60Hz Noise

[Yardov]

Hi Brandon,

Have you tried to swap your 5620? In my 5660 I see 60 Hz peaks at about -85 dBc. This is with the most simple configuration: CW source » 5660 demo panel.

Try not to lock any clocks, try to do the test as simple as possible.

Remember you can contact National Instruments to have your board repaired (in case needed).

[cochenob]

I am using the "Generate Single Tone" VI as well.  I don't think I got as good as -85dBc though...

Unfortunately...for my application, I desire to make some relative phase measurements from I and Q samples...so locking the two together is necessary.

As I've mentioned...I'm not sure whether or not the board/chassis needs repaired...there's noise in my system...or I'm just looking at the fundamental limit of the devices.

I appreciate your insight...thanks!

---

BC

[Yardov]

Hi Brandon,

a) Can you define if you tried the simple test setup? Probably the noise is being introduced by the chassis and switching the module to the other chassis as suggested before will help you define that.

b) If you are looking at phase measurements, you should know that the internal LOs of the RF devices (5600 and 5610) cannot be locked together. This thread speaks little more about that same problem.

[cochenob]

In regards to question a)....please see some of my previous posts in this thread.

In regards to question b)...my apologies here.  I wasn't specific in my verbage.  Locking the up and downconverters removes the frequency offset, not the phase offset.  Locking the two produces a constant, but non-zero, phase offset.  Regardless...if I measure what this constant, but non-zero, phase offset is without a DUT, and then the phase offset with the DUT, I should be able to measure the relative phase added by the DUT.  This is what is described in one of my previous topics you linked to above.

The problem I'm seeing is that the constant, but non-zero, phase offset isn't exactly constant....there's a good deal of noise on it.  Again...see my previous posts in this topic for a more detailed description.

[Yardov]

Hello Brandon,

Sorry for asking things you stated before,
Have you been able to see the 567x signal in a spectrum analizer (or scope) not linked to the PXI chassis to see if this artifact is present? Basically I want to confirm that the artifact is comming from the generation side or it being introduced by the acquisition side.

Assuming is on the acquisition side, it is highly unlikelly that both modules are bad. Therefore you can open the example:
C:\Program Files\National Instruments\LabVIEW 8.2\examples\Spectral Measurements Toolset\SMT Examples for niScope\Interactive Examples

You can get to them from:
Start->Programs->National Instruments->Spectral Measurements->LabVIEW Support->Spectral Measurements Examples Folder

called SMT Configured Spectrum Example for niScope.vi

Inside this example you can configure the center frequency, span and resolution. If you could then generate a sine wave signal (2V pk-pk) and send this to the digitizer (5620). This application will display the spectrum and probably display the same noise, however it will guarantee that the problem is with the 5620.

Finally, perform an acquisition of a 50 ohm termination so that we have noise across the entire frequency bands. Do you see anything unusual?

Note: Each test you perform, try to make sure no other unused modules are on the chassis. This way we can be sure to isolate them from the system.N

[Yardov]

Hi Brandon,

Any news there?
Best,

[cochenob]

Been tied up with some other projects this week, so I haven't yet tried all of these suggestions.

I have tried to power the chassis from an isolated supply, and saw the same noises...so it would appear that it's something internal to the chassis.

A week or so ago I looked at the 567x into both a scope and a seperate spec analyzer.  I saw a little bit of 60Hz noise around my carrier on sthe spec analyzer, but it seemed sufficiently down in the noise enough.  (At least in comparison to the spectrum I saw when I acquired with the 5660 demo panel....see one of my attachments in a previous posting in this topic).  Of course...who knows how much the demod process may stand to promote that little noise I see on the source side.

I haven't yet tried to look at an outside source into the 5660 downconverter, which would be a good to test.  As would your suggestion about by-passing the downconverter and going directly into the digitizer (at least that's what it sounds like your suggesting).

I have looked at the entire spectrum with a load on the input...and nothing looks out of the ordinary.

I'd be interested to see if anyone else using a 5660 and 567x has seen these problems (i.e. send a single tone into the downconverter, and narrow up the span around your tone to see what low freq noise is there.)  Likewise...if anyone has had problems with frequency locking the two devices together...and what noises are associated with doing that.

Stand by for more results!!!  Thanks again for your help.

---
BC

[Yardov]

Hi Brandon,

------
I haven't yet tried to look at an outside source into the 5660 downconverter, which would be a good to test.  As would your suggestion about by-passing the downconverter and going directly into the digitizer (at least that's what it sounds like your suggesting).
------

Yes correct. I am starting to think that the issue might be with the acquisition side.
I look forward,