Phase measurement with order analysis

[Elnin83]

Hello,

It's my first request so ... sorrry for mistakes and misunderstanding.

I'm facing with a problem using a example call OAT High Speed Tacho.lvproj. It was adapted with my hardware :

•     NI -cRIO-9074 (slot 1 pin damaged)
•     NI-9234 slot 2 : ai0 used
•     NI-9205 slot 3 : ai23 differential mode used

The primary goal of my RT application is a phase measurement between a tacho signal and accelerometer signals. The tacho signal is not a counter but an analog sensor with sinus output.

I simulate the same signal with Waveform generator Agilent 33210A.The sinus parameters are : frequency 5 or 10 Hz, amplitude 1V or 2V, no offset. A Y coupling with BNC cables are used to send this signal to both analog input (tacho & accelerometer). I 'm aware about the problem of input delay caused by Delta-Sigma converter of NI-9234.So I decide to test with a zero phase signal to know how many samples are requested for compensation.

My problem is when I'm running the [RT] Cont Acq (AnalogTach).vi, I see different values in average phase indicator, even changing the amplitude or frequency of input sinus. The accelerometer seems to be not synchronize with a zero threshold detector for the tacho signal.

Parameters of the UI :

Threshold : 0V  ; Hysteresis : 0.5V

Sampling Frequency : 5120 samp/s ; No sample: 5120 (1s of signal)

9221 rate (it's 9205 rate) : 400 ticks (360 is the lowest I can get)

When I take a close look to the FPGA vi, I think that's all loops are functional, even if there are no trigger when tacho input passes zero. So what's wrong ?

Thanks.

[dsb@NI]

I expected to see a a vibration signal and a tachometer signal in your pictures, but only saw one sine wave. If you are measuring the phase difference between the analog tachometer (sine wave generated externally) and noise on a vibration channel, you should expect to see fluctuating measurements. "T" your tachometer signal into both the tachometer input channel and the vibration input channel,and see what you get.

[Elnin83]

Thanks for your reply Doug,

It's what I've done. The tachometer (ai23 of NI-9205) and vibration input chanel (ai0 of NI-9234) are the same signal. The red line is the ai1 NI-9234 which was not used during the test. The upper graph show waveform (output from WaveformCrio library)and the lower the ai0 after convert_to_even_angle_vi

According to OAT process, the tachometer info is used to resample the vibration signal into even angle signal, then OAT get magnitude & Phase give the right phase between tacho and vibration signal (1st order average phase). I was expected to see a value not equal to zero (because time delay of DSA NI-9234 was not compensate, sample correction =0 on picture) but I was surprised to see -20.53°, then stop - restart and see another value and so on.

My first problem was how synchronize a analog input from DSA 9234 and and analog input of SAR NI-9205 with FPGA Loop. It seems that the sample clock can not be shared between C-modules, Right ?

So I've found this example in NI.com : http://www.ni.com/example/31301/en/

I made few modifications on it and start with phase compensation. I understand the tach event count and the FPGA seems to be right for me. But I don't understand my result. I was expected a 40 samples correction (formula say 40/fs + 3.2 microsecond). For 10Hz signal input, it equivalent to 14°.

Is there any other method to synchronize both analog inputs from DSA and SAR modules in a FPGA ?