开云体育

On Sun, Feb 14, 2021 at 05:03 PM, Ed Krome wrote:

NWDZ board, like it says in the graphs. All calibrated on the same board with
a single cable, as stated. And results were repeatable; I ran it several
times. Are the board standards 100% accurate? Who knows? But the results were
remarkably consistent over specific ranges.

Ed,

Sorry that this discussion went so far off topic. It would be really interesting to get an answer to your initial question. From reading your posts and looking at your data I think it might be one of the following.

1. You can see in the photos of your board and the other board there is quite a difference in the layout of the cal loads. You want the reference plane to be right where the device under test (DUT) is connected. On your board it looks like the open and short are not at the same location as the load. This is fairly critical for accurate measurements. Also the DUT should be in the same position as the load especially for higher frequency measurements.

2. Your cap plot is pretty typical. Inductance will make the "apparent capacitance" rise with frequency. In your plot you went to 300 MHz. and if you tried 600 MHz. you will see the circuit at its self -resonant frequency. After that you effectively have an inductor. The exact frequency can be easily determined by plotting the S11 phase angle - it will abruptly change at resonance. I have attached a plot I did of a 10 pf cap with short and long leads which illustrates this point.

3. It is hard to figure out what is going on with your inductor plot. It should not be abruptly changing at 30 MHz. You do get considerable inductance change with ferrite inductors versus frequency because the complex permeability changes. With powdered iron very little and with air-core types it is self capacitance which has the most effect. It would be interesting to see the S11 phase plot and the resistance plot over this frequency range.

Regards - Roger

Hi Roger
Man, it sure did go off topic. (Bypass caps; where did that come from? Should be a whole thread on its own.) All I wanted was for someone to tell me how to use nano to measure caps and inds. Well, it forced me to do it right; get some data. I now understand more about component values and frequencies. And as long as I measure parts at very low frequencies, it appears that I get good data. And on three different machines. Now I'm gong to build some test fixtures for real world components and see what happens.
But... I now have all kinds of questions about why those components seem to retain their marked values at high frequencies (if they didn't, you couldn't calculate resonant frequencies), but that's another topic.
BTW: part of the weirdness in the inductor plots (sudden knee in curves) is due t coarseness in frequency steps. But finer steps wouldn't change the general trends. It's a mystery to me.
--
Ed K9EK

Instead of taking it personally and flying off the handle how about you offer an explanation as to why the Drake engineers put an electrolytic in this spot in the circuit? I posed this to Garey K4OAH (SK) who had as along and storied a career as you. He called it a "buzz kill" without further explanation I left it at that.

73

-Jim
NU0C

On Mon, 15 Feb 2021 18:24:37 +0000
"David Eckhardt" <davearea51a@...> wrote:

On second thought, forget it. Those of us who have been there, done that
know better. I have a well outfitted RF lab covering to 21 GHz and know
how to use the equipment. I'm through replying to your posts. Go find
someone else to flame......

Reapectfully submetted:

Dave - W?LEV

On Mon, Feb 15, 2021 at 6:17 PM David Eckhardt <davearea51a@...>
wrote:

OK. I'm stupid and really don't know what I'm talking about. I spread
snake oil, saucery, witchcraft, and magic as applied to RF. Smith charts
are just disguised wiggie boards. And, of course, the earth is flat and
all celestial movements are determined by epicycles and epispheres. I
spent 10 years in RF design and the last 30 in EMC/RFI. Color me stupid,
dumb, incapable of handling DC Ohm's Law and don't know which end of the
soldering pen to grab.

I do not appreciate being attacked for my posts. I am working on data and
will post to this group.

Electrolytic, tantalums do NOT make good bypasses for high frequencies. I
proved that to myself decades ago in RF design and following in control of
EMC/RFI. Go build a preamp using only SMD 'electrolytics'. Control RF
emission of a ?P ringing at 3.6 GHz with only SMD 'electrolytics'. I am
taking data on that and will present it to this group. Please hold off on
flaming me further until that work is complete?

Dave - W?LEV

On Mon, Feb 15, 2021 at 2:43 AM Jim Shorney <jshorney@...>
wrote:

Drake TR7/TR4310.

73

-Jim
NU0C

On Mon, 15 Feb 2021 03:31:07 +0100
"Dragan Milivojevic" <d.milivojevic@...> wrote:

Forgot to ask, what's this legendary piece of equipment,
it has transistors so it can't be that good ?

On Mon, 15 Feb 2021 at 03:27, Jim Shorney <jshorney@...>

wrote:

Yeah, I remembered that about the one on the left after I sent the

pic.

The one on the right he called a "buzz kill".

73

-Jim
NU0C

On Mon, 15 Feb 2021 02:07:44 +0000 (UTC)
"Bob Albert via groups.io" <bob91343@...> wrote:

The left hand one is bypassed with a .01 so that doesn't count. I

am

puzzled by the other one; what's it supposed to do?

