AnsweredAssumed Answered

MSOX2024A Frequency Measurment Instability

Question asked by jstern on Apr 13, 2013
Latest reply on Apr 17, 2013 by algoss

Like • Show 0 Likes0
Comment • 3

I recently upgraded to the MSOX2024A and am getting familiar with the scope. My only DSO comparison is to my Rigol DS1102D. I noticed in the Agilent 30-day Training Signals trial, that on measurements there seems to be much more cursor 'jiggle' particularly on the X axis (Freq) and much more than I have seen on my Rigol. The Rigol measurement function can read out Freq to two significant digits like the MSOX2024A. Using the MSOX2024A wave gen with a 10.00Mhz signal, it jiggles from 9.95-10.05. I thought maybe the problem was the Wave Gen option but I took a high-imp 10.0000000 Mhz signal (from a GPS-Rb locked HP SigGen) and its reads a rock steady 10.00 Mhz on the Rigol while it bounces around on on the MSOX2024A. The Rigol is in same normal capture mode as the 2000x. Is this normal behaviour for the MSOX2024A?

thanks
Jerry

No one else has this question

Outcomes

hithesh123 Apr 15, 2013 1:09 AM

Jerry,

This is not instability. Look up the specs of the wavegen. It has 0.5nsecs jitter.
Jitter is just like noise.
It adds/subtracts to your waveform period (T+/-Jitter). If you do the math for a 10MHz wave -
Period T = 100nSecs. With Jitter the period will be 100.5nsecs and 99.5nsecs
This gives a frequency of 9.95MHz and 10.05MHz, which is what the MSO is displaying.
The RIGOL is a less accurate instrument.

Actions

jstern @ hithesh123 on Apr 15, 2013 4:00 AM

Thanks for the explanation. I am NOT trying to disrepsct the MSOX2024A, its a great scope with many, many more features than my Rigol. But with regards to the frequency measurement and jitter, how is the Rigol not as accurate? The defininton of accurcy : "The ability of a measurement to match the actual value of the quantity being measured" It seems the Rigol frequency measurement has less Jitter and its measurement displays more of the actual value than the MSOX2024A?"

thanks
Jerry

Actions

algoss

@ jstern on Apr 17, 2013 8:13 AM

Ok, it's a little more complicated than accuracy, etc.

A lot depends on the BW, noise, etc. of the scope, and how the scope makes the measurement. I don't know anything about the specifications of the Rigol scope, or how it makes the measurement. It may, for instance, limit its display of frequency measurements to 3 or 4 significant digits.

The Agilent scope makes its frequency (and other time-related) measurements by looking at the time between calculated crossing points. It uses the captured voltage values and does interpolation between then to display the waveform and find the crossing points. Noise on the signal, either from the target or within the scope (and there will ALWAYS be some), causes the captured points to move up and down. This vertical noise makes the calculated crossing time move horizontally. Since both crossing points are moving randomly, this makes the calculated frequency change.

I don't know if the 2000X uses cursor to make the measurement, but I'll assume it does. There is a listed specification in the DataSheet for cursor measurements:
+/-(time base * reading) +/- (.0016 * screen width) +/- 100 psec

Since 2 cursors would be used to make this measurement, this would be doubled, the frequency result would be the inverse. Each of the three separate values used in the calculation would have 'normal' distributions, so I could see a reasonable amount of variation in the results.

Higher BW Agilent scopes further specify the timing inaccuracies using slew rate, since a slower slew rate would increase the effect of the noise on the measurement.

The short answer is, as you get more resolution, you will bee more variability in the answer.

Al

Actions

3 Replies

hithesh123 Apr 15, 2013 1:09 AM
Mark Correct
Correct Answer
Jerry,

This is not instability. Look up the specs of the wavegen. It has 0.5nsecs jitter.
Jitter is just like noise.
It adds/subtracts to your waveform period (T+/-Jitter). If you do the math for a 10MHz wave -
Period T = 100nSecs. With Jitter the period will be 100.5nsecs and 99.5nsecs
This gives a frequency of 9.95MHz and 10.05MHz, which is what the MSO is displaying.
The RIGOL is a less accurate instrument.
Like • Show 0 Likes0
Actions
- jstern @ hithesh123 on Apr 15, 2013 4:00 AM
  Mark Correct
  Correct Answer
  Thanks for the explanation. I am NOT trying to disrepsct the MSOX2024A, its a great scope with many, many more features than my Rigol. But with regards to the frequency measurement and jitter, how is the Rigol not as accurate? The defininton of accurcy : "The ability of a measurement to match the actual value of the quantity being measured" It seems the Rigol frequency measurement has less Jitter and its measurement displays more of the actual value than the MSOX2024A?"
  
  thanks
  Jerry
  Like • Show 0 Likes0
  Actions
  - algoss @ jstern on Apr 17, 2013 8:13 AM
    Mark Correct
    Correct Answer
    Ok, it's a little more complicated than accuracy, etc.
    
    A lot depends on the BW, noise, etc. of the scope, and how the scope makes the measurement. I don't know anything about the specifications of the Rigol scope, or how it makes the measurement. It may, for instance, limit its display of frequency measurements to 3 or 4 significant digits.
    
    The Agilent scope makes its frequency (and other time-related) measurements by looking at the time between calculated crossing points. It uses the captured voltage values and does interpolation between then to display the waveform and find the crossing points. Noise on the signal, either from the target or within the scope (and there will ALWAYS be some), causes the captured points to move up and down. This vertical noise makes the calculated crossing time move horizontally. Since both crossing points are moving randomly, this makes the calculated frequency change.
    
    I don't know if the 2000X uses cursor to make the measurement, but I'll assume it does. There is a listed specification in the DataSheet for cursor measurements:
    +/-(time base * reading) +/- (.0016 * screen width) +/- 100 psec
    
    Since 2 cursors would be used to make this measurement, this would be doubled, the frequency result would be the inverse. Each of the three separate values used in the calculation would have 'normal' distributions, so I could see a reasonable amount of variation in the results.
    
    Higher BW Agilent scopes further specify the timing inaccuracies using slew rate, since a slower slew rate would increase the effect of the noise on the measurement.
    
    The short answer is, as you get more resolution, you will bee more variability in the answer.
    
    Al
    Like • Show 0 Likes0
    Actions

MSOX2024A Frequency Measurment Instability

Outcomes

Related Content

Recommended Content