How to detect the response signal's delay time during measuring the frequecny response?

[Roker]

Hi experts:

I am  a Labview new user, and I am now handling the frequecny response analyzing.

attached please find my stimulate signal and my response signal graph,then how can I get the delay time ,for example, if I sweep the frequency from 100Hz to 10kHz,cause there is always a unstable time delay and some noise,how can I target the 100Hz-10kHz signal from the response signal?

I was trying to use the express vi Crosscorelation.vi to retrieve the delay time ,but I don't know how to use it....

thanks very much!

Best Regards

Roker

[johnsold]

Roker,

I am not fully understanding what you are asking.  Do you want to measure the time delay between the stimulus signal and the response signal as a function of frequency?  Do you want the phase response as a function of frequency? Something else?

If my first question is what you want, then what is the expected range of delays?  Because your graphs have different time scales it is hard to estimate from those images.

Lynn

[Roker]

Hi Lynn,

thanks very much for your relpy and your attention 🙂

I have generated a stimulating signal(20Hz---20kHz,1/3 octave, and 10 cycles for every frequency point,as the attaeched file stimulate signal.jpg),and play it to a Bluetooth earphone(has A2DP profile) by a Bluetooth dongle,then record the response signal by using a couple of artificial ear(the response signal is as the attached file response signla.jpg).then I will get the sub signal according to the frequnecy resolution (1/3 octave) and the duration of every frequency point(for example,20Hz's duration is (1/20)*10, 40Hz's duration is (1/40)*10.....) by using the Vi named Get WaveForm Subset.vi. and then calculate the amplitude of every frequnecy point to analyze the Frequecny response.

and now the question is that the response signal has a variational time delay, I can not determine what the 20 Hz signal position is,what the 40Hz position is ....and so on.

I am not professional in this, so ,could you please give your comments on this issue ,or could you please advise another method to anlyze the frequnecy response?

Thanks very much

Best Regards

Roker

[johnsold]

Roker,

Thank you for the more detailed explanation.

Perhaps the first thing to try is to send the stimulus in bursts rather than continuously.  Send the 10 cycles of 20 Hz followed by period of no signal.  Then send the 40 Hz signal and pause again.  Of course this makes the total test take longer, but right now you are merely trying to get something to work.  Make the no signal periods long enough that you are sure you are not getting data segments confused.  You could start at 1 second for example.

Another approach is to generate a synchronizing signal between each segment.  Choose signal which will always get through and which can be easily discriminated from the others.  A multi-tone burst might be good.  Something like 1 kHz and 3.3 kHz simultaneously. If you need phase continuity, select a duration which gives an integer number of cycles for both frequencies.

Ideally you would record the stimulus and the response simultaneously (on separate channels), but if the delay is due to OS latencies for the Bluetooth parts, you might still have a problem.

After you get the signals recorded in a manner which allows you to know which part is a response to which stimulus, then you can proceed to the analysis.

Lynn