--
*Dave - W?LEV*
*Just Let Darwin Work*

Hi,
Roger have shown two plots with increase of capacitance with frequency, as the leads become longer the capacitance seems grows. It is caused by inductance but as I understand the model of real capacitor, the series capacitance extracted from it should not change. Should it be a influence of series inductance on measurement accuracy? I mean with the frequency increase the inductance impedance becomes much larger than the capacitance so we are getting bigger ratio between of them destroying the capacitance measurement?

--
Regards,

Slawek/SP9BSL

Slawek,
The inductance is in series with the capacitance and as you approach series resonance the net impedance drops and if it is still below the resonance, the effect is a lower capacitive reactance which is equivalent to a larger capacitance.
Gary
W9TD

Thank you Garry for the explanation. It is clear now.

Slawek
SP9BSL

To test out my nanoVNA-H over LF to about 50MHz I dug out a SMD airwound inductor of just over 240nH. I measured it in three ranges and stitched the results together to allow a log plot from about 100kHz to 50MHz.

The results below are quite good I think. The nanoVNA struggles a bit at 100kHz but from about 500kHz onwards the nanoVNA does a very good/stable job of measuring the inductance of the SMD coil and the red series resistance trace looks good too with a fairly smooth response across the frequency range. The resistance trace starts to misbehave a bit above 35MHz but this is a very good result for a low cost VNA.

Jmr
What are you using to make your plots?
Gary
W9TD

I'm using Eagleware Genesys. This is old RF CAE software from 2004. I've had my nanoVNA-H for quite a while now and measuring the inductance of coils across LF through about 50MHz is what I mostly use it for. It can also give a fairly good indication of the coil Q as long as the Q of the inductor isn't too high.

Do you import the s2p from the NanoVNA then to analyze and plot?
Gary
W9TD

Here's a nanoVNA-H measurement of a Micrometals T50-6 powdered iron toroid with 22 turns spread over about 270degrees of the toroid. The inductance should be about 2.2uH and you can see the nanoVNA does a good job again. It also has a reasonable stab at measuring the series resistance although the data gets a bit noisy above 20MHz. I keep meaning to add averaging to my PC tools. The nanoVNA is much more powerful than a typical low frequency LCR meter because it can measure the inductance across a wide frequency range. The inductance at the design frequency can usually be measured quite accurately. It often manages to give a reasonable estimate of Q for powdered iron toroids like this.

On Fri, Mar 19, 2021 at 12:40 PM, Gary W9TD wrote:

Do you import the s2p from the NanoVNA then to analyze and plot?
Gary
W9TD

Yes, except in this case I'm making a 1 port s11 measurement with the nanoVNA. I use a very basic s1p data dump tool I wrote in VB and then I import the dumped s1p file into Genesys for analysis. It just takes a few seconds to do it all as I have an inductance template in Genesys for this type of analysis.

To give the nanoVNA a tougher test I tried measuring the Q of a small 50uH inductor across LF to 3MHz. In the plot below it struggles to indicate the Q as the data is very noisy and uncertain. The measurement would benefit from some averaging but it indicates a peak in the Q somewhere around 450 at about 700kHz. Clearly this inductor is at its best across 300kHz to just over 1MHz where it delivers an impressive Q factor. I think the nanoVNA is doing very well here although it would be best to measure the Q using a decent Q meter or by using the nanoVNA in a different way to measure Q.

On Fri, Mar 19, 2021 at 05:45 AM, jmr wrote:

Here's a nanoVNA-H measurement of a Micrometals T50-6 powdered iron toroid
with 22 turns spread over about 270degrees of the toroid. The inductance
should be about 2.2uH and you can see the nanoVNA does a good job again. It
also has a reasonable stab at measuring the series resistance although the
data gets a bit noisy above 20MHz. I keep meaning to add averaging to my PC
tools. The nanoVNA is much more powerful than a typical low frequency LCR
meter because it can measure the inductance across a wide frequency range. The
inductance at the design frequency can usually be measured quite accurately.
It often manages to give a reasonable estimate of Q for powdered iron toroids
like this.

If you make the measurement using the NanoVNA app by OneOfEleven you will see that it has adjustable sweep averaging and data smoothing built-in. You can also import the s1p file for post processing. The graphing capability is quite good with many scaling options. Once I starting using this tool I stopped using NanoVNA Saver. The calibration routine is very nice as well.

Roger

Thanks. I'll probably have to do a firmware upgrade to start playing with the latest third party tools. There seems to be a lot of choice for nanoVNA firmware and I'm not sure who provides the most robust firmware upgrade for my nanoVNA-H (2.8") at the moment. There are a few niggly bugs with my existing (old!) firmware.