Poll

Which one is better for my application? (measuring 2nd and 3rd Harmonic distortions of a sensor from 10KHz to 100MHz)

Siglent SSA3021X ($1200, produced in 2018)
1 (33.3%)
HP 8560E (good condition but without calibration)
2 (66.7%)

Total Members Voted: 3

Author Topic: A spectrum analyzer for measuring linearity  (Read 2564 times)

0 Members and 1 Guest are viewing this topic.

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
A spectrum analyzer for measuring linearity
« on: March 26, 2021, 02:30:25 pm »
Hello
I have designed a new sensor in 2018 so I need to have a spectrum analyzer to measure the harmonics and noise of the sensor. Unfortunately I haven't accessed to a Spectrum Analyzer yet!

I analyzed all circuits in theory and then simulated and built the sensors. In simulation the SFDR is from 70dBc to 90dBc. the results are very good for my application. For example if the input signal is 0dB the second harmonic distortion is 80dB or 85dB or something similar (it depends on the frequency). One of my sensors covers from 10KHz to 50MHz but another sensor can cover from 10KHz to 100MHz and all the outputs are 50 Ohm, 1Vp-p and the SNR is from 75 to <87 (it depends on the bandwidth). now I want to measure the HD2, HD3, THD, AC response (magnitude of the bode plot) of my sensors.

I apply 10KHz to my sensor and check the output response to measure HD2 (second harmonic distortion), HD3 (third harmonic distortion), THD then I will increase the input frequency upto 100MHz step by step (for example 100KHz, 200KHz, 300KHz,...., 1MHz,...) to plot HD2, HD3, THD over frequency axis, another test is only to plot magnitude of the bode plot of the sensors, this is all of I want :) finally I will publish the results in a journal.

I see most of good spectrum analyzers have Amplitude accuracy better than 1dB and it's very good for me, 90dBc/Hz of phase noise is very good, which spectrum is good for me? Old HP8560a, HP856x, E440x series or etc?
Plus I've found HP3588a dynamic signal analyzer as an interesting choice but it's boatanchor!!:
https://accusrc.com/uploads/datasheets/4629_3588A.pdf

Thanks
« Last Edit: March 29, 2021, 08:00:59 pm by xzswq21 »
❤ ❤
 

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 6126
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #1 on: March 26, 2021, 09:22:10 pm »

What is your budget for this?
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #2 on: March 26, 2021, 09:25:36 pm »
I have taken some screenshots from the datasheets, please help me to choose a proper spectrum analyzer for my application, (you may know better models):

For HP8560E:




For HP8560a:




For HP3588a Dynamic signal analyzer:


plus HP3588a has "low distortion mode"
https://accusrc.com/uploads/datasheets/4629_3588A.pdf
The phase noise of HP3588a is only -105dBc/Hz @1KHz

For Siglent SSA3000 series:


Phase Noise Performance of HP Spectrum analyzers:
« Last Edit: March 28, 2021, 08:40:59 am by xzswq21 »
❤ ❤
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #3 on: March 27, 2021, 08:17:21 pm »

What is your budget for this?
Do you have any solutions?! :)
❤ ❤
 

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 6126
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #4 on: March 27, 2021, 10:19:31 pm »

If money is no object, www.keysight.com is your friend!  :D
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 8996
  • Country: us
  • Retired, now restoring antique test equipment
Re: A spectrum analyzer for measuring linearity
« Reply #5 on: March 27, 2021, 11:36:16 pm »
Will your sensor drive 50 ohms?  Otherwise, you need to characterize the distortion of a voltage probe or preamp as well as the analyzer.
 
The following users thanked this post: xzswq21

Offline xmo

  • Regular Contributor
  • *
  • Posts: 193
  • Country: us
Re: A spectrum analyzer for measuring linearity
« Reply #6 on: March 28, 2021, 04:10:42 am »
Since you want to measure individual harmonic levels and also THD, an instrument with the THD function would seem preferable to a basic instrument that lets you do the hard work.

I suggest you choose several candidate instruments.  Then download the brochures, specifications sheets, and operating manuals for each and compare them vs. price of available units.

A couple instruments that can do THD are the Agilent ESA-E and the R&S FSP.

 
The following users thanked this post: xzswq21

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #7 on: March 28, 2021, 06:46:32 am »

If money is no object, www.keysight.com is your friend!  :D

 :D :D if I had a lot of money I could have measured the quality of the sensors in 2017, but thanks for your recommendation

Will your sensor drive 50 ohms?  Otherwise, you need to characterize the distortion of a voltage probe or preamp as well as the analyzer.

yes the output of the sensors all are 50 Ohm and all cables are RG316.


