RF Synchronization Suite

[Andy_Hinde]

Hi Betty,

The RF multi-channel reference software does indeed perform repeated single-shot acquisitions. The example VI I attached to my previous post performs continuous acquisition to illustrate the point, but the reference software performs distinct single-shot acquisitions. The Record option for the reference software does perform continuous acquisition, but streams the data to disk with no in-line analysis or calculations being done, as this woudl slow the streaming process.

This is because when I choose CW in the generation front panel, I get changing horizontal lines for the I and Q vs t waveforms. On the other hand, I get 2 points (since I deal with 2 channels) which both move in circles for the I vs Q plot.

Exactly - the single-channel phase of each channel changes from one-acquisition to the next since they start at a different point relative to the LO/reference clocks, but the phase delta vs time plots will show the phase difference between these points as a function of time, which wil be constant.

May I also ask if you have any suggestions on how I can fix my I and Q vs t waveforms when I choose Arb in the generation front panel. The resulting I and Q vs t waveforms are changing sinusoids which I thought to indicate frequency offset between the generator and analyzer.

So there may be some confusion here, which I'll take as feedback for better documentation of the generation reference software front panel. The 'Arb' mode selection when using the generation reference software is intended to show how to program the generators when something other than a sine wave (CW) signal at the center frequency is desired. For this purpose, there is an option for double sideband, upper sideband, or lower sideband. Double sideband generates a multi-tone signal with a sine wave above and a sine wave below the carrier frequency. Upper sideband produces just the upper tone, etc... So, since you're generating a tone/tones that are not centered at the center frequency, it is the same as simulating a frequency offset and the results you're seeing are expected. Keep in mind that these arb waveforms could be anything, I simply copied the NI-RFSG shipping example called RFSG Arbitrary Waveform Generation for my chouce of arbitrary waveforms.

You can change this if you want - go to case "(25) Configure Generators", and the sub case "Arb" in the block diagram. You'll see sine patterns used as the I and Q waveforms written to generator onboard memory. If you wanted a sine wave at the center frequency for each channel, you could instead write an array of 1's for the I waveform and an array of 0's for the Q waveform (instead of the sine patterns). This would cause the "Arb" mode of the generator reference software to act just like the "CW" mode.

So to summarize, what you're seeing is expected behavior for the "Arb" mode and you can change the waveform types in the "Arb" mode if you like. And the next version of this reference software will better document the waveforms being generated by the "Arb" mode.

Regards,

Andy Hinde

RF Systems Engineer

National Instruments

[Betty Boop]

Thank you very much! 😄

Since I want to generate sine wave at the center frequency for each channel, I followed your advice:

          You can change this if you want - go to case "(25) Configure Generators", and the sub case "Arb" in the block diagram. You'll see sine patterns used as the I and Q waveforms written 

          to generator onboard memory. If you wanted a sine wave at the center frequency for each channel, you could instead write an array of 1's for the I waveform and an array of 0's for the Q 

          waveform (instead of the sine patterns). This would cause the "Arb" mode of the generator reference software to act just like the "CW" mode.

Attached is a screen shot of the revised code following your suggestion. However, whenever I try to change the amplitude of I and Q (by changing the array elements), I don't see any noticeable change in the I and Q vs t plots. Why is this so, Sir? 🙂

On the other hand, when I use the CW generation mode, the phase delta is not constant as seen in the attached file from my previous reply. I understand that I'm seeing expected behavior - changing horizontal lines - in the I and Q vs t plots since the demo program is configured for distinct single-shot acquisitions. Like what I said in my last reply, I already had my generator and analyzer both locked to the same reference clock source (they are in the same chassis and both set to PXI_CLK reference clock source in the front panel of generation and analyzer VIs) so that there is no frequency offset. Do you have any suggestions on how I can fix the phase delta to make it constant?

Thank you very much for your time! 😄 

[Betty Boop]

Hello!

This is about my second question in my last reply.

Attached is a pdf containing the results I get in using the CW mode/Arb mode (which now functions like the CW mode after replacing the sine patters with an array of 1's for the I waveform and an array of 0's for the Q waveform in case "(25) Configure Generators" sub case "Arb").

