Generating waveform with trigger using 5671

[cxl6288]

Hi, all

I am using 5671 to generate a GFSK signal with a trigger. The script for the generation is as following:

script trigGFSK
   repeat forever
      generate GFSK marker0(0)
   end repeat
end script

The marker is output through PFI0 as a trigger to syn the RFSA.

The length in time of GFSK signal is 128us. Thus the trigger should repeat every 126us. However, the measured trigger period is about 10-15ms. This long trigger period make our BER test very slow.

Is this caused by software or hardware? Thanks for the help!

Chen

[Robert_F]

Hello Chen,

How many elements are in your waveform and what is the actual IQ Rate?  There is a great shipping example that does exactly what you want. 

(Help >> Find Examples >> Hardware Input & Output >> Modular Instruments >> NI-RFSG >> Scripts >> Marker Events.vi) 

This example has 64 elements in the array with an actual IQ Rate of 100MS/s which means that the markers should occur every 0.64us.  Run this example and make sure that the markers are distributed correctly. 

[cxl6288]

Hi, Robert,

I run the example. In the example, the IQ rate is 1MS/s with 64 elements. The measures trigger period is 64 us. This is correct.

In my case, the IQ rate is 8.3333MS/s with 1152 elements. Thus the trigger period should be 1152/8.33333M=138us. However the measured period is 10ms. We noted this problem because the BER test is mush slow.

Thank you,

Chen

[Robert_F]

Hello Chen,

I was testing out the same example with an IQ rate of 8.33MS/s and 1152 elements.  I connected PFI0 to a 5142 digitizer and used the NI Soft Front Panel (SFP) to measure the markers.  I was getting very weird results just like you were explaining.  Sometimes 2.37ms and 17.31ms. 

However after playing with time/div on the SFP, I was able to measure the marker correctly.  The time/div and record length is actually going to determine the sampling rate of the digitizer.  Since the record length is held constant as the time/div was adjusted, it determined the sampling rate.  This could definitely explain the weird dt between the markers. 

The markers were occurring in between when the digitizer was sampling.  This is why it appears that the dt between the markers are not correct.  When I changed the time/div to 50us, it allowed the digitizer to sample the PFI line fast enough to account for the correct dt. 

Here is a screenshot of the SFP and the shipping example.

 

If you still unable to get the marker dt to be correct, can you post a screenshot of your VI and oscilloscope. 

Message Edited by Robert F on 01-29-2008 11:13 AM

[cxl6288]

Hi, Robert,

I run NI example with an IQ rate of 8.33MS/s and 1152 elements, and get correct marker dt (138us). I measured the marker using a Agilent digital scope.

In my VI, the marker dt is varied from 10-15 ms for different data pattern.

Attachment is the VI we are using for GFSK generation. Run the VI with the default setup, the marker dt should be 238us. The measured dt is 750us. The only difference in our application of this VI is that, we use a special data pattern of size 36 bit, rather than a PN sequence.

Thank you,

Chen

[Robert_F]

Hello Chen,

Set the repeat waveform until tolerance is met to false.  This VI is making sure that your signal is phase continuous however the 5142 should be doing this for you.  Once you set this value to false your marker dt should be correct.

Message Edited by Robert F on 01-30-2008 06:18 PM

[cxl6288]

Hi, Robert,

It works. I get correct trigger dt.  Thank you so much.

Chen