开云体育

On Sun, Dec 22, 2019 at 05:56 PM, Gary O'Neil wrote:

. I would expect this to be true if the standards used in both were set equal
to ideal.

Being mathematical identical the standards being NOT ideal have identical impact, even the uncertainty regions will be identical.
The formula I provided for comparison and is proven to be identical is the the standard 3 term calibration formula as also used in nanoVNA saver and TAPR-VNA.
I have not yet checked if this formula is also used (without simplification) in the nanoVNA itself. For sure the nanoVNA does not allow to use fully characterized standards and therefore the mathematics inside the nanoVNA is somewhat simpler.

--
NanoVNA Wiki: /g/nanovna-users/wiki/home
NanoVNA Files: /g/nanovna-users/files
Erik, PD0EK

GIN&PEZ;

Am I correct in my interpretation of this that incorporating standards with correction has - in this particular case - moved the measured result closer to the higher of the uncertainty boundary limits? The implication being that the use of nominal values for the standards potentially minimizes the degree of uncertainty in the final measurement.

Is it possible that there is a sign error in how the HP correction terms are coded? Alternating signs would tend toward convergence on a nominal value, rather than divergence in a signed direction.

--
73

Gary, N3GO

Thanks Jeff;

Too many new tools tied to learning about this. I don’t know which one to learn first. :-)

I’m seeing universal agreement on the equivalence of the computations, so I’m comfortable that I can set anchor for my understanding at this point.

I think the nuance that GIN&PEZ may be attempting to communicate; is the influence or perhaps lack of influence of calibration on the uncertainties in the final result.

They utilize what they define as differential error regions and intervals to predict the limits of uncertainty, and I think they are comparing measurements using nominal value estimates for the NanoVNA’s OEM standards with measurements made using high quality HP standards and their respective characterized correction values. Those results are also evaluated to assess their containment within the predicted error region.

--
73

Gary, N3GO

Erik;

The most recent posts by GIN&PEZ is shedding some new light on this. I am seeing additional data that leads me to believe that my understanding still remains somewhat off track. I also see hints of how the traditional approach may be different.

It raises a few more questions for me.

The calibration mathematics is making sense though, and I think I’m not too far from understanding the process in better detail.

--
73

Gary, N3GO

On Sun, Dec 22, 2019 at 11:32 AM, <erik@...> wrote:

In /g/nanovna-users/message/8494 they provide the proof they
are mathematically the same.

The GIN&PEZ equation certainly seems to give the same result as the standard one-port 3-term correction equations. Here's a quick Matlab example I wrote, using the GIN&PEZ terminology and values pulled out of the air for short, open, load, and measured DUT gamma:

MATLAB CODE

S = -0.9; % Actual Gamma of Short
O = 0.9; % Actual Gamma of Open
L = 0.1+1i*0.1; % Actual Gamma of Load

s = -0.8+1i*0.1; % measured Gamma of Short
o = 0.85-1i*0.15; % measured Gamma of Open
l = 0.2; % measured Gamma of Load

g = 0.5+1i*0.5 % measured Gamma of DUT

% G will be the actual (calculated) Gamma of the DUT.

% First, Gamma calculated according to GIN&PEZ
Gginpez = (S*(L-O)*(g*s+l*o)+L*(O-S)*(g*l+o*s)+O*(S-L)*(g*o+s*l))/...
((L-O)*(g*s+l*o)+ (O-S)*(g*l+o*s) + (S-L)*(g*o+s*l))

% Now, Gamma calculated per 3-term one-port correction formulas:
% (see equations 1, 5, 6, and 7 in the User Guide down-loadable from:
% )

% First, four preliminary equations...
D = S*(O-L)*s + O*L*o - O*L*l - (O*o - L*l)*S;

e00 = -((O*l-L*l)*S*o - (O*L*o - O*L*l - (L*o-O*l)*S)*s)/D;

e11 = ((O-L)*s - S*(o-l) + (L*o - O*l))/D;

de = -((S*(o-l)-O*o + L*l)*s + (O*l-L*l)*o)/D;

% And then the final equation calculating Gamma per 3-term correction:
G3term = (g-e00)/((g*e11)-de)


*** The Results ***

Gginpez = 0.3288 + 0.7535i

G3term = 0.3288 + 0.7535i


Note: This is not a mathematical proof, but for this example the two methods of error correction give identical results.

Which raises the question: For a one-port VNA, is there an advantage to using the GIN&PEZ equation over the standard 3-term error correction equations?

- Jeff, k6jca

I believe it is soldered, haven't bothered to look.

--
Terry VK5TM

I found this
2.8 inch 240*320 color TFT LCD touch screen with ili9341 driver IC touch panel | eBay

|
|
|
| $18.00 | |

|

|
|
| |
2.8 inch 240*320 color TFT LCD touch screen with ili9341 driver IC touch...

驱动器IC：ILI9341. 带SPI的 0.91英寸80x160 sma ll tft lcd. S TFT lcd display fo. TFT TRANSMISSIVE. composed of a color TF...
|

|

|


But it costs as much as a new nanoVNA.
Bob

I looked at it and it's not clear just how it replaces the original screen.? Is there a connector you just swap, or what?? I would give it a shot to make for a bigger display.
I have had some trouble with AliExpress as well.? I am putting them on the bottom of my list of suppliers, along with Amazon.? So far Ebay is on top.
Bob K6DDX

Found it on the english Aliexpress site, but when I went to pay they started wanting jpeg scans of my credit card, drivers license and other stuff.

