do this too. (and a lot more easy to trigger)
www.pa4tim.nl my collection measurement gear and experiments Also lots of info about network analyse
www.schneiderelectronicsrepair.nl  repair of test and calibration equipment
https://www.youtube.com/user/pa4tim my youtube channel
 

Offline Rosendorfer

  • Regular Contributor
  • *
  • Posts: 50
Re: FFT @ high precision mode with HAMEG HMO vs Rigol DS2000 scopes.
« Reply #54 on: August 22, 2012, 01:08:13 pm »
Dear PA4TIM

What can I say, it is absolutely great work You have been kind to share with us...

And second thing that comes to me is that Your Hameg HMO3522 is /just as R&S scope supposed to be /  "Scope of the Art" ... Measurements capability and resolving power of that machine is just amazing.
For myself having Rigol1052E that what Your Hameg can do and clearly show at FFT's is almost "wonders"....

I have to reconsider few things as seems, that what HMO is showing at "normal" FFT is, as far as I know, "bit more" that should in theory be possible in "normal" 8 bit mode.
Well clearly  that what Hameg can show even at "Normal" mode is way(!) above that what small Rigol can possibly show.

Any way to try find out, if whenever (!) FFT is used, HMO turns some kind of special/extra FFT's HR mode.
That possibly could "explain" extremely high resolving power of HMO FFT's at any "mode". And olso could bit explain "lack of expected" major differences between normal and HR FFT mode.
But just a fought... But well..as for me HMO FFT's at any mode is HR..!!!

Thanks again for taking Your time to share Your measurements with us.....

Rosendorfer



 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf