AnsweredAssumed Answered

Phase locking multiple NA's

Question asked by kknox on Apr 23, 2007
Latest reply on Apr 24, 2007 by kknox

Like • Show 0 Likes0
Comment • 4

I'm setting up an far field antenna range, and have an option of several NA's (5062a, 8752a,8714). Trying (2) analyzers - 1 @ range antenna and 1 @ the AUT side.

The problem is that phase measured at the receive end 'drifts' with time. For example, it starts at 0degrees, and in 15mns will be 10degrees off. Have simulated in the lab so far using a common 10MHz source (disciplined source & also an 8341B synthesizer). I have tried various combinations of NA's as well as using the 8341B as an RF source - they all eventually drift.

Also varied the IF BW settings - no luck.

Any thoughts?

Thank you,

Kurt

No one else has this question

Outcomes

4 Replies

dgun Apr 23, 2007 11:02 AM
Mark Correct
Correct Answer
I'm not exactly sure of your setup, but locked, synthesized sources will have phase drift. It's inherent in the frac-n design. However, the frequencies should drift together, which is the primary purpose of locking.

10 degrees in 15 minutes seems high, but it depends on the frac-n used in the synthesizers. (I think the VNAs you mention all have frac-n synthesizers.)

The only way I know of to prevent this is to use the same LO fed to all the VNAs involved in the setup. If you're going to do that, you might as well use a single VNA with external coupler/receiver access.
Like • Show 0 Likes0
Actions
kknox Apr 23, 2007 11:29 AM
Mark Correct
Correct Answer
As far as setup, reflection port on source is directly connected to transmission port on 2nd analyzer, and a common 10mhz source was connected to ext ref on both analyzers. I tried combinations such as:

(2) 8752A's
(1) 8752A & (1) 5062A
(1) 5062A & (1) 8714A

Also tried the 8341B as RF source (RF out to transmission port on NA) and connected its 10mhz ext ref to that of the NA).

Our ultimate end goal is to use (2) gps disciplined clocks - 1 at each NA supplying the 10mhz ref signal locked to GPS, and then separate the NA's on the order of 1000ft. In the lab, the GPS units frequencies are matched - we can't discern a freq difference between them and the 8341B.

From what you stated, i will look into external coupler/receiver access options.

One of the design restraints was the actual physical obstacle of running a cable 1000' - not to mention loss..

Thank you,

Kurt
Like • Show 0 Likes0
Actions
Dr_joel Apr 23, 2007 1:08 PM
Mark Correct
Correct Answer
Hi, the problem you describe is well known to us. The trouble is, the 10 MHz locking locks the internal clock of the VNA to the 10 MHz input. But, the lock is through an analog circuit, so temerature can cause the output of the phase detector to change, causing the phase to drift. This can occur at the 10 MHz lock phase detector or at the frac-N phase detector. You will find that some units will be very stable after warmup, but I've seen 8340s drift for about 2 hours (It has 3 loops so lots of opportunities). On the other hand, the new PSGs are really really stable, like 5 degrees per hour of drift.

For many measuremnt scenarios, drift in phase is not a problem as you can normailze at the first point, take fast data, and not worry about the drift. You can get away with this using a reference path, which is sometimes used in antenna ranges for just such purposes.
Like • Show 0 Likes0
Actions
kknox Apr 24, 2007 9:26 AM
Mark Correct
Correct Answer
Out of the analyzers I have available, the 8752 has the aux input - but I don't think it is possible to get a reference path in there. From what I understand, that port is for measuring ac or dc voltage for display purposes. I imagine the way to tackle this problem, in an antenna range setup, is to use another antenna at the aut side for the reference path. Then offset the aut phase data with the reference horn phase data.

I would then need a 4-channel NA, correct?

Thank you for the help guys, i don't think i can spend another 8hrs staring at the phase drift - although it is somewhat soothing like ocean waves .

Kurt
Like • Show 0 Likes0
Actions

Phase locking multiple NA's

Outcomes

Related Content

Recommended Content