They can stick that where the sun doesn't shine.

Terry VK5TM
( )

I was just going by the guy who said it was, specifically when he said it was a way to get QST issues.

Dave?? AB7E

Did you actually look at TinyCat/LibraryThing that W0RW uses? It is not a download or copying service.




73

-Jim
NU0C


On Sun, 22 Dec 2019 19:39:21 -0700
"stone_ridge_road" <xdavid@...> wrote:

the guy who thinks W0RW isn't engaging in copyright abuse

Not all.? I just think it's funny how easily people will rationalize their behavior, and I'm referring to the guy who started this thread as well as the guy who thinks W0RW isn't engaging in copyright abuse. ? Some of the other comments on free use have been informative for me, but I still think it's telling that nobody has thought to simply ask the author of the article for a free summary of his conclusions.

I don't know what links you referring to either .... I don't believe I've posted any relative to this thread.

73,
Dave?? AB7E

Hello again Erik;

It seems we are in complete agreement at my current level of understanding. This all being new to me, I've not yet attempted to dig into the details of how the calibration coefficient polynomials enter into the mix, but more than that, I now don’t see why they are used, much less how they are implemented.

Have you been able to identify where this approach diverges from those that employ the coefficients, why they are used, and where they differ? I am not familiar with the other algorithms, and since I’m also not a programmer, the tedium of extracting the existing algorithm(s) from the firmware source listings would be tedious for me and prone to error and misinterpretation.

GIN&PEZ have stated that the technique is not new, and if done carefully will yield the same results. I would expect this to be true if the standards used in both were set equal to ideal. I believe the NanoVNA meets that criteria, since there is no provision given for defining the standards.

So my understanding is that the standards uncertainties that seed the solution are applied as a fundamental part of the calculation... for example where S,O, & L are defined as ideal in the proof, they are defined as the standards in use, and the s, o, & l measurements establish the frequency response of the aforementioned and accurately defined standards. The result being an accurately corrected computational result, void of any need for further correction. This certainly appears to be the case, and compensating the results with additional bias following this first calculated result would seem to introduce errors and uncertainties in the outcome rather than improve accuracy.

My thinking has been that the process being used in VNAs today is To measure the standards, compute the results, then correct the computer results with a polynomial algorithm.

Is there something here that you think I may be missing or misunderstanding in my interpretation of their work.

--
73

Gary, N3GO

@Bob Albert

I understand that... It was more about accuracy over a smaller calibration range.

Thanks...

@Larry Rothman

I thought I had read about the "measurement points"
That was what led me to asking the question....

Thanks :)

Still, the calibration is based on the value of the load used for calibrating.? If it's 51 Ohms you will have SWR of 1.02 with a 50 Ohm load.
However, this is usually a trivial error.
Bob

Wow, Dave! You got in the bus very late, didn’t you?

NONE of us are interested in this new QEX article for VNA theory! It’s about the empirical comparison of the NanoVNA versus other higher grade VNAs. Do you get that difference? I’ve got about a GB of books and articles (including the ones you’ve put links to instead of actually reading this thread) on VNA theory, tutorials, user & service manuals, etc.

Way back on Saturday Night Live, Emily Litella used to get in the bus late, too. But when she caught on, she exclaimed, Never mind!

What’s in your wallet? Lol

Yes, for that narrow range, you will have a set of cal values.The Nanovna natively scans 101 points and if you narrow the freq range, it will still scann 101 points in the narrow range so the interpolation between the points should be better than if the unit needed to interpolate using the 50k to 900M range over 101 points.?
Clear as mud?



On Sun, 22 Dec 2019 at 7:50 PM, Dave H<xwebsubs@...> wrote: Hello,

I'm new to VNA's and have just received my nanoVNA

All the YouTubes show calibrating from 50Khz to 900MHz

If you set the range to say 45MHz to 55Mhz, and then calibrate.

will it provide more accuracy over that range, than using the 50khz to 900MHz calibration.

Thanks.. :)

Yes, you need to place the choke in its position ON THE INTENDED AMPLIFIER chassis. In this manner the exact distributive affects of shunt C are properly noted.

And with all do respects to the VNA, I would much rather do this measurement with a GDO.

Or... convert your vna into the equivalent of a GDO by exciting the ch 0 port into a small link coil which would utilize the measurement system as a coupled transformer. Not unlike the GDO, coupling must be kept to a minimum so that reduction in Q affect of the choke under test is minimized.

Note, you will need to measure both the anti resonate as well series resonate points as you described. In the series mode, the choke is shorted. Again, a direct connect of this arrangement to ch 0 would seem problematic. Consider the vna system is placing 50 ohms in series with the choke under test!

I would encourage you to correlate the measurements with the vna and with the GDO technique before applying 4000 volts to that nice tube!

I would also consider if you have one... the classic chokes like the old National Radio R175 which had strong series resonate problems in the 15 meter band. They would burst into flames! See if the vna points out the alarm. You should be able to pick up both the series resonate and parallel resonate signature on the vna-Smith plot. I know the GDO does pick up these signatures and I have modified those chokes based on the GDO results to fix their problems.

Alan

Hello,

I'm new to VNA's and have just received my nanoVNA

All the YouTubes show calibrating from 50Khz to 900MHz

If you set the range to say 45MHz to 55Mhz, and then calibrate.

will it provide more accuracy over that range, than using the 50khz to 900MHz calibration.

Thanks.. :)