Measuring the subtract of two phases

[Dimon87]

Hi. I use the PCI-6221 device: 1 output and 2 input channels. I connected output and 2 input channels together with the wire. I put on the block diagram 2 extract signal tone information elements and connected them to the input channels. Then I wanted to subtract 2 phases. I generated the sine signal. The subtract was  0, but when I was changing the amplitude or the frequancy of the sine signal the subtract didn't become 0. It was 65,-23,... . Why does happen it? How can I solve this problem?

Thank You.

[tnek]

Hi Dimon87,

The difference in phase that you are seeing is probably due to the interchannel delay between channels. This occurs because all of the channels on this board are multiplexed together so there is a difference in time from when one channel is sampled to the next. This article gives an expanded definition for interchannel delay.

Ideally, when making phase measurements a simultaneous sampling board should be used such as the S-Series DAQ cards so the phase information is preserved between channels. Each channel on these boards will have a dedicated A/D converter. However, there are ways to correct for this phase difference. One way is to use any of the Align functions in the signal processing palette. In addition, examples on how to use this function can be found in the Example Finder. This article gives a few more details about these examples. Let me know if this helps.

Regards,

Kent

Applications Engineer

[Dimon87]

Hi, thank You for reply. Can I decrease the interchannel delay using special functions in DAQ palette?

[Dimon87]

As I understand, interchannel delay=1/convert rate. How can I get to know what is the convert rate in the PCI-6221. So if I can minimize the convert rate, then the subtract of the two phases will be corrected. How can I do this?

The sample rate of 1 AI =250kS/s in PCI-6221. I use 5 AI and now the sample rate is 50kS/s. Is this correct? 

Thank You.

[tnek]

Hi Dimon87,

The interchannel delay can be set to an extent using the DAQmx Timing property nodes. However, you will run into a minimum interchannel delay as mentioned in this article since this device you are using has a multiplexer. This means that the channels cannot be sampled at the same time and some amount of interchannel delay will exist. More information about the convert clock can be found here if you are interested. If you wanted to try to account for this delay, you can use the align waveform functions. Also, your calculation of maximum sample rate is correct. This value can also be determined programatically as documented here. Hopefully this clarifies a few things.

Regards,

Kent

Applications Engineer

[Dimon87]

Hi. I set the convert rate=250kS/s, calculate the interchannel delay, translate it to the phase and subtract it and all is OK, but the dependence upon the amplitude of the output signal is remained.

I set the sampling rate of the output signal =800kS/s, and the sampling rate of the input is 50kS/s, then there were bad processes with the waveform.

How can I correct them?

Thank You  

[tnek]

Hi Dimon87,

What do you mean by "bad processes"? From what have mentioned, everything should be fine as long as all of the sampling considerations are taken into account. What is the frequency of the signal that you are generating?

Regards,

Kent

Applications Engineer

[Dimon87]

Hi, I have the problem. I work with PCI-6221, 1 output and 3 inputs.  From the LV generator I send the sine wavefrom with the ampl=2V to the input of the RLC circuit. The output (capacitor) I connect to the input of the ADC. In LV I measure the RMS of the output (from capacitor) signal and the subtract of 2 phases (input, output). I change the frequancy of the generator. At first the RMS increases then after the peak it decreases with hops, not fluently. Also the subtract is changed. It is wrong

What I should do?

I send my VI.

Thank You.

[tnek]

Hi Dimon,

I took a quick look at your code and everything looks good. Does this behavior also occur when you just feed the output straight to the input? We should see the correct RMS values doing this. If this is correct, you may want to check your external circuit to make sure everything is placed as expected.

Regards,

Kent

Applications Engineer

[RavensFan]

First, when you post a VI, please post one that has meaningful text names for your controls.  What you have is nothing but goofy, garbage, accented characters that don't mean anything.  Even your subVI (which you didn't post) has a problem with its name.  It's impossible to know exactly what your controls and clusters mean.

I think your problem may be related to aliasing and Nyquist theorem.  You are increasing the frequency of your analog output and adjusting your analog output parameters accordingling.  But your analog input is always humming along at the same data acquisition rate.  As your output frequency increases, so does your input frequency, but you aren't adjusting your data acquisition parameters to compensate.  So your acquired waveform no longer adequately represents the actual waveform.

Message Edited by Ravens Fan on 09-24-2008 11:27 PM