开云体育

On Aug 7, 2019, at 3:21 PM, tuckvk3cca <tuckvk3cca@...> wrote:

I am not a fan of calibrating with an impedance load that is too high. This is a limitation of the design. Theoretically we have a wheatstone bridge. 2 legs are 50 ohms and one leg is fixed at 50 ohms. I have a RLB that allows this to be set to any other value which is great. If you do not have this and calibrate with too high a value you eat into the dynamic range of the instrument, no amount of math correction can help you improve that. For accurate work one should really have a good RLB with adjustable standard and a good solid vector phase meter. Everything else are gimmicks.
Call me old fashion.


This business is quite déjà vu for me. As a undergraduate student I took a challenge to use a standard Wheatstone bridge with 10 or 12 ohm standards. The exercise was to measure to milliohm accuracy four orders of magnitude down. It was a tough exercise but I won. Nowadays you can buy 1 milliohm resistors with 1 or even 0.1% accuracy, but 5% is available on EBay, I did not have that luxury then.
Then I found an excellent paper on Wheatstone bridges that thought me a lot that I did not know. Still have the article somewhere I think. Drink deep or not taste the Pierian spring!

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Peter Gottlieb
Sent: Wednesday, August 7, 2019 9:07 PM
To: [email protected]
Subject: Re: [nanovna-users] Measurement challenge

How about running a cal with a known higher impedance load? Would that give a better chart? I don’t think you would need super accuracy for this application.


Peter

On Aug 7, 2019, at 3:04 PM, alan victor <avictor73@...> wrote:

After you get comfortable with the calibration process you may want to
consider measurement of components whose |Z| is much greater (or
smaller) than 50 ohms. Grab a couple of known parts out of the junk
box. I am working on a couple of tube amplifiers and in need of RF chokes and elements for a PI matching network. The desired chokes are 1-3 mH and the inductor for the PI match, 6 uH.

However any range where the magnitude of Z is several hundred ohms or
more and gamma gets to the edge of the chart will challenge the quality of your cal.

Currently I am measuring these devices in the SHUNT mode, simply as a one port.
Give it a go at HF, confine yourself below 30 MHz. See if you can get numbers that make sense.
Start with something whose value you know!

Alan

On Aug 7, 2019, at 3:04 PM, alan victor <avictor73@...> wrote:

After you get comfortable with the calibration process you may want to
consider measurement of components whose |Z| is much greater (or
smaller) than 50 ohms. Grab a couple of known parts out of the junk
box. I am working on a couple of tube amplifiers and in need of RF chokes and elements for a PI matching network. The desired chokes are 1-3 mH and the inductor for the PI match, 6 uH.

However any range where the magnitude of Z is several hundred ohms or
more and gamma gets to the edge of the chart will challenge the quality of your cal.

Currently I am measuring these devices in the SHUNT mode, simply as a one port.
Give it a go at HF, confine yourself below 30 MHz. See if you can get numbers that make sense.
Start with something whose value you know!

Alan

On Aug 7, 2019, at 3:04 PM, alan victor <avictor73@...> wrote:

After you get comfortable with the calibration process you may want to consider measurement
of components whose |Z| is much greater (or smaller) than 50 ohms. Grab a couple of known parts
out of the junk box. I am working on a couple of tube amplifiers and in need of RF chokes and elements
for a PI matching network. The desired chokes are 1-3 mH and the inductor for the PI match, 6 uH.

However any range where the magnitude of Z is several hundred ohms or more and gamma gets to the
edge of the chart will challenge the quality of your cal.

Currently I am measuring these devices in the SHUNT mode, simply as a one port.
Give it a go at HF, confine yourself below 30 MHz. See if you can get numbers that make sense.
Start with something whose value you know!

