Author Topic: Bode Plot Computational Time for various DSOs  (Read 5539 times)

0 Members and 1 Guest are viewing this topic.

Offline mawyattTopic starter

  • Super Contributor
  • ***
  • Posts: 4116
  • Country: us
Bode Plot Computational Time for various DSOs
« on: November 07, 2022, 04:32:59 pm »
For a given set of Bode parameters, say 100Hz to 100KHz, dec sweep with 20pts/dec, dynamic scaling ON, how do the various DSOs compare time-wise?

Our particular interest is with the various Siglent flavors, like the SDS2000+P, SDS2000HD and SDS6000, but also interested in how other brands perform, including firmware level.

This might be a good means to keep tabs on the various firmware updates also, as the OEMs improve things. Know that different algorithms will play into this, some with very good frequency selectivity will require more computational load, thus longer timeframes, so not exactly an ideal comparison and tabulation.

Also know a few don't feel the DSO Bode Plot capability is useful, very likely because they don't have, nor experience, with a DSO with a good Bode Plot implementation, so please refrain from commenting. Our interest is from those that actually use and value this feature, as we've found it quite useful indeed, albeit slow, so we view relative timescales useful information!!

Best,

Edit: Added for SDS2102X+ and SDS2104X+  Parameters: Simple, Single Sweep, Decade, 100-100KHz, 20 pts/dec, Int AWG

AWG output to CH1 with BNC cable, another BNC to BNC cable from CH1 to CH2.

SDS2102X+  HW 02-08   FPGA 2021-07-16   U-OS 5.0   SW 1.3.9R6      Time   1:10

SDS2104X+  HW 02-04   FPGA 2022-04-06   U-OS 5.0   SW 1.5.2R1      Time   1:45
« Last Edit: November 07, 2022, 07:05:37 pm by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 
The following users thanked this post: RoGeorge, Martin72

Online Martin72

  • Super Contributor
  • ***
  • Posts: 7018
  • Country: de
  • Testfield Technician
Re: Bode Plot Computational Time for various DSOs
« Reply #1 on: November 07, 2022, 04:45:10 pm »
Hi,

For the HD I can do it maybe tomorrow, for the X+ next week at work and for the 6000....I´m out... ;D
Then finding rigol mso5000 and keysight users...
Lecroy didn´t have bode, tek ??
GwInstek, Hantek, Micsig, Owon ???

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Bode Plot Computational Time for various DSOs
« Reply #2 on: November 07, 2022, 05:33:05 pm »
RTM3004: 12 seconds
BTW, I also played around a bit with the SMD test fixture to repeat the test with the capacitors (1pf and 1nf), and on the RTM3004 the results show similar accuracy.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Online Martin72

  • Super Contributor
  • ***
  • Posts: 7018
  • Country: de
  • Testfield Technician
Re: Bode Plot Computational Time for various DSOs
« Reply #3 on: November 07, 2022, 05:48:14 pm »
What did you use as a DUT ?

Offline balnazzar

  • Frequent Contributor
  • **
  • Posts: 417
  • Country: it
Re: Bode Plot Computational Time for various DSOs
« Reply #4 on: November 07, 2022, 05:49:22 pm »
GwInstek, Hantek, Micsig, Owon ???

No FRA on those, AFAIK.
 

Offline TopQuark

  • Frequent Contributor
  • **
  • Posts: 326
  • Country: hk
Re: Bode Plot Computational Time for various DSOs
« Reply #5 on: November 07, 2022, 05:56:32 pm »
SDS2000X HD, software version 1.2.0.2, FPGA version 2022-07-08

Parameters: Simple, Single Sweep, Decade, 100-100KHz, 20 pts/dec, Ext AWG

1k ohm and 100nF low pass filter (note the circuit type will affect how many times the dynamic scaling will have to adjust the vertical scale)

1 min 53 sec.
 
The following users thanked this post: Martin72

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Bode Plot Computational Time for various DSOs
« Reply #6 on: November 07, 2022, 06:11:53 pm »
GwInstek, Hantek, Micsig, Owon ???

