AnsweredAssumed Answered

VNA insertion phase measurment for cable assembly

Question asked by shan on Apr 21, 2008
Latest reply on Apr 30, 2008 by shan

Like • Show 0 Likes0
Comment • 10

Hi I m a new member. I m a testing technician and I use E8801A model Agilent PNA.
Can anyone please explain how to do the phase measurment and delay using this VNA in order to phase match cable assembly in a group of cable assemblies.

I briefly explains what I did.
1.Start Freq 360 MHz Stop 380 MHz
2.VNA was set to do S11,S21,S22, and S21 phase measurment.
3.Measurment: S21, Format: Phase, Marker 370 MHz
4.Did full 2 port Calibration and connect DUT (Cable).
5.Displayed SAW tooth form graph.
6.On electrical delay menu following settings were displayed
    Delay =0ns
    Velocity factor = 1 (actual cable has VF 0.77)
    Cuttoff frequecy 450000Hz
    Coax was ticked
    Wave guide was unticked
7.Phase offset menu setting: 0

8. selected marker function menu and select delay.
9. A flat line was now oberserved with phase value in degrees. What will be the explanation of this value oberserved. Why it changes every time I select marker function menu and press delay tab.

9. Now electrical delay menu displayed 6.5ns.

How do I use or interpret this graph to phase match cable assemblies to each other within +/-2.5 degress.
Looking forward to anyone's reply with detailed explanation
Thanking you

No one else has this question

Outcomes

10 Replies

odanzy Apr 21, 2008 4:14 AM
Mark Correct
Correct Answer
You might try this process:

[list=1:gyl7682g]
Calibrate and measure the "golden cable"

With the phase display shown, push the current data into Memory (Data->Memory)

Set the trace math function to "Data/Memory"

Connect the "test cable" and you will now see the difference between the golden cable and test cable. This display will show you the difference in degrees between the two cables.

http://na.tm.agilent.com/pna/help/WebHe ... ations.htm

Actions

shan Apr 21, 2008 11:26 PM

Thanks a lot Mr Danzy for you time in replying my question.

Could you pls explain more about this.

1. I performed what you told exactly.
2. Choose one Golden cable
3. Calibrate analyser
4. DUT (Cable) connected.
5. Observed a value -56 deg at 380MHz for S21 Phase measurement
6. Open Math menu box
7. Push Data>>Mem
8. didn't not tick 8510 check box
9. Data/Mem option was selected.
10. Data,mem trace was selected. to display both traces
What format does the calculation is being done. :?:
11.Now display reads 102 deg at 380MHz for data plot and -56 deg at 380MHz for memory plot. :?
12 My understanding it should be zero or 1 as data plot is normalised now for the same cable. :?
13 Since the result is confusing for the same cable I didn't compare other cables to the golden cable.

Could you also pls explain how to measure the electrical length of the cable assembly using E8801A model PNA where correct VOP value is not known. :?:
Thanking you

Actions

odanzy

Apr 22, 2008 3:55 AM

First, only the data trace need be displayed as the memory trace will show the data that you pushed into memory. This is why the memory trace still read -56 degrees. For the current setup (as you have written) the data trace would be the active measurement/memory trace which corresponds to the difference in phase delay between the two cables.

Regarding your question on what data is used, it is the complex data for both data and memory traces that are used for the calculation.

What IFBW did you use for measurement and calibration? I recommend at least 1 kHz with step sweep enabled.

How long is your cable? You may have to increase the number of points to make sure that you are not aliasing. More on Measuring Electrically Long Devices.

After the data/memory are you still seeing phase wraps on the screen (on the data trace)? If you wiggle the cable do you see the data jump around everywhere?

Actions

Dr_joel

Apr 22, 2008 8:32 AM

"
Hi I m a new member. I m a testing technician and I use 8. selected marker function menu and select delay.
9. A flat line was now oberserved with phase value in degrees. What will be the explanation of this value oberserved. Why it changes every time I select marker function menu and press delay tab.