Alan

On Aug 7, 2019, at 12:50 PM, tuckvk3cca <tuckvk3cca@...> wrote:

Why are you starting at 900MHz? I would start from 1MHz and scan upwards and see how the magnitude and phases change as you increase frequency or equivalently how they rotate on the Smith chart. There are issues about cable loss too other than kinks in the adapters as you go up to 900MHz. Calibration, calibration is all about taking these factors into account. Sent from my Samsung Galaxy smartphone.
-------- Original message --------From: Peter Gottlieb <hpnpilot@...> Date: 07/08/2019 18:36 (GMT+01:00) To: [email protected] Subject: Re: [nanovna-users] Some basic load measurements I will perform some more measurements. One issue is that I only have a N male OSL (on a 7mm to N adapter on the VNA) so the BNC and SMA loads have to be via adapters. They are precision ones but are bound to have some effect, if only phase, at 900 MHz. Peter> On Aug 7, 2019, at 11:56 AM, tuckvk3cca <tuckvk3cca@...> wrote:> > For one port it should be sufficient. When you get a good match the return loss is low so adequate dynamic range is necessary. I would also check both the RL and it's phase. The latter is most tricky . I wish there is such a thing as a standard zero degree phase dummy. I often find I get good RL : 40dB or so but very lousy phases. Sent from my Samsung Galaxy smartphone.> -------- Original message --------From: Peter Gottlieb <hpnpilot@...> Date: 07/08/2019 17:29 (GMT+01:00) To: [email protected] Subject: Re: [nanovna-users] Some basic load measurements Is the basic OSL calibration sufficient? Or do I need to do something special?Peter> On Aug 7, 2019, at 11:18 AM, tuckvk3cca <tuckvk3cca@...> wrote:> > I don't understand you Peter, the 8753 is a network analyzer, which will> give you S11 the reflection coefficient. The log of its magnitude will be> the return loss. It should be able to convert your readings into complex> impedance R+jX too. I have the equivalent of the minivNA which only> calibrates with one point, an open load. I still get 38 to 40 dB return loss> at HF for some of my better 50ohm loads. I use the j-VNA software which> converts all magnitudes and phase measurements to S parameters of R+jX etc.> Most of the time I just live with this simple calibration procedure. On> transmission just a through pass for baseline calibration works for most> applications. For a full S parameter set measurement one does need to work> much harder in terms of calibration. > > -----Original Message-----> From: [email protected] <[email protected]> On Behalf Of Peter> Gottlieb> Sent: Wednesday, August 7, 2019 4:29 PM> To: [email protected]> Subject: Re: [nanovna-users] Some basic load measurements> > If you can tell me a simple way to set up to measure return loss I can do> that for the various loads. I can do the smith chart but it is very hard to> see in a photo as the loads are all just tiny dots near the center on the> horizontal line. Perhaps easier to see if I change display colors or save> onto the floppy instead. What I call the cheap BNC is an old mil spec> terminator probably designed for HF and it gives a nice little spiral around> the center. > > > Peter> >> On Aug 7, 2019, at 5:48 AM, tuckvk3cca <tuckvk3cca@...> wrote:>> >> That is expected Peter. Have you got the imaginary parts or phases? Best> to plot on a Smith chart too and see how your samples rotate from low to> high frequency. 1.02 swr corresponds to about 40dB return loss which is> very respectable at 900MHz.Sent from my Samsung Galaxy smartphone.>> -------- Original message --------From: Peter Gottlieb> <hpnpilot@...> Date: 07/08/2019 04:43 (GMT+01:00) To:> [email protected] Subject: [nanovna-users] Some basic load> measurements Resending from website as it didn't seem to go through as a> message. Also my pasted table from Excel lost formatting so I tried to fix> it to be more readable.I just did some very simple resistance and SWR> measurements using a HP 8753ES with 85046A, resistance was measured using a> calibrated Agilent 34401A in 4 wire mode.I did a very basic one port 3 point> cal using a Anritsu OSL which is specified to over 3 GHz.I took measurements> at 900 MHz.Load R ohms SWR SWR notesOSL> 50.052 1.001 FlatCheap BNC 51.104 1.908> Sloping up with freqNano load 49.044 1.019> FlatNarda 12.4 GHz 49.536 1.018 FlatTiny SMA> 50.787 1.009 FlatI am guessing there is some significant> reactive component in the BNC terminator. All three of the SMA loads showed> a flat SWR with frequency so I'm thinking they all have a minimal reactive> component.The difference in resistances while keeping SWR low was a bit of a> surprise to me. The load that came with the Nano is over an ohm off of the> load I used to calibrate yet the SWR remains at a low 1.019. Why is this? I> did the math and surprisingly this is indeed correct, per calculation the> SWR should be 1.021 vs my measured 1.019. I'd say this is darn close seeing> one measurement is DC resistance and the other is at 900 MHz.So my> conclusion is that SWR is not a sensitive number to see resistance> differences.Once I read some of the references cited I can do some more> advanced measurements.Peter>> >> > > > > > > > > >

