[100pts] Signal analysis - need progressive material to read

[100pts] Signal analysis

Hello Edaborders,

I'll give 20pts for the best reply to each question.


I have no experience nor ideas how to analyse signals. I wanna measure, decide and prove whether a signal is good or bad.


1- For the moment I plotted the frequency of my signal vs time. The signal is stable at 12 MHz. Actually, according to the figure, it is 12 Mhz +/- 0.04 %.
Is it a good signal ?

2- Oscillations are between 12.005 MHz and 11.995 MHz. The corresponding period is respectively 80.32 ns and 80.33 ns.
Can I say that the jitter of the signal 0.01 ns ?

3- Can I extract other usefull informations from the figure ?

4- What are the creteria that let me say that the signal is good ?

5- Will the study of this signal in the frequency domain bring more information ?

Could someone refert me to the necessary progressive material to read in order to understand that issue.

[100pts] Signal analysis

You need to define some specifications and then compare this signal versus your specs. Then, and only then, you can decide wether your signal is "good" or not!

2- Yes.
5- Yes, it will allow you to see if there are systematic patterns.

Re: [100pts] Signal analysis

Hi all,


1- Good or bad depends on your application requirements
2- The maximum jitter is 0.01ns but you'd better find also the other jitter's figures
3- As it is ? No.
4- You have to consider jitter and noise
5- Sure.

[100pts] Signal analysis

I don't understand, how the frequency measurement shown in the plot could be derived from a real signal. Can you clarify?

Re: [100pts] Signal analysis

Hi all,

Thanks JoannesPaulus and Advares for your replies. Could you please elabore the last two points ?

Please feel free to post your comments and replies.

FvM said:

I don't understand, how the frequency measurement shown in the plot could be derived from a real signal. Can you clarify?

Click to expand...

Thanks FvM. I'm sure that other member like you didn't get my questions.
The signal In the figure comes really from a real signal. It represents the Frequency vs Time of a PLL output in a very small time window when the PLL is locked. (Please se figure below)

[100pts] Signal analysis

Thank you for the clarification. So I think, that the displayed graph has to be smoothed respectively low-pass filtered to represent a meaningful signal parameter. Details e.g. shorter than a signal period must be rather considered as noise, they can't represent a frequency. This has be considered, when calculating e.g. a jitter. As another problem, a jitter specification must be referred to a period, e. g. cycle-to-cycle or specified long-term.