Contact an Expert

Thread: Understanding VNA specifications.


Permlink Replies: 14 - Pages: 1 - Last Post: Dec 5, 2012 11:05 AM Last Post By: SOLT_guy
SOLT_guy

Posts: 268
Registered: 05/30/09
Understanding VNA specifications.
Posted: May 20, 2012 2:04 PM
Click to report abuse...   Click to reply to this thread Reply
This is a general question about understanding VNA specs.

I own a 8753C (but I have to use other network analyzers for my job) and I have to correspond with customers and management with respect to the accuracy of the network analyzer measurements. Therefore, my replies have to be exact and cannot be "generally" correct because these measurements may be associated with calibration laboratory tests or some other application with exacting standards.

If I look at the receiver instrument specifications under the section, "Phase Characteristics" - Frequency Response Deviation from Linear, there is a spec + or minus 3 degrees from 300KHz to 3 GHz. Under this spec. I see a graph.

I need to interpret this graph with regard to what can be formally asserted with respect to actual phase angle measurements.

From the graph I can see that there is a 2 degree deviation from linear at 6 GHz (see page 19 - 8753C instrument specifications section of the user manual) and about one degree deviation at 3 GHz.

Let's say I ran a performance test and my performance test data output was exactly the same as the graph.

Can I tell a customer who wants to know the accuracy of my phase measurements that my phase measurement is within plus or minus 2 degrees at 6 GHz and one degree at 3 GHz (data from a test)?

Or am I only entitled to say that Agilent specifies its NA phase measurements at plus or minus 3 degrees (that is, if this interpretation of "linear phase characteristics" is equivalent to the phase angle "accuracy" measure of a phase measurement).

Is the term "linear phase characteristics" equivalent to the phase angle "accuracy" measure?

Please help me define exactly what these specifications are really saying.

Thank you for your past replies.

Edited by: SOLT_guy on May 21, 2012 4:11 AM

Edited by: SOLT_guy on May 21, 2012 4:13 AM

Edited by: SOLT_guy on May 21, 2012 3:54 PM

Edited by: SOLT_guy on May 21, 2012 3:57 PM

Edited by: SOLT_guy on May 30, 2012 1:07 AM
Dr_joel


Posts: 2,745
Registered: 12/01/05
Re: Understanding VNA specifications.
Posted: May 30, 2012 11:28 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
SOLT_guy wrote:
This is a general question about understanding VNA specs.

I own a 8753C (but I have to use other network analyzers for my job) and I have to correspond with customers and management with respect to the accuracy of the network analyzer measurements. Therefore, my replies have to be exact and cannot be "generally" correct because these measurements may be associated with calibration laboratory tests or some other application with exacting standards.

If I look at the receiver instrument specifications under the section, "Phase Characteristics" - Frequency Response Deviation from Linear, there is a spec + or minus 3 degrees from 300KHz to 3 GHz. Under this spec. I see a graph.

If I recollect correctly, this is for raw measurements. So it is unrelated to the results you get with calibrated measurements
I need to interpret this graph with regard to what can be formally asserted with respect to actual phase angle measurements.

From the graph I can see that there is a 2 degree deviation from linear at 6 GHz (see page 19 - 8753C instrument specifications section of the user manual) and about one degree deviation at 3 GHz.

Let's say I ran a performance test and my performance test data output was exactly the same as the graph.

Can I tell a customer who wants to know the accuracy of my phase measurements that my phase measurement is within plus or minus 2 degrees at 6 GHz and one degree at 3 GHz (data from a test)?

You must add the uncertainty or error of the test method.
Or am I only entitled to say that Agilent specifies its NA phase measurements at plus or minus 3 degrees (that is, if this interpretation of "linear phase characteristics" is equivalent to the phase angle "accuracy" measure of a phase measurement).
Again, this is uncorrected response.
Is the term "linear phase characteristics" equivalent to the phase angle "accuracy" measure?
No, it is exactly what it says: it is the deviation from linear phase. But phase accuracy should include the linear phase effects. Adding a length of line to your test set port will not appreciable change the deviation from linear phase (it will and substantially only linear phase, that is phase that changes directly with frequency), but will certainly change the absolute phase number.
Please help me define exactly what these specifications are really saying.