You read my mind, Mike - I second the TDR function!

Maybe the next version of the NanoVNA could also have a broadband noise
source to act as a pseudo 'tracking generator' so we can do additional
swept measurements?

Just a thought.

Larry

On Wed, Aug 7, 2019 at 11:43 AM, Mike Brown wrote:

TDR. Other VNAs seem to have the feature. It would really help me with an
iffy feeder I have running down the garden.

On Wed, 7 Aug 2019, 16:34 SPACE OPQA, <spaceopqa@...> wrote:

What about coax cable length calculation?

And j+/-? :D

On Wed, 7 Aug 2019 at 15:10, dk1vi <dk1vi@...> wrote:

There are several features missing in current version:
* Two permanent markers. There is only one marker displayed when

hovering

with mouse over graph. Not so easy to make a screen shot and still
displaying marker value.

* With two markers a delta marker funktion would be fine.

* Sometimes the plots are rather noisy, so an averaging funktion

would

help.

Ernst

On Aug 7, 2019, at 11:56 AM, tuckvk3cca <tuckvk3cca@...> wrote:

For one port it should be sufficient. When you get a good match the return loss is low so adequate dynamic range is necessary. I would also check both the RL and it's phase. The latter is most tricky . I wish there is such a thing as a standard zero degree phase dummy. I often find I get good RL : 40dB or so but very lousy phases. Sent from my Samsung Galaxy smartphone.
-------- Original message --------From: Peter Gottlieb <hpnpilot@...> Date: 07/08/2019 17:29 (GMT+01:00) To: [email protected] Subject: Re: [nanovna-users] Some basic load measurements Is the basic OSL calibration sufficient? Or do I need to do something special?Peter> On Aug 7, 2019, at 11:18 AM, tuckvk3cca <tuckvk3cca@...> wrote:> > I don't understand you Peter, the 8753 is a network analyzer, which will> give you S11 the reflection coefficient. The log of its magnitude will be> the return loss. It should be able to convert your readings into complex> impedance R+jX too. I have the equivalent of the minivNA which only> calibrates with one point, an open load. I still get 38 to 40 dB return loss> at HF for some of my better 50ohm loads. I use the j-VNA software which> converts all magnitudes and phase measurements to S parameters of R+jX etc.> Most of the time I just live with this simple calibration procedure. On> transmission just a through pass for baseline calibration works for most> applications. For a full S parameter set measurement one does need to work> much harder in terms of calibration. > > -----Original Message-----> From: [email protected] <[email protected]> On Behalf Of Peter> Gottlieb> Sent: Wednesday, August 7, 2019 4:29 PM> To: [email protected]> Subject: Re: [nanovna-users] Some basic load measurements> > If you can tell me a simple way to set up to measure return loss I can do> that for the various loads. I can do the smith chart but it is very hard to> see in a photo as the loads are all just tiny dots near the center on the> horizontal line. Perhaps easier to see if I change display colors or save> onto the floppy instead. What I call the cheap BNC is an old mil spec> terminator probably designed for HF and it gives a nice little spiral around> the center. > > > Peter> >> On Aug 7, 2019, at 5:48 AM, tuckvk3cca <tuckvk3cca@...> wrote:>> >> That is expected Peter. Have you got the imaginary parts or phases? Best> to plot on a Smith chart too and see how your samples rotate from low to> high frequency. 