No FRA on those, AFAIK.
GW Instek has FRA on the models with an internal generator. But not many have those units as GW Instek released the more elaborate models later on.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 
The following users thanked this post: balnazzar

Offline mawyattTopic starter

  • Super Contributor
  • ***
  • Posts: 4116
  • Country: us
Re: Bode Plot Computational Time for various DSOs
« Reply #7 on: November 07, 2022, 07:04:00 pm »
What did you use as a DUT ?

In our case just a pass thru BNC cable and another BNC cable to the AWG output. I'll add to OP.

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Offline mawyattTopic starter

  • Super Contributor
  • ***
  • Posts: 4116
  • Country: us
Re: Bode Plot Computational Time for various DSOs
« Reply #8 on: November 07, 2022, 07:08:55 pm »
SDS2000X HD, software version 1.2.0.2, FPGA version 2022-07-08

Parameters: Simple, Single Sweep, Decade, 100-100KHz, 20 pts/dec, Ext AWG

1k ohm and 100nF low pass filter (note the circuit type will affect how many times the dynamic scaling will have to adjust the vertical scale)

1 min 53 sec.

A quick test with Ext AWG by LAN added ~ 5s from using internal AWG, so seems the X+ and XHD are about the same.
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Offline Fgrir

  • Regular Contributor
  • *
  • Posts: 164
  • Country: us
Re: Bode Plot Computational Time for various DSOs
« Reply #9 on: November 07, 2022, 07:30:23 pm »
RTB2004 with FW version 2.400, BNC cables from Aux Out to Ch1 and Ch1 to Ch2.  Bode plot for 100Hz to 100KHz, 20pt/dec takes 9s.
 

Offline colorado.rob

  • Frequent Contributor
  • **
  • Posts: 420
  • Country: us
Re: Bode Plot Computational Time for various DSOs
« Reply #10 on: November 07, 2022, 07:47:25 pm »
What accounts for the 2 orders of magnitude time difference to acquire a Bode plot between Siglent and R&S scopes?
 

Online Martin72

  • Super Contributor
  • ***
  • Posts: 7018
  • Country: de
  • Testfield Technician
Re: Bode Plot Computational Time for various DSOs
« Reply #11 on: November 07, 2022, 08:16:44 pm »
SDS2504X HD, 1.2.1.1

Quote
Parameters: Simple, Single Sweep, Decade, 100-100KHz, 20 pts/dec, Int AWG

AWG output to CH1 with BNC cable, another BNC to BNC cable from CH1 to CH2.

Done the settings, let it run...
And then checking it three times because:

3:35min.... ;)

Online Martin72

  • Super Contributor
  • ***
  • Posts: 7018
  • Country: de
  • Testfield Technician
Re: Bode Plot Computational Time for various DSOs
« Reply #12 on: November 07, 2022, 08:29:12 pm »
Pics:
Setting, final measure.
Interesting:

10Hz-100Hz : 2:03min
100Hz- 1Khz : 1:04min
1Khz-100Khz : 28sec.

EDIT:

Didn´t start from 100Hz like mawyatt mentioned.
Here the time from 100Hz- 100Khz, new measured:

1:45min

« Last Edit: November 07, 2022, 08:36:36 pm by Martin72 »
 

Offline mawyattTopic starter

  • Super Contributor
  • ***
  • Posts: 4116
  • Country: us
Re: Bode Plot Computational Time for various DSOs
« Reply #13 on: November 07, 2022, 08:45:04 pm »
What accounts for the 2 orders of magnitude time difference to acquire a Bode plot between Siglent and R&S scopes?