Since you want to measure individual harmonic levels and also THD, an instrument with the THD function would seem preferable to a basic instrument that lets you do the hard work.

I suggest you choose several candidate instruments.  Then download the brochures, specifications sheets, and operating manuals for each and compare them vs. price of available units.

A couple instruments that can do THD are the Agilent ESA-E and the R&S FSP.



Actually a basic/low distortion spectrum analyzer is good, I will measure the level of each harmonics and then calculate the THD. the picture you attached is interesting :) thank you very much, FSP3 spectrum analyzer seems good
https://cdn.rohde-schwarz.com/pws/dl_downloads/dl_common_library/dl_news_from_rs/187/n187_fsp.pdf

the below pic is for E4402B (ESA-E series):
https://www.keysight.com/us/en/assets/9018-03184/technical-specifications/9018-03184.pdf



the below pic is for E4403B (ESA-L series):

« Last Edit: March 29, 2021, 08:01:42 pm by xzswq21 »
❤ ❤
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #8 on: March 29, 2021, 04:29:06 am »
Finally today I want to buy a Siglent SSA3021x or HP8560E, which one is better? (I found a R&S FSP3 but it's expensive) Thanks
« Last Edit: March 29, 2021, 05:33:42 am by xzswq21 »
❤ ❤
 

Offline Bud

  • Super Contributor
  • ***
  • Posts: 7276
  • Country: ca
Re: A spectrum analyzer for measuring linearity
« Reply #9 on: March 29, 2021, 05:43:16 am »
Mentioning together Siglent and HP is an atrocity. In any case, the HP8560E will give you more dynamic range to measure 2nd harmonic before the device's own distortion kicks in.
Facebook-free life and Rigol-free shack.
 
The following users thanked this post: xzswq21

Offline Wuerstchenhund

  • Super Contributor
  • ***
  • Posts: 3088
  • Country: gb
  • Able to drop by occasionally only
Re: A spectrum analyzer for measuring linearity
« Reply #10 on: March 29, 2021, 07:14:30 am »
I have designed a new sensor in 2018 so I need to have a spectrum analyzer to measure the harmonics and noise of the sensor. Unfortunately I haven't accessed to a Spectrum Analyzer yet!

I apply 10KHz to my sensor and check the output response to measure HD2 (second harmonic distortion), HD3 (third harmonic distortion), THD then I will increase the input frequency upto 100MHz step by step (for example 100KHz, 200KHz, 300KHz,...., 1MHz,...) to plot HD2, HD3, THD over frequency axis, another test is only to plot magnitude of the bode plot of the sensors, this is all of I want :) finally I will publish the results in a journal.

I see most of good spectrum analyzers have Amplitude accuracy better than 1dB and it's very good for me, 90dBc/Hz of phase noise is very good, which spectrum is good for me? Old HP8560a, HP856x, E440x series or etc? Plus I've found HP3588a dynamic signal analyzer as an interesting choice:

https://accusrc.com/uploads/datasheets/4629_3588A.pdf

What do you think about HP3588a for my application?! HP8560 (a or e) or HP3588a?!

With that little information it's mostly up to what one can pull out of thin air.

First of all, you need to define what your requirements are. What's the frequency range that this sensor covers? What are the expected RF parameters (and have you confirmed them through simulation)? Because if you don't know how good the sensor is you don't know how good your test equipment must be, and then you're looking at big $$$ test equipment to find that out.

Also, since we're talking about some kind of sensor, have you considered the performance of the stimuli equipment and the impact it will have on your measurements?

Once it's clear what performance you expect from your UUT (and what limits there are because of your stimuli equipment) you can start narrowing down to suitable equipment that is good enough to confirm the UUT's RF performance.

Now, as far as test equipment is concerned, your budget is a very important factor. Because just buying some antique HP SA for a few bucks from ebay will not guarantee that you get the results you need. Test equipment ages, and pretty much everything which carries the HP label is now >23 years old and in many cases will have reached the limits of its service life. So instead of a functional test instrument you can rely on you might well end up with a project on its own (and fixing SAs is difficult if you don't have the equipment for it).

Lastly, since it seems you're mostly interested in THD, a THD analyzer might be a better alternative than a spectrum analyzer.