1.02 swr corresponds to about 40dB return loss which is> very respectable at 900MHz.Sent from my Samsung Galaxy smartphone.>> -------- Original message --------From: Peter Gottlieb> <hpnpilot@...> Date: 07/08/2019 04:43 (GMT+01:00) To:> [email protected] Subject: [nanovna-users] Some basic load> measurements Resending from website as it didn't seem to go through as a> message. Also my pasted table from Excel lost formatting so I tried to fix> it to be more readable.I just did some very simple resistance and SWR> measurements using a HP 8753ES with 85046A, resistance was measured using a> calibrated Agilent 34401A in 4 wire mode.I did a very basic one port 3 point> cal using a Anritsu OSL which is specified to over 3 GHz.I took measurements> at 900 MHz.Load R ohms SWR SWR notesOSL> 50.052 1.001 FlatCheap BNC 51.104 1.908> Sloping up with freqNano load 49.044 1.019> FlatNarda 12.4 GHz 49.536 1.018 FlatTiny SMA> 50.787 1.009 FlatI am guessing there is some significant> reactive component in the BNC terminator. All three of the SMA loads showed> a flat SWR with frequency so I'm thinking they all have a minimal reactive> component.The difference in resistances while keeping SWR low was a bit of a> surprise to me. The load that came with the Nano is over an ohm off of the> load I used to calibrate yet the SWR remains at a low 1.019. Why is this? I> did the math and surprisingly this is indeed correct, per calculation the> SWR should be 1.021 vs my measured 1.019. I'd say this is darn close seeing> one measurement is DC resistance and the other is at 900 MHz.So my> conclusion is that SWR is not a sensitive number to see resistance> differences.Once I read some of the references cited I can do some more> advanced measurements.Peter>> >> > > > > > > >

TDR. Other VNAs seem to have the feature. It would really help me with an
iffy feeder I have running down the garden.

On Wed, 7 Aug 2019, 16:34 SPACE OPQA, <spaceopqa@...> wrote:

What about coax cable length calculation?

And j+/-? :D

On Wed, 7 Aug 2019 at 15:10, dk1vi <dk1vi@...> wrote:

There are several features missing in current version:
* Two permanent markers. There is only one marker displayed when hovering
with mouse over graph. Not so easy to make a screen shot and still
displaying marker value.

* With two markers a delta marker funktion would be fine.

* Sometimes the plots are rather noisy, so an averaging funktion would
help.

Ernst

What about coax cable length calculation?

And j+/-? :D

On Wed, 7 Aug 2019 at 15:10, dk1vi <dk1vi@...> wrote:

There are several features missing in current version:
* Two permanent markers. There is only one marker displayed when hovering
with mouse over graph. Not so easy to make a screen shot and still
displaying marker value.

* With two markers a delta marker funktion would be fine.

* Sometimes the plots are rather noisy, so an averaging funktion would
help.

Ernst

There are several features missing in current version:
* Two permanent markers. There is only one marker displayed when hovering
with mouse over graph. Not so easy to make a screen shot and still
displaying marker value.

* With two markers a delta marker funktion would be fine.

* Sometimes the plots are rather noisy, so an averaging funktion would
help.

Ernst

Did do a calibration right after the upgrade and save it to position 0
(zero). It shows the problem with or without a calibration.