9. Now electrical delay menu displayed 6.5ns.

How do I use or interpret this graph to phase match cable assemblies to each other within +/-2.5 degress.
Looking forward to anyone's reply with detailed explanation
Thanking you
"

The Marker->delay function basically computes the group delay around a 20 percent span of the trace centered at the marker. That is, it uses the phase value from 10 percent above the marker minus the phase value for 10 percent below the marker, divided by the delta frequency (20 percent of the span frequency) and computes the delay, then does an electrical delay offset (mathematically) to flatten the phase trace about the marker position.

Each time you push the button, a new computation is made, and if the phase trace has any noise on it, a slightly different phase will be read above and below the marker and a slightly different electrical delay offset will be computed.

I recommend, instead of using a golden cable, to use the electrical delay offset value to normalize the trace to an "ideal" slope. Choose a value for electrical delay to flatten the phase response of the cable, and compare other cable values to this ideal curve. But, if your cables are very lossy, the phase will not be a straight line, so in that case, a golden cable may be a better choice. If you do a data/mem, you must set the delay value (under the scale menu) back to zero, because the d/m function also normalizes out the phase.

If you have more questions, it might be best to post a plot of your results. Use the "file, print to file" function and select the PNG format for the smallest file.

Actions

shan Apr 24, 2008 1:14 AM

Thanks for your reply Mr Danzy and Mr Dunsmore,

VNA was calibrated to full 2 port usinng 85052D.
Frequency 360 - 380 MHz
IFBW= 40KHz
201 points
Sweep time 27ms
Pls clarify these are correct setting :?:

I have attached the 2 files for your reference. These are observed form 2 cables.

File 1: refernce cable.
          data/mem plot shows -86 milli degress
          Pysical Cable length is 143.6 inch

File 2: Test cable.
          data/mem plot shows -12.5 degress
          Pysical Cable length is 147 inch

Last time I was confused with milli degree sign.
I can see the phase differnce correctly now. Thanks

Can you also please explain how to meaure the electrical length of the cable where VOP is 0.7 using this PNA. :?: . What will be the VNA setting?

Looking forward to your answer.
Thanking you.

Attachments

phase cable 1.png15.6 KB
phase cable 3.png15.5 KB

Actions

odanzy

Apr 24, 2008 4:10 AM

I recommend using a 1 kHz IFBW, that should clean up some of the noise that you are seeing with a 40 kHz IFBW.

Also based on a 12.5 degree difference in phase that should only equate to about 0.8 inch difference in cable length at 370 MHz using a 0.7 VOP, not the 3.4" that you are seeing. This means that while the approximate VOP is 0.7 there is some uncertainty in this number and one reason why phase is used vs exact length.

Actions

Dr_joel

Apr 24, 2008 7:29 AM

"
Thanks for your reply Mr Danzy and Mr Dunsmore,

Can you also please explain how to meaure the electrical length of the cable where VOP is 0.7 using this PNA. :?: . What will be the VNA setting?

Thanking you.
"

Very simple. Change the display format to "delay". It will show the electrical length directly, expressed as a time (e.g. 93 ps). Turn smoothing on to 20% for cleanest signal. convert to physical length by mulitplying time delay * 300e6 m/s to get length in meters. This is the electrical length. Multiply by VOP to get physical length.

Also, OJ is right to say use 1 kHz BW. Also, I would recommend using a wider span so that your measurement span of interest is less than 80 percent of the display width (smoothing works best in the middle 80 percent).

Actions

shan Apr 29, 2008 12:30 AM

Thanks for the feed backs.

Does model E8801A have the option to measure delay :?:
Because can't find format "delay" in the menu :!:
Can find only format "group delay".
Thanks.

Actions

odanzy

Apr 29, 2008 3:49 AM

Delay and Group Delay are the same format. Delay is the terminology that was used with many of the older analyzers (8753, 8720, 8510) probably due to screen real estate, where Group Delay is spelled out on the newer analyzers.

Actions

shan Apr 30, 2008 12:37 AM

Thanks.
One last question please.
Should I enter the VOP value 0.7 of the cable or dielectric constant of the cable anywhere. :?:

Actions

VNA insertion phase measurment for cable assembly

Outcomes

Attachments

Related Content

Recommended Content