How to disable antialias filter with PXI5124

[JoeFriedchicken]

Hi Shiying,

You haven't been very specific about what kind of measurements you're making. All we know is you want to measure frequency content as high as 20 MHz, but you don't say how large the AC signal is that you're measuring.

My recommendation for you is to use one of the NI-Scope shipping examples: "niScope EX Configured Acquisition.vi". You can open this from the Example Finder: Help menu » Find Examples... and then Hardware Input and Output » Modular Instruments » NI-Scope » General. This VI will allow you to fully configure the scope for whatever signal you're measuring.

If you want to test whether or not the 20 MHz filter is working, set the "Max. Input Frequency" to 20 MHz, and the "Min. Sample Rate" to 200 MHz, and then try to measure a signal with frequency content higher than 20 MHz (30 MHz like Thomas suggested, or higher) and see if the measurement attenuates at those higher frequencies. For example, if you send a 1 V peak-to-peak signal at 50 MHz, then you should see that same signal on the scope, but at a lower amplitude. If you do, then the filter is working. The higher the frequency, the greater the attenuation.

[Shiying]

Hi Joe,

My application is to measure the output noise peak to peak voltage of the UUT (Unit Under Test).  Depend on different UUT, the DC output value is in the range of 3 - 35 V.  The AC noise value (peak to peak) should be less than 1V in the 20 MHz bandwidth.  What is your suggestion for the "Configure Vertical" and "Configure Horizontal"?  Thank you.

Shiying 

[Shiying]

Joe,

You said that 1V peak to peak 50 MHz signal can't be attenuated with the 20MHz filter on.  Then how low the amplitude can be attenuated?  Thank you.

Shiying

[JoeFriedchicken]

Hi Shiying,

I didn't say that a 50MHz signal cannot be attenuated with the 20MHz filter. I said that the signal will have a smaller amplitude: "For example, if you send a 1 V peak-to-peak signal at 50 MHz, then you should see that same signal on the scope, but at a lower amplitude. If you do, then the filter is working. The higher the frequency, the greater the attenuation."

When configuring your vertical and horizontal parameters, you need to keep a few things in mind.

niScope Configure Vertical [1]:

1. vertical range -- this is the maximum peak-to-peak amplitude you expect to measure. It sounds like you need to set this to 1 Volt.
2. vertical offset -- this is the DC component of the signal. This will change according to your UUTs.

niScope Configure Horizontal Timing [2]:

1. min sample rate -- set this to at least 40 MHz to meet Nyquist limits for your 20MHz measurements. Setting it to 10x higher than the highest frequency content you expect will help give you a good shape of the waveform, although the 20MHz filter will attenuate the higher frequencies.

niScope Configure Chan Characteristics [3]:

1. max input frequency -- set this to 20 MHz or lower to turn on the 20 MHz filter. If you set it to 0, the default 60MHz anti-alias filter will be enabled instead.

[1] niScope Configure Vertical
NI High-Speed Digitizers Help >> Programming >> Reference >> NI-SCOPE LabVIEW Reference Help >> VIs

[2] niScope Configure Horizontal Timing
NI High-Speed Digitizers Help >> Programming >> Reference >> NI-SCOPE LabVIEW Reference Help >> VIs

[3] niScope Configure Chan Characteristics
NI High-Speed Digitizers Help >> Programming >> Reference >> NI-SCOPE LabVIEW Reference Help >> VIs >> Configuration

[Shiying]

Thank you, Joe.

I did some experiments as below:

Experiment #1:

1) configure channel characteristics - max input frequency = 20 MHz

2) configure 200MHz sample rate

3) AC coupling (I can't set DC offset due to various DC output)

4) set immediate trigger

result: p-p = 840mv (same reading as without adding 20MHz max input frequency).  There are many very sharp (high frequency) noise can be observed in the fetched waveform.  These noises are too narrow in time (X axis) to be detected with 200MHz sample rate.  At the same time, these sharp noises repeat themselves in low frequency (<20MHz).  It maybe why the filter seems not working.

I setup to compared with Tektronix digital scope (1.25GHz sample rate) at the same time.  Tektronix reading is 840mv at full bandwidth and 420mv at 20MHz bandwidth.  So it looks like the built-in 20 MHz noise filter in PXI 5124 didn't work in our application.

Experiment #2:

Since the max real time sample rate in PXI 5124 is 200MHz, I tried to setup RIS (up to 4GHz sample rate).  Unfortunately, RIS doesn't work with immediate trigger.  I got error message from the scope.  But I can't use other trigger type.  So I had to give up.

Experiment #3:

Add a Butterworth Filter.vi with sampling frequency = 1.25GHz (compare with Tektronix) after the acquisition (Fetch data).

with the order = 2, p-p reading is 420mv (mach the Tektronix)

with the order = 4, p-p reading is 260mv

with the order = 8, p-p reading is 190mv

Also I found out that the higher sampling frequency was used , the lower p-p reading was obtained.

I need advise to select the sampling frequency and the filter order in the Butterworth Filter.vi.

Shiying

[JoeFriedchicken]

Hi Shiying,

I'm concerned about your scope not appearing to filter at its input. If you would like to investigate this further, please attach your test VI and any screenshots that show the unexpected behavior.

If you're facing a time deadline and want to move forward, you can filter your data in software using the Butterworth Filter VI. The sample rate you should pass to the VI is the sample rate at which the data was collected. The order you should use is the one that best filters your data for your needs. A second-order filter is generally good and will be faster to perform in software than higher order filters.

[Shiying]

Hi Joe,

I will discuss with you about the scope issue later or talk to our local NI rep..  Meanwhile I have to move forward.  I am going to use Butterworth Filter with the order of 2 (I am looking into the system requirement to find out what the best curve fit is for our application) with 1.25GHz sampling frequency.  I can't use 200MHz although the real sample rate of my scope is 200MHz because it didn't attenuat p-p value at high frequency signal.

Shiying