Thank you for your past replies.

Edited by: SOLT_guy on May 21, 2012 4:11 AM

Edited by: SOLT_guy on May 21, 2012 4:13 AM

Edited by: SOLT_guy on May 21, 2012 3:54 PM

Edited by: SOLT_guy on May 21, 2012 3:57 PM

Edited by: SOLT_guy on May 30, 2012 1:07 AM

SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Understanding VNA specifications.
Posted: Jun 3, 2012 9:35 AM   in response to: Dr_joel in response to: Dr_joel
Click to report abuse...   Click to reply to this thread Reply
Attachment Bird Article.pdf (759.4 KB)
Dear Dr. Dunsmore;

Thank you for your reply, it was most enlightening.

With regard to the measurement of phase angle accuracy, I read a two page article (published by Bird corporation), entitled, "VSWR: Voltage Standing Ratio." The article says on page 2, second paragraph down:

"The slotted line is still the most accurate instrument, yielding both VSWR and phase angle information"

I have uploaded the article so that you may verify its authenticity. The article does not appear to be dated.

My question is, was an independent study ever published which verified that the modern network analyzer makes more accurate phase measurements than the slotted line? I realize that the "workmanship" of each particular unit may determine its actual measurement accuracy, but I am referring to a general case whereby the units under test are random samples ("off the shelf") that are identified by the manufacturer and model number.

I know that an independent study was conducted comparing dedicated TDR units vs. NA TDR units and it was implied that the NA TDR units output less uncertainty in their data measurements than dedicated TDR units under specific conditions ("Comparision of VNA and TDR Measurement Uncertainty Using Coaxial Cables," by Paul Pino, W.L. Gore & Associates Inc., Landenberg, PA - Cables & Connectors 2010 A Special Supplement to Microwave Journal).

Was a study ever presented that compared slotted line phase angle accuracy measurements as compared to the phase angle measurements using a, relatively, modern network analyzer?

Edited by: SOLT_guy on Jun 3, 2012 9:36 AM
Dr_joel


Posts: 2,745
Registered: 12/01/05
Re: Understanding VNA specifications.
Posted: Jun 5, 2012 6:06 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
I do not know of any study of the accuracy of slotted lines. I doubt that anyone has seriously used slotted lines as measurements for more than 30 years, since the invention of modern error correction techniques. Slotted lines have many areas that are open to error including probe variation along the line and detector issues, which I would suppose are very difficult to analyze and create error bounds for.
SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Understanding VNA specifications.
Posted: Jun 5, 2012 6:34 PM   in response to: Dr_joel in response to: Dr_joel
Click to report abuse...   Click to reply to this thread Reply
Dear Dr. Dunsmore;

Thank you very much for your feedback.

I would tend to agree with your opinions regarding the inherent potential for errors which may arise from probes, detectors, and other devices of an analog nature. But, I do believe that becoming too dependent upon digitized interpretations of data has its pitfalls. Analog devices may not be perfectly accurate but their output as a "ballpark" figure is very reliable.

I was hoping that someone out there may want to contribute to our discussion.

I do believe a comparision had to have been done at some time, unfortunately, its conclusions may only be relevant for the time frame that it was published.

It is probable that a time frame existed when the network analyzer was in its developmental stage while the "slotted line" was probably a "mature" technology. I would still like to read the findings of the article, or the study, with regard to these two measurement devices.

Again, thank you for your feedback and, hopefully, someone will contribute to this discussion.
jwl

Posts: 2
Registered: 06/05/12
Re: Understanding VNA specifications.
Posted: Jun 5, 2012 9:54 PM   in response to: Dr_joel in response to: Dr_joel
Click to report abuse...   Click to reply to this thread Reply
On the contrary, I have used a slotted line for measurements at 100 GHz within the last 12 years. However, I would not recommend it as it is very tedious and requires a lot of attention to detail. The probe is usually coupled to a direct detector which generally does not have the sensitivity of a heterodyne system (even when a chopped/synchronously detected scheme is used). There is a compromise between sensitivity and distortion of the standing wave pattern when you adjust the probe depth, and that depends on the magnitude of the VSWR. You need to do measurements at several points along the slotted line so you can extrapolate to the end of the line, etc. And all that gets you is one frequency point.

I have used this method at ~30 GHz and ~100 GHz to do things like measure horn reflection coefficients and the dielectric constant of epoxy, but tedious is the word that springs to mind. I do not think that I ever achieved anything like the accuracy possible with a modern (or even not-so-modern) VNA. The only reason, in my mind, for using a slotted line is if you have one and you can't get your hands on a VNA ;-)

The best (actually, only) reference I know of is the old Rad. Lab. series from the 1940's and 50's (probably Marcuvitz) ...
SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Understanding VNA specifications.
Posted: Jun 13, 2012 4:18 AM   in response to: jwl in response to: jwl
Click to report abuse...   Click to reply to this thread Reply
Dear JWL;

Thank you for your reply. I tend to wait some time before I reply so that everyone who is interested in the topic can reply to the last message. I got really busy and I couldn't reply to this message until now. I hope that you are still able to read this reply. Sorry for the belated reply.

I tend to believe that although the slotted line may be inconvenient to operate, I tend to believe that one can still learn from the data that this device outputs.

The experimental data outputs may be associated with a mathematical analysis which has yet to be published, or, at least popularized. What are your thoughts on this topic?
Dr_joel


Posts: 2,745
Registered: 12/01/05
Re: Understanding VNA specifications.
Posted: Jun 13, 2012 12:09 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
I would guess everything of interest on slotted lines was published 50 years ago.
dhamilton

Posts: 144
Registered: 02/01/08
Re: Understanding VNA specifications.
Posted: Jun 15, 2012 12:33 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
SOLT_guy,

Your best source for documentation would be some NIST publications concerning the measurement analysis in this discipline. You could find the information you need to concerning Error correcting VNA's vs the slotted line technique. I believe one of the first ones published was about 1986 or 1987 shortly after the 8510 series was introduced. I believe the also had several 8409 systems which was an automatic 8410 system which took up three rack systems to work from 100 MHz - 18 GHz. I had used this system, an it took a lot of effort to get it to have repeatable measurements. You had to be a very skilled metrologist. I know we had gotten .005 dB/10 dB repeatablity in magnitued and I believe +/- 2 deg to 12.4 GHz and +/- 3 deg @ 18 GHz. We had compared measurements with a VM7 system and NIST artifacts to ensure we had quality results.
SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Understanding VNA specifications.
Posted: Jun 24, 2012 2:54 PM   in response to: jwl in response to: jwl
Click to report abuse...   Click to reply to this thread Reply
Dear JWL;

Thank you for your reply.

I am actually thinking about purchasing a working slotted line system for academic reasons. Write back to me privately if you plan on selling your working system.
SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Understanding VNA specifications.
Posted: Jun 24, 2012 2:58 PM   in response to: dhamilton in response to: dhamilton
Click to report abuse...   Click to reply to this thread Reply
Dear DHamilton;

Thank you for your reply and I must apologize for my belated reply. I get so busy sometimes that I can't log onto this web site for weeks.

One quick question:

On average, with regard to your experimental comparision of phase angle data from slotted lines vs. network analyzer phase angle measurements, as compared to the NIST loads, can you tell me which data values were more consistent, slotted line or network analyzer?

Thank you again for your input.
dhamilton

Posts: 144
Registered: 02/01/08
Re: Understanding VNA specifications.
Posted: Jun 24, 2012 4:01 PM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
For most users the error correcting VNA will give the most repeatable results. A skilled metrologist in using slotted lines can get similar results. However, as JWL stated it takes a lot of practice to get repeatable results. Once you are able to do that, then you can rely on your measurements being correct. It is good that you are willing to demonstrate to new students this technique as so they may understand the measurement technique and how it is accomplished. I know when I started this field you had to demonstrate you skills manually before you could even touch an automated system. This way you can better determine the source of poor readings, then resolve them.
SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Understanding VNA specifications.
Posted: Jun 25, 2012 6:00 PM   in response to: dhamilton in response to: dhamilton
Click to report abuse...   Click to reply to this thread Reply
Dear DHamilton;

Thank you for your reply.

"Repeatibility" is not the highest priority for me. I am seeking "accuracy" for my measurements. The NA has a lot of features such as averaging, smoothing, etc.. which output good data but I want a test instrument that will return an "accurate" measurement even if operator skill, attention to detail, and more time are required.

I was curious as to which test instrument output data "closest" to the NIST specified phase angles for a given load.

Also, it is my belief that it may be possible that the slotted line may output some relevant physical information which may not be obvious when using a network analyzer. This belief is only hypothetical. Since I do not own a slotted line, I will not know the answer to this question until I purchase one. Maybe you can give me a hint about what I am missing. Hopefully, JWL will someday sell his system at a reasonable price and I will be lucky enough to see his ad and purchase it.

There are so many slotted line parts that you can buy on e-Bay but they are all worthless unless you obtain every part. "Cobbling" together working systems can be such a waste of time and money. I have learned from experience, believe me.

It was nice to hear from you DHamiliton, hopefully I will hear from you again.
Metatron

Posts: 11
Registered: 05/02/12
Re: Understanding VNA specifications.
Posted: Dec 4, 2012 8:21 AM   in response to: SOLT_guy in response to: SOLT_guy
Click to report abuse...   Click to reply to this thread Reply
Interesting topic. I use to teach the use of slotted lines long ago. They were very cumbersome to use and were most accurate when measuring a single frequency point. As accurate as they were, they are no match for a VNA in terms of accuracy or repeatability. The VNA's major errors are corrected by using a calibration kit. The slotted line errors included the mating of the various waveguide flanges and adapters, the depth setting of the detector probe, the linearity accuracy of the detector in the square-law region, the tuning of the 415E SWR meter to the 1kHz modulation frequency, the 415E attenuator and meter accuracy, the skill of the technician to use the tuning stubs, and the quality of the termination at the end of the line itself.

It was eventually replaced for a more practical approach of using a signal generator, and 415E for swept-frequency measurements. By using a rotary vane attenuator for RF Substitution, or a 415E and crystal detector for AF Substitution, and a 415E and Piston attenuator for IF Substitution. Each method had its advantages and disadvantages, but all were quicker, cheaper, and simpler to use.

Then came the VNA, a very expensive approach, but one that is more accurate and quicker than all the other methods.
SOLT_guy

Posts: 268
Registered: 05/30/09
Re: Understanding VNA specifications.
Posted: Dec 5, 2012 11:04 AM   in response to: Metatron in response to: Metatron
Click to report abuse...   Click to reply to this thread Reply
At the time I wrote this post I didn't own a slotted line.

I bought a couple of slotted lines since then (3 total). I really have to say, I like the slotted line. It is much easier to understand the effects of high frequency signals upon a load with a slotted line. Further, although VNA's may be more accurate (that is, if you are working with a VNA that is in spec.), it is far more difficult to detect errors in measurement with a VNA. I like reliable measurements and the slotted line offers the user the ability to observe where the correct phase angle locus must lie. This feature can't be beat.

The problem with buying slotted lines now is that you can't get parts. People on ebay are selling junk that isnt' working and, further, the sellers don't even know how to determine if the slotted line can even work or not. Therefore, noobies are really stuck when they make their first purchase. Also, old mechanical test instruments can break easily and no one is qualified to repair the repair the test instrument properly- at least, I couldn't find one. I know all this from experience.

Although I went through 2 broken slottted lines, I found it was well worth the "learning" investment.

I did not like the HP 415E system. I found it to be an unreliable system judging it by modern standards. I am certain in its heyday it was the talk of the town, but not anymore. I actually bought one on EBay but it was unable to make measurements on receipt. So I returned it. I look through the manual and did not care for how it was written. By the way, have you ever calibrated the HP 415E?

Edited by: SOLT_guy on Dec 5, 2012 11:04 AM

Point your RSS reader here for a feed of the latest messages in all forums