But first you need to do your homework and specify your requirements.
« Last Edit: March 29, 2021, 07:21:19 am by Wuerstchenhund »
 
The following users thanked this post: xzswq21

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #11 on: March 29, 2021, 07:34:09 am »
I analyzed all the circuits in theory and then simulated and built the sensors. In simulation the SFDR is from 70dBc to 90dBc. the results are very good for my application. For example if the input signal is 0dB the second harmonic distortion is 80dB or 85dB or something similar (it depends on the frequency). One of my sensors covers from 10KHz to 50MHz but another sensor can cover from 10KHz to 100MHz and all the outputs are 50 Ohm, 1Vp-p and the SNR is from 75 to <87 (it depends on the bandwidth). now I want to measure the HD2, HD3, THD, AC response (magnitude of the bode plot) of my sensors.
« Last Edit: March 29, 2021, 08:12:04 pm by xzswq21 »
❤ ❤
 

Offline srb1954

  • Super Contributor
  • ***
  • Posts: 1124
  • Country: nz
  • Retired Electronics Design Engineer
Re: A spectrum analyzer for measuring linearity
« Reply #12 on: March 29, 2021, 12:17:53 pm »
I analyzed all circuits in theory and then simulated and built the sensors. In simulation the SFDR is from 70dBc to 90dBc. the results are very good for my application. For example if the input signal is 0dB the second harmonic distortion is 80dB or 85dB or something similar (it depends on the frequency). One of my sensors covers from 10KHz to 50MHz but another sensor can cover from 10KHz to 100MHz and all the outputs are 50 Ohm, 1Vp-p and the SNR is from 75 to <87 (it depends on the bandwidth). now I want to measure the HD2, HD3, THD, AC response (magnitude of the bode plot) of my sensors.

I apply 10KHz to my sensor and check the output response to measure HD2 (second harmonic distortion), HD3 (third harmonic distortion), THD then I will increase the input frequency upto 100MHz step by step (for example 100KHz, 200KHz, 300KHz,...., 1MHz,...) to plot HD2, HD3, THD over frequency axis, another test is only to plot magnitude of the bode plot of the sensors, this is all of I want :) finally I will publish the results in a journal.
If you are applying an external excitation signal to your sensor you are going to need find a very clean signal source if you are hoping to measure distortion products below -80dBc for 100MHz signals. Ideally, your signal generator should have distortion products at least 10dB below than of the sensor under test for you to be able to have any confidence in the measured distortion figures of your sensor.

You might find the selection of a suitable signal generator a bigger problem than choosing a spectrum analyser.
 

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 6126
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #13 on: March 29, 2021, 02:26:37 pm »
I analyzed all circuits in theory and then simulated and built the sensors. In simulation the SFDR is from 70dBc to 90dBc. the results are very good for my application. For example if the input signal is 0dB the second harmonic distortion is 80dB or 85dB or something similar (it depends on the frequency). One of my sensors covers from 10KHz to 50MHz but another sensor can cover from 10KHz to 100MHz and all the outputs are 50 Ohm, 1Vp-p and the SNR is from 75 to <87 (it depends on the bandwidth). now I want to measure the HD2, HD3, THD, AC response (magnitude of the bode plot) of my sensors.

In theory, practice doesn't matter.  In practice, it does!
 
The following users thanked this post: Vovk_Z

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #14 on: March 29, 2021, 02:29:10 pm »
In theory, practice doesn't matter.  In practice, it does!

In practice I need a low distortion Spectrum Analyzer :)
« Last Edit: March 29, 2021, 08:03:35 pm by xzswq21 »
❤ ❤
 

Offline SilverSolder

  • Super Contributor
  • ***
  • Posts: 6126
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #15 on: March 29, 2021, 02:40:35 pm »

Is it an option to use a notch filter of some kind, to get rid of the fundamental, so you can measure the device using less esoteric equipment?
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 8996
  • Country: us
  • Retired, now restoring antique test equipment
Re: A spectrum analyzer for measuring linearity
« Reply #16 on: March 29, 2021, 03:40:36 pm »
That’s the normal method for audio distortion measurement, where the residual harmonics (after nulling) can be analyzed further in a SA.  Perhaps a passive RLC bridged-tee filter would work in his frequency range?  There is a reasonable amount of literature on this subject.
 

Offline xzswq21Topic starter

  • Frequent Contributor
  • **
  • Posts: 300
  • Country: 00
Re: A spectrum analyzer for measuring linearity
« Reply #17 on: March 29, 2021, 07:00:21 pm »
my brother told me Siglent SA is new, light weight, with USB port, LCD and etc and it's lovely! but if your are concerned about the distortion, HP8560E is a reference Spectrum Analyzer and has better Second harmonic distortion, I wanted to buy Siglent SA tonight but I'm skeptical to buy it! just look at:

SSA3021X distortion:


HP8560E distortion:

« Last Edit: March 29, 2021, 07:24:40 pm by xzswq21 »
❤ ❤
 

Offline Wuerstchenhund

  • Super Contributor
  • ***
  • Posts: 3088
  • Country: gb
  • Able to drop by occasionally only
Re: A spectrum analyzer for measuring linearity
« Reply #18 on: March 30, 2021, 08:05:30 am »
I analyzed all the circuits in theory and then simulated and built the sensors. In simulation the SFDR is from 70dBc to 90dBc. the results are very good for my application. For example if the input signal is 0dB the second harmonic distortion is 80dB or 85dB or something similar (it depends on the frequency). One of my sensors covers from 10KHz to 50MHz but another sensor can cover from 10KHz to 100MHz and all the outputs are 50 Ohm, 1Vp-p and the SNR is from 75 to <87 (it depends on the bandwidth). now I want to measure the HD2, HD3, THD, AC response (magnitude of the bode plot) of my sensors.

Simulation is fine, as it gives you some idea as to what performance you can expect, but especially in the world of RF just using simulation has its limits.

But leaving that aside, you'd be looking at a signal source which can deliver SHD of better than 95dB for your measurements to gain any realistic results. And as other members said, finding a suitable low distortion RF source will be more difficult than a suitable SA.

In theory, practice doesn't matter.  In practice, it does!

In practice I need a low distortion Spectrum Analyzer :)

Yes, you do. And neither the Siglent SSA (an entry-level device designed for low costs) nor the HP 8560E (a quarter of a century old SA designed for portable operation) are low distortion analyzers.

Since your budget seems to be a secret, have you considered just renting proper test equipment?
« Last Edit: March 30, 2021, 09:39:46 am by Wuerstchenhund »
 

Offline Shiv

  • Regular Contributor
  • *
  • Posts: 56
  • Country: es
Re: A spectrum analyzer for measuring linearity
« Reply #19 on: March 30, 2021, 11:10:36 am »
Hello
I have designed a new sensor in 2018 so I need to have a spectrum analyzer to measure the harmonics and noise of the sensor. Unfortunately I haven't accessed to a Spectrum Analyzer yet!

I analyzed all circuits in theory and then simulated and built the sensors. In simulation the SFDR is from 70dBc to 90dBc. the results are very good for my application. For example if the input signal is 0dB the second harmonic distortion is 80dB or 85dB or something similar (it depends on the frequency). One of my sensors covers from 10KHz to 50MHz but another sensor can cover from 10KHz to 100MHz and all the outputs are 50 Ohm, 1Vp-p and the SNR is from 75 to <87 (it depends on the bandwidth). now I want to measure the HD2, HD3, THD, AC response (magnitude of the bode plot) of my sensors.

I apply 10KHz to my sensor and check the output response to measure HD2 (second harmonic distortion), HD3 (third harmonic distortion), THD then I will increase the input frequency upto 100MHz step by step (for example 100KHz, 200KHz, 300KHz,...., 1MHz,...) to plot HD2, HD3, THD over frequency axis, another test is only to plot magnitude of the bode plot of the sensors, this is all of I want :) finally I will publish the results in a journal.

I see most of good spectrum analyzers have Amplitude accuracy better than 1dB and it's very good for me, 90dBc/Hz of phase noise is very good, which spectrum is good for me? Old HP8560a, HP856x, E440x series or etc?
Plus I've found HP3588a dynamic signal analyzer as an interesting choice but it's boatanchor!!:
https://accusrc.com/uploads/datasheets/4629_3588A.pdf

Thanks

-70 to -90 is a big difference.

What about the signal generator?

To measure -90dBc you should use a very, very clean signal ...

Y our cal lab some time ago, we needed to use filters to help cleaning the input signal and reducing the carrier on Analizer.

Maybe you could do this:
Very clean signal @ 10 MHz (<90dbC is very hard to find...or  HP + LP filters to make it clean)
Low Span measurement at 10 MHz
HP filter 15 MHz before the analyzer.
Low Span measurement at 20 MHz
Calibrate the HP filter response
Discount filter response to harmonic measurement.

So you don´t deal with Analyzer generated harmonics and dynamic range and just matters the noise floor.


BUT: if you want to check it at some frecuencies you will need a trailer full of attenuators. (We only take 2 or 3 special measurements)

I would go for external calibration or equipement rental

And trully I don´t know what to use to get so clean MHz.



« Last Edit: March 30, 2021, 11:24:06 am by Shiv »
 
The following users thanked this post: Wuerstchenhund


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf