Interchannel delay correction

[dariodicara]

Hi, I've read many comments about Interchannel delay but I haven't  still understood if it’s possible to correct this error.

I want to submit my problem: I must measure the phase angle difference between to different signal composed by a fundamental and a series of Harmonics. I want to know  the delay between each harmonic of the first signal and the correspondent of the second signal. I acquire this two signal in different mode by means of a DAQCard-AI-16XE-50 that is a multiplexer acquisition card (with only one A/D converter). I understood that I've an interchannel delay but I want to correct it.

I've already find a method to correct this delay but I think that it is not enough accurate. I want to measure the phase delay for each harmonic and so I use a AI Acquire Waveforms.vi and the Ex.ExtractSingleTone.vi to measure Phase and Amplitude for each harmonic. After I make the difference between the phase angle of the two signal for each harmonic. To correct the interchannel delay I carry out a first measure driving the two channel with the same signal and I use the phase delay measured for the fundamental as the interchannel delay. In this way I correct the phase angle delay between the two channel subtracting the interchannel delay (measured as I’ve explained) times the order of the harmonic.

The interchannel delay measured in this way is 0.89 deg when I acquire 10 kS/s (1000 sample) gain 2. I chose this frequency because in the specification is written:

Max sampling rate:       200 kS/s (single-channel),

20 kS/s guaranteed (scanning; gain = 1, 2, 10),

17 kS/s (scanning; gain = 100)

My first question is: Can I increase the sample frequency? I’ve tried but the situation become worst and so I think that 10 kS/s =20kS/2channel is the limit. Is that right?

Second question if instead of measure the interchannel phase delay, can I impose the iterchannel delay in AI Config.vi. In the help of this VI they say that I can do this if my card has both scan clock and channel clock but I can’t understand if my device has channel clock. Or better I think that this Cars hasn’t channel block.

Third question can I use Align Waveform (continuous).vi to solve my problem? But in the example called correct for MIO inter channel delay (continuous).vi it shows four waveforms and for each waveform the sampled points correspond to different time (because of the interchannel delay) instead when I zoom my waveforms the sampled points of the two signal correspond to the same instant (as there wasn’t interchannel delay). Do I have make some mistakes in using those VIs?

Finally does anybody have a better solution for my problem?

I thank whoever wants to reply me,

I’m sorry if there is some mistakes in my English,

Dario

  

[Chuck_81]

Hello Dario,

if you are interested in minimizing (or totally removing) your inter-channel delay, you may want to consider the following. Since you know the value of your delay (that can be computed as explained here), you can either correct for it in the frequency domain (if you are doing frequency domain analysis) or even process your signals in the time domain using appropriate filtering.

Moreover, if you choose to correct for the inter-channel delay in a post-processing operation, it may actually be better to maximize (!) your inter-channel delay in order to minimize inter-channel cross-talk caused by the physical settling time of your board. The AI-Config VI provides an input where you can control the interchannel delay (as a multiple of 5 us in your case) so with 4 channels configured you could enter for example 10 us.

Hope this helps,

best regards,

Fabio

[dariodicara]

Thank you a lot for your quick answer,

I’ve changed the interchannel delay in AiConfig.vi but I couldn’t try it until Monday because actually I’m aware from my laboratory. However on Monday or Tuesday I will say you if it will work. I have some doubts because the help of AiConfig.vi said that modifation of interchannel delay is possible only for the card that has two clock: scan and channel clock. And in the specification of my card there isn’t mentioned the channel clock.

However I’ve only a question you have said 5 us = 1/200kS/s, but the specification of my card says the the single channel acquisition frequency is 200kS/s, instead for multichannel acquisition the said that only 20kS/s are garanteed. Moreover when I measure the phase angle error between the same signal acquired by two different channel I found 0.89 deg that for 50Hz is 49 us. Infact I think that I should chose as interchannel delay 50 us or 100 us. Moreover I’ve read that my card as this Settling time for full-scale step:(Gain = 1, 2, 10) 50 us max to ±2 LSB. Then I must set (if it is possible) interchannel delay > 50 us. Is it right?

I will post the results of my experimental tests on Tuesday. I will try to use also Align Waveform (continuous).vi to solve this problem and after I will let you know the results.

Best regards,

Dario