Eventually figured it out and got it working. Apparently the new Aug 2
firmware does not actually change the frequency limit. It just leaves it
at what is was (800 or 900 MHz)

The only firmware that seems to be able to actually change the 900 MHz
limit to 800 MHz is the older June 22 firmware
(nanoVNA_800_aa_20190722.dfu)

Once the older software sets the limit to 800 MHz, the new Aug 2 firmware
ch with change it to 4 channels and leaves it at 800 MHz.

The new Aug 2 firmware ch also leaves the user area in a state of disarray
such that it will actually crash. It was the phase display that was
causing all the spikes. Finally managed to get the user area set up from
the front panel. It is working great now on 800 MHz 4 channel.

Thanks for all you help,

Jim K.

On Aug 7, 2019, at 11:18 AM, tuckvk3cca <tuckvk3cca@...> wrote:

I don't understand you Peter, the 8753 is a network analyzer, which will
give you S11 the reflection coefficient. The log of its magnitude will be
the return loss. It should be able to convert your readings into complex
impedance R+jX too. I have the equivalent of the minivNA which only
calibrates with one point, an open load. I still get 38 to 40 dB return loss
at HF for some of my better 50ohm loads. I use the j-VNA software which
converts all magnitudes and phase measurements to S parameters of R+jX etc.
Most of the time I just live with this simple calibration procedure. On
transmission just a through pass for baseline calibration works for most
applications. For a full S parameter set measurement one does need to work
much harder in terms of calibration.

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of Peter
Gottlieb
Sent: Wednesday, August 7, 2019 4:29 PM
To: [email protected]
Subject: Re: [nanovna-users] Some basic load measurements

If you can tell me a simple way to set up to measure return loss I can do
that for the various loads. I can do the smith chart but it is very hard to
see in a photo as the loads are all just tiny dots near the center on the
horizontal line. Perhaps easier to see if I change display colors or save
onto the floppy instead. What I call the cheap BNC is an old mil spec
terminator probably designed for HF and it gives a nice little spiral around
the center.


Peter

On Aug 7, 2019, at 5:48 AM, tuckvk3cca <tuckvk3cca@...> wrote:

That is expected Peter. Have you got the imaginary parts or phases? Best

to plot on a Smith chart too and see how your samples rotate from low to
high frequency. 1.02 swr corresponds to about 40dB return loss which is
very respectable at 900MHz.Sent from my Samsung Galaxy smartphone.

-------- Original message --------From: Peter Gottlieb

<hpnpilot@...> Date: 07/08/2019 04:43 (GMT+01:00) To:
[email protected] Subject: [nanovna-users] Some basic load
measurements Resending from website as it didn't seem to go through as a
message. Also my pasted table from Excel lost formatting so I tried to fix
it to be more readable.I just did some very simple resistance and SWR
measurements using a HP 8753ES with 85046A, resistance was measured using a
calibrated Agilent 34401A in 4 wire mode.I did a very basic one port 3 point
cal using a Anritsu OSL which is specified to over 3 GHz.I took measurements
at 900 MHz.Load R ohms SWR SWR notesOSL
50.052 1.001 FlatCheap BNC 51.104 1.908
Sloping up with freqNano load 49.044 1.019
FlatNarda 12.4 GHz 49.536 1.018 FlatTiny SMA
50.787 1.009 FlatI am guessing there is some significant
reactive component in the BNC terminator. All three of the SMA loads showed
a flat SWR with frequency so I'm thinking they all have a minimal reactive
component.The difference in resistances while keeping SWR low was a bit of a
surprise to me. The load that came with the Nano is over an ohm off of the
load I used to calibrate yet the SWR remains at a low 1.019. Why is this? I
did the math and surprisingly this is indeed correct, per calculation the
SWR should be 1.021 vs my measured 1.019. I'd say this is darn close seeing
one measurement is DC resistance and the other is at 900 MHz.So my
conclusion is that SWR is not a sensitive number to see resistance
differences.Once I read some of the references cited I can do some more
advanced measurements.Peter

Are the newer model VNA's going to have updated firmware?? Hal/WB4AEG