See this other Bode related thread just started, my have some clues to the large time discrepancy between various Bode Implementations.

https://www.eevblog.com/forum/testgear/bode-plot-torture-test/

Best,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Online Andreas

  • Super Contributor
  • ***
  • Posts: 3350
  • Country: de
Re: Bode Plot Computational Time for various DSOs
« Reply #14 on: November 07, 2022, 09:04:06 pm »
Hello,

PicoScope 5444A with FRA 4PicoScope 0.7.3b  100Hz-100kHz 20 steps/dec -> ~5sec

with best regards

Andreas
 
The following users thanked this post: Martin72

Online Martin72

  • Super Contributor
  • ***
  • Posts: 7018
  • Country: de
  • Testfield Technician
Re: Bode Plot Computational Time for various DSOs
« Reply #15 on: November 07, 2022, 10:08:58 pm »
First of all I really don´t have a problem with it, but....
It´s remarkable how short others can perform it while the siglents take their time....
In this case minutes against seconds...
On the siglent you got the feeling, they´re switching completely in a different mode, the whole scope becomes a FRA.. - When bode menu is only activated (not starting), you can´t do anything else on the scope until you leave the bode mode.
And leaving the bode mode takes also time....
Won´t believe it´s somekind of performance lack.


Offline Someone

  • Super Contributor
  • ***
  • Posts: 5154
  • Country: au
    • send complaints here
Re: Bode Plot Computational Time for various DSOs
« Reply #16 on: November 07, 2022, 10:41:38 pm »
It´s remarkable how short others can perform it while the siglents take their time....
In this case minutes against seconds...
On the siglent you got the feeling, they´re switching completely in a different mode, the whole scope becomes a FRA.. - When bode menu is only activated (not starting), you can´t do anything else on the scope until you leave the bode mode.
And leaving the bode mode takes also time....
Won´t believe it´s somekind of performance lack.
The widely variable time is somewhat related to performance and the design tradeoffs that the manufacturer has chosen:
https://www.eevblog.com/forum/testgear/faster-fra-from-scope-by-external-control-over-visa/

As an example for this thread, that Keysight DSOX1102G takes 42 seconds for the above requested 100-100k, 20pts/decade measurement. But 4 seconds of that is the autoscaling of the input attenuators. Swap to a script running on a (slow) computer and you can have a similar quality answer in 14 seconds, but now with the user being able to choose how much averaging to apply at each frequency step.

Want to go faster? A chirp excitation should be able to do this in a few seconds.
 

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 7462
  • Country: hr
Re: Bode Plot Computational Time for various DSOs
« Reply #17 on: November 07, 2022, 10:45:15 pm »
What accounts for the 2 orders of magnitude time difference to acquire a Bode plot between Siglent and R&S scopes?

Taking 5 averages at every frequency.
Biggest impact is at low frequencies. It takes time to gather enough samples of a signal that has period of 10 ms.

Benefit is that it works with noisy signal (like switchers) and gets few dB better dynamic range..

MSOX3104T did it in 35sec with simple DUT (few resistors and capacitors). It would have been a bit slower if  DUT was exercising autoranging.
But it gave up when pushed to small signals.....

Siglent's implementation at this moment doesn't win speed competition. But is very accurate, has best dynamic range, has capability to do multichannel Bode plot (one input and 3 outputs in 4 ch scope) has automatic measurements, best frequency range, works with outside AWG (interesting things can be done there with large signal and channel locking) and has very good accuracy...
« Last Edit: November 07, 2022, 11:00:49 pm by 2N3055 »
"Just hard work is not enough - it must be applied sensibly."
Dr. Richard W. Hamming
 

Offline gf

  • Super Contributor
  • ***
  • Posts: 1425
  • Country: de
Re: Bode Plot Computational Time for various DSOs
« Reply #18 on: November 07, 2022, 11:12:46 pm »
It´s remarkable how short others can perform it while the siglents take their time....

There is always a trade-off between measurement duration and dynamic range (noise floor).
Comparing the duration is IMO rather meaningless if the achieved dynamic range is not compared as well.
On the other hand, if the dynamic range of the DUT happens to be much lower in the frequency span of interest, then aiming for a too large dynamic range can result in wasted time.
 