I would just like to ask why I get inconsistent results. For example, the phase delta is only constant for some frequency. I can’t see a pattern from the obtained results, that’s why I’m having a hard time in trying to fix the phase delta to make it constant in 915MHz which is the frequency I’m working on.

Also, notice my comment in the attached file. What I mean by “overtakes the channel” is that one channel is moving faster compared to the other channel (I deal with 2 channels only) as seen in the IQ plot and phase vs t plot. Just like the phase delta, the behavior of the channels in the mentioned plots is dependent on the frequency I set.

The settings in the front panel are the default ones except for the frequency.

Thank you so much for your time! 😄

[Betty Boop]

Hello!

I would just like to correct the 'PHASE VS TIME' description of the 1GHz. It should be '2 CHANGING HORIZONTAL LINES', not 2 circling points. However, one does not overtake the other horizontal line like in the 915MHz.

Thank you! 🙂

[Betty Boop]

I need help please 😞

[Eric_L]

Hello Betty Boop,

Unfortunately Andy has not responded to your post but we will be able to help you with your questions. I have looked over the post in this tread. Would you be able to post a copy of the VI that you are using? Also, would you be able to post some screenshots of the behavior you are currently seeing?

Eric

[Betty Boop]

Hello!

Thank you for responding. 🙂 The results I am currently getting are summarized in the attached file. From what I learned from Andy, the “correct” plots are those under the 1GHz. However, my application uses 915MHz frequency. As seen in the attached file, setting the frequency to 915MHz results to “incorrect” plots (compared to those when 1GHz is used). What I mean by “overtake” is that one channel moves faster than the other channel. For example, for the IQ plot, there are 2 points representing the two channels. If the slower channel travels 1 complete circle (since the 2 points travel in circular path), the faster channel is able to travel more than 1 circle. The “space” between the 2 channels in the plots is not constant since one channel moves faster. More importantly, I am not getting a constant phase delta which is really problematic.

Also attached are screenshots of the IQ, Phase vs time and Phase delta vs time plots I’m getting. However, since these are static, we can’t see the actual overtaking of one channel.

I'm using the NI RF Phase-Coherent Multi-Channel suggested by Andy in post no. 2 of this thread.

Thank you very much for your time! 😄

[Andy_Hinde · Accepted solution]

Hi Betty,

Sorry for my absence recently, I have been travelling a lot recently.

First and foremost, I recently posted an updated version of the RF Multi-Channel and RF Phase-Coherent Multi-Channel reference architectures on the NI RF Multi-Channel site on NI Community. Here is a link to the (currently) latest version 1.1 of the phase-coherent multi-channel software:

https://decibel.ni.com/content/docs/DOC-19433

Next, in the hopes of communicating more efficiently, I've recorded a demo of a four channel phase-coherent generation system transmitting into a four channel phase-coherent acquisition system (4x4 MIMO). Each Tx xhannel is cabled to a Rx channel individually. I could also have used a single signal with a 4 Way 0 degree splitter (each output of the splitter cabled to a Rx channel), the only difference would be some insertion loss from the splitter, and that the phase deltas for the three slaves relative to the master in the acquisition VI would initally have been all very close to zero degrees, dependent on the phase tracking specs of the splitter.

In any case, the goal of this video is to demonstrate what I see and what is expected behavior for this software and hardware setup. If you could view this video and listen to the audio, and then let us know if what you are seeing matches or is different from this, that would be most helpful.

Video link:

http://screencast.com/t/hvHVSJoe

Regards,

Andy Hinde

RF Systems Engineer

National Instruments

[Betty Boop]

Hello, Sir Andy! 😄

The latest version which is the NI RF Phase-Coherent Multi-Channel 1.1 worked perfectly.
Thank you very much for all the help! 😄

For the record, the link below contains a video of the resulting plots when I used the older version of the software.
This is set to 915MHz frequency.
Observe how one channel "overtakes" the other channel which is visible in the I vs Q and Phase vs time plots.

http://www.youtube.com/watch?v=gmQ8CkBztpw&feature=youtu.be

Anyway, the latest version is working properly regardless of the frequency I set.
Thank you very much for your time! 😄
I learned a lot from this thread! 😄
God Bless! 😄

[Andy_Hinde]

Hi Betty,

I am glad to hear this is resolved for you! Enjoy!

Regards,

Andy Hinde

RF Systems Engineer

National Instruments