The following users thanked this post: Someone, 2N3055, Martin72

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Bode Plot Computational Time for various DSOs
« Reply #19 on: November 07, 2022, 11:31:38 pm »
It´s remarkable how short others can perform it while the siglents take their time....

There is always a trade-off between measurement duration and dynamic range (noise floor).
Comparing the duration is IMO rather meaningless if the achieved dynamic range is not compared as well.
On the other hand, if the dynamic range of the DUT happens to be much lower in the frequency span of interest, then aiming for a too large dynamic range can result in wasted time.
I agree. I have the feeling the engineers at Siglent geeked out on getting the lowest noise floor while losing sight of what is also important: speed.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline BillyO

  • Super Contributor
  • ***
  • Posts: 1847
  • Country: ca
Re: Bode Plot Computational Time for various DSOs
« Reply #20 on: November 08, 2022, 12:02:50 am »
I agree. I have the feeling the engineers at Siglent geeked out on getting the lowest noise floor while losing sight of what is also important: speed.
Speed?  Is it a race or an exercise in getting good results?
Bill  (Currently a Siglent fanboy)
--------------------------------------------------
 
The following users thanked this post: Martin72

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Bode Plot Computational Time for various DSOs
« Reply #21 on: November 08, 2022, 12:33:19 am »
I agree. I have the feeling the engineers at Siglent geeked out on getting the lowest noise floor while losing sight of what is also important: speed.
Speed?  Is it a race or an exercise in getting good results?
Well, if you look in the other thread Mawyatt has created, you'll see that the same results can be obtained in a matter of seconds. There really is no need to go slow. When I'm tweaking a circuit, I don't want to wait for over a minute to see the results of the changes. Heck, you can simulate a circuit faster nowadays. IOW: There is a tradeoff between getting the lowest noise floor and making a feature that is useable. In most cases you'll be working in situations where there is more than enough headroom.
« Last Edit: November 08, 2022, 12:38:33 am by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Someone

  • Super Contributor
  • ***
  • Posts: 5154
  • Country: au
    • send complaints here
Re: Bode Plot Computational Time for various DSOs
« Reply #22 on: November 08, 2022, 12:41:34 am »
I agree. I have the feeling the engineers at Siglent geeked out on getting the lowest noise floor while losing sight of what is also important: speed.
Speed?  Is it a race or an exercise in getting good results?
Well, if you look in the other thread Mawyatt has created, you'll see that the same results can be obtained in a matter of seconds. There really is no need to go slow. When I'm tweaking a circuit, I don't want to wait for over a minute to see the results of the changes. Heck, you can simulate a circuit faster nowadays.
Well, referring to random other thread isn't really helping when you put it out of context like that (that thread doesn't mention the elapsed time of the test). Just to ruin it further I'm not sure your LARGE (in size) contributions to that thread were actually comparable to what mawyatt provided as their example:
https://www.eevblog.com/forum/testgear/bode-plot-torture-test/
Fast = short time captures and less noise rejection, cant get around that when keeping other conditions the same.
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: Bode Plot Computational Time for various DSOs
« Reply #23 on: November 08, 2022, 12:52:34 am »
I guess you missed the fact that the RTM3004 offers at least the same results in the same noise / disturbance conditions while needing about 10 times less time. It is not my fault mawyatt created two threads to discuss closely related topics.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline Someone

  • Super Contributor
  • ***
  • Posts: 5154
  • Country: au
    • send complaints here
Re: Bode Plot Computational Time for various DSOs
« Reply #24 on: November 08, 2022, 01:09:14 am »
I guess you missed the fact that the RTM3004 offers at least the same results in the same noise / disturbance conditions while needing about 10 times less time. It is not my fault mawyatt created two threads to discuss closely related topics.
Lol, it cant get the same results that much faster. There isn't enough time capture the data that would suppress the interference. I'm guessing your configuration was not comparable to the original post... in that other thread where you could be discussing this rather than trying to blame others for your inability to maintain context.  :clap:
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf