开云体育

?A new utility has been introduced in SimSmuth.

This utility can be used to connect to, calibrate, initiate measurements, write touchstone files, and import them for further analysis.

A quick overview is here



Ward
Ae6ty

On 7/14/21 12:02 PM, WB2UAQ wrote:

This is a long thread so this might have been mentioned already. I bet I am way out of line. Sorry.

Why can't the measurement be made with a 50 ohm network analyzer? We're just looking at impedance so it can easily be converted to SWR or other formats mathematically referencing 75 ohms. Maybe this is what the formulas show above are doing?

It is not easy to fabricate a 75 ohm OSL cal kit using discrete components. If this is an HF app maybe.

Yes, you could do it with a 50 ohm analyzer and mathematically convert.

I'm not sure a 75 ohm OSL kit would be super difficult. You'd have to choose a 75 ohm connector family (i.e. BNC, TNC, F, I don't know what others are available, probably there's a 75 ohm N) and then do the open and short and load - probably use a pair of 150 ohm or a quad of 300 ohm SMT resistors.? You could probably *buy* a 75 ohm termination that would be "good enough" (e.g. video systems and cable TV use them by the gazillion- surely there are higher and lower quality available).


I don't know of any 75 ohm systems above L-band (CATV and DBS satellite LNB outputs).

This is a long thread so this might have been mentioned already. I bet I am way out of line. Sorry.

Why can't the measurement be made with a 50 ohm network analyzer? We're just looking at impedance so it can easily be converted to SWR or other formats mathematically referencing 75 ohms. Maybe this is what the formulas show above are doing?

It is not easy to fabricate a 75 ohm OSL cal kit using discrete components. If this is an HF app maybe.

This is the most recent video on my presentation:
Nano VNA Part I The Basics (not much math)
Nano VNA Part II The Under the Hood Approach (Advanced topics)

I'll also include the latest PDF of the slides - enough for a few hours of discussion

The video listed below was recorded a few days ago by the Leicester Radio Society (UK)
TU G8PGO David Carter.

I'm learning as I am playing, so here are the latest two videos
The Part I is just the "basics" without any of the math. That video was also recorded.
If you spot any errors on either video please let me know. This is rather technical stuff.
The ending on both presentations was a bit abrupt since I ran out of time.


!AtLJBaL57qLbg5sz6VcyxJdPeQaCrw?e=RkNpYx

have fun watching
de k3eui Barry
k3euibarry@...

I test ojisankoubou Z normalization code, adaptate it to NanoVNA

Calibrate NanoVNA by 50 Om load
Measure 25 Om (SWR = 2 and dot at left smith)

After add DUT impedance as 25 Om and calculate how it should look
And see SWR = 1.0 and dot at center of smith

Screenshot from my H4

And quite a number of special applications use something entirely different
than either 50 or 75-ohms, not to mention 30-Ohms. Take the 'flash' lasers
at Los Alamos and Phillips lab, to name a couple I'm a bit familiar with.
They use slab lines to establish a low impedance in the vicinity of 2-Ohms,
or less. Why? That's the easiest way they can obtain the fast fise times
at full power.

Dave - W?LEV

On Wed, Jul 14, 2021 at 3:25 AM Jim Allyn - N7JA <jim@...>
wrote:

Actually, there is a good reason for 75 ohm coax. If you plot cable loss
versus impedance, you will find that minimum loss occurs at 75 ohms. When
you've got miles and miles of coax strung everywhere, as cable TV companies
do, you want to minimize loss to reduce the number of booster amplifiers
needed. In addition, since 75 ohm coax has lower capacitance per unit
length, it has better high frequency response than 50 ohm cables. Best
power handling ability occurs at 30 ohms impedance. I don't know that any
industry uses a 30 ohm system impedance, but - 50 ohms is a good compromise
between minimum loss at 75 ohms and best power handling at 30 ohms. (All
this is the best I remember it, there might be a bit more to it that I have
forgotten over the years.)

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

On 7/14/21 12:25 AM, Uwe Lange wrote:

On Tue, Jul 13, 2021 at 03:27 PM, David Eckhardt wrote:

So what if the "bull's eye" on the Smith Chart is not appropriately placed
in the center and appears to the right of center halfway between the center
and where the 100-ohm circle intersects the real axis? I know that is
75-ohms. So, what is the problem??

There are two issues:
First the one from OP: he built at 50/75ohm pad and calibrated the NanoVNA with
a 75ohm load. The pad transforms every impedance on its 75ohms side into the
50ohms world of the NanoVNA and the reported impedances are then off which
confuses the OP.

A resistive pad does NOT transform from a reference impedance of 75 ohms to a reference impedance of 50 ohms.? A transformer would (over some frequency range).

All a resistive pad (of whatever loss) does is provide a 50 ohm impedance on port A when port B is terminated in 75 ohms, and provide a 75 ohm impedance, when port A is terminated in 50 ohms. It does not, for instance, provide a 100 ohm impedance when Port B is terminated in 150 ohms.

There are several solutions to this: The probably easiest is to not use the pad
in the first place and calibrate with a 50ohm load. This would also avoid the
5.6dB loss of dynamic range incurred by the pad.

Right, and then you can just change the way the smith chart displays and VSWR, etc. are computed so they turn the 50 ohm measurement into some other reference impedance.

Another option is to keep the pad and calibrate with a 75ohm load but *inform*
the NanoVNA that the cal load was 75ohm. The NanoVNA then has a chance to
display correct impedance values. My patch to NanoVNAsaver achieves this.

Sure, you could do this.? But why not just calibrate with a 75 ohm load, leaving out the pad (which in this case would reduce dynamic range).

And then, you'd need to change some other code to make sure that displays and the calibration use the 75 ohm reference (and calibration) impedance.

Here is a good explanation






On Wed, Jul 14, 2021 at 5:25 AM Jim Allyn - N7JA <jim@...>
wrote:

Actually, there is a good reason for 75 ohm coax. If you plot cable loss
versus impedance, you will find that minimum loss occurs at 75 ohms. When
you've got miles and miles of coax strung everywhere, as cable TV companies
do, you want to minimize loss to reduce the number of booster amplifiers
needed. In addition, since 75 ohm coax has lower capacitance per unit
length, it has better high frequency response than 50 ohm cables. Best
power handling ability occurs at 30 ohms impedance. I don't know that any
industry uses a 30 ohm system impedance, but - 50 ohms is a good compromise
between minimum loss at 75 ohms and best power handling at 30 ohms. (All
this is the best I remember it, there might be a bit more to it that I have
forgotten over the years.)





--

Kenneth Hansen
Owner
ScanAm Technical Services, LLC

Actually, there is a good reason for 75 ohm coax. If you plot cable loss versus impedance, you will find that minimum loss occurs at 75 ohms. When you've got miles and miles of coax strung everywhere, as cable TV companies do, you want to minimize loss to reduce the number of booster amplifiers needed. In addition, since 75 ohm coax has lower capacitance per unit length, it has better high frequency response than 50 ohm cables. Best power handling ability occurs at 30 ohms impedance. I don't know that any industry uses a 30 ohm system impedance, but - 50 ohms is a good compromise between minimum loss at 75 ohms and best power handling at 30 ohms. (All this is the best I remember it, there might be a bit more to it that I have forgotten over the years.)

On Tue, Jul 13, 2021 at 03:27 PM, David Eckhardt wrote:

So what if the "bull's eye" on the Smith Chart is not appropriately placed
in the center and appears to the right of center halfway between the center
and where the 100-ohm circle intersects the real axis? I know that is
75-ohms. So, what is the problem??

There are two issues:
First the one from OP: he built at 50/75ohm pad and calibrated the NanoVNA with
a 75ohm load. The pad transforms every impedance on its 75ohms side into the
50ohms world of the NanoVNA and the reported impedances are then off which
confuses the OP.

There are several solutions to this: The probably easiest is to not use the pad
in the first place and calibrate with a 50ohm load. This would also avoid the
5.6dB loss of dynamic range incurred by the pad.

Another option is to keep the pad and calibrate with a 75ohm load but *inform*
the NanoVNA that the cal load was 75ohm. The NanoVNA then has a chance to
display correct impedance values. My patch to NanoVNAsaver achieves this.

The way I structured the patch also makes the value of the cal load resistance
the center of the Smith chart but this is a mere *side effect*. The patch could
be more elaborate to leave the center at nominal 50ohms but still allow to
calibrate with something different than 50ohm.

Which brings me to the second issue (and my primary interest):
To my best knowledge neither the NanoVNA firmware variants nor NanoVNAsaver
allow the user to specify a cal load resistance other than 50ohms.
Larry Benko has a good article ()
why that might be important to you. My quick'n'dirty patch now gives you that
option for NanoVNAsaver but at the cost of redefining the center of the Smith
chart.

I hope this explains better what I was doing here.

I'd tend to agree with you that 50ohms is fine as a standard, who needs 75?

However, as a reader of this thread I'd have to say that far from belly aching the active particpants in the thread are working to address what they see as a problem. You or I or others might disagree with the problem statement but if they think its a problem what's wrong with them trying to fix it?

All you need to do is make a measurement, reverse the DUT and make a 2nd measurement.

What you proposed is fraught with trouble and would not work without a lot of effort. The DUT would see 25 ohms rather than 50 unless you used splitters instead of the tees. There are probably other issues. Don't do it. Just reverse the DUT.

Have Fun!
Reg

All of the RF world standardizes on 50-Ohms as a system impedance (with
very few special applications)................EXCEPT the television
industries. Personally, I'm perfectly happy with a 50-ohm center to my
Smith Charts and embedding my 'creations' in a 50-Ohm system impedance for
evaluation. If you really want something else of the Smith Charts, use
normalized impedances as a tool. On rare occasions I make measurements in
a 75-Ohm system, usually for others. My test instruments - mostly HP -
play to the standard of 50-Ohms, although the 8753C has an option to
measure in a 75-Ohm system (which I never use) as does the Rigol SA. I've
been embroiled in a 50-Ohm system impedance for some four to five decades
with the rest of the RF world. If the television industry wants to use
75-ohms, have at it. I have little interest in a non-standard "standard".
So what if the "bull's eye" on the Smith Chart is not appropriately placed
in the center and appears to the right of center halfway between the center
and where the 100-ohm circle intersects the real axis? I know that is
75-ohms. So, what is the problem?? If you work exclusively with the
television 'standard' of 75-Ohms, go buy a far more expensive instrument
that will address that need. After all, for nominally $100, the NANOVNAs
offer more bang for the amateur's buck (and the professionals) than any
other piece of test equipment at similar prices. Even a good DMM can't
address the RF world to the extent of the NANOs!!! So....... p l e a s
e...... stop belly aching? ????

Just my two Pecos

Dave - W?LEV

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

last update for today:
I found another spot in the calibration part that had 50ohm hard-coded. The diff attached to this post also corrects for that.

for those who want to play along at home (with Linux):
git clone
cd nanovna-saver
wget <insert here the link to patch file attached to this post>
patch -p1 < ./50_to_75ohms.patch
pip3 install .
python3 -m nanovna-saver

If there is interest I'd go through the pain to turn this into a pull request to try to merge it into NanoVNAsaver.
This would still require more work since all tests in the project have 50ohm ref impedance hard coded as well.
I'm not even sure the patch has a chance of landing in the official project: Changing the ref impedance
has a good chance to shoot yourself into the foot ...

On Tue, Jul 13, 2021 at 04:15 AM, Larry Rothman wrote:

Gentlemen - if you look in the forum's file section, you will find NanoVNA & H
(not H4) firmware for 75 ohm by forum member Xenomorph.

Hi Larry,
Thanks, I already saw your previous comment and finally checked out the wiki.
Since I use the NanoVNAs almost exclusively together with a computer because of the larger screen a 75ohm firmware doesn't appeal to me. Even with that firmware the NanoVNAsaver would still display/calculate everything to 50ohm. In addition changing the reference impedance to a new value requires a recompilation of the firmware (and the 75ohm firmware is already quite old). I always wanted to try running a NanoVNA with a Z_0 different than 50ohm so mikek's question gave me the final push to try it.

I quickly made a female 75ohm SOL cal set from 3 female SMA PCB edge connectors, the 75ohm load is two 0805 150ohm resistors in parallel which measures 75.36ohm with a multimeter. I then used my patched NanoVNAsaver and went through a SOL calibration with that set from 50kHz to 250MHz. Finally I measured the 75ohm load and my 50ohm calibration load. The 75ohm is displayed properly in the center of the Smith chart while the 50ohms is unsurprisingly now left of the center. Marker values show the correct impedance. All tests done with a NanoVNA V2Plus4

On 7/13/21 9:19 AM, Jim Lux wrote:

On 7/13/21 7:29 AM, stephenmahrer@... wrote:

Simple 75 to 50 and 50 to 75 Ohm minimum loss matching pads.
These are lossy by nature, about 5.7dB.

Best!

Steve Mahrer

Bear in mind that a resistive pad might provide "minimum loss" but is not an "impedance transformer" for making measurements.

That is, it's not the same as changing the reference impedance on your plots - for an obvious case, a 50 ohm load with a 75 system is a S11 = -14dB (VSWR 1.5:1), but if you look at a 50 ohm load through a MLP, you see a S11 of -25dB and a VSWR of 1.1


????Series???? Shunt???? Impedance seen by Test???? gamma 75 gamma 50???? VSWR 75 VSWR 50
????|S11| (75)???? |S11| (50)
Load???? 43.3333???? 86.6666

ok that's ugly.. I attach the xls


But the point is that a MLP presents a good match at the design impedance to both sides, it's not like a 75:50 ohm transformer.

On 7/13/21 7:29 AM, stephenmahrer@... wrote:

Simple 75 to 50 and 50 to 75 Ohm minimum loss matching pads.
These are lossy by nature, about 5.7dB.

Best!

Steve Mahrer

Bear in mind that a resistive pad might provide "minimum loss" but is not an "impedance transformer" for making measurements.

That is, it's not the same as changing the reference impedance on your plots - for an obvious case, a 50 ohm load with a 75 system is a S11 = -14dB (VSWR 1.5:1), but if you look at a 50 ohm load through a MLP, you see a S11 of -25dB and a VSWR of 1.1


Series Shunt Impedance seen by Test gamma 75 gamma 50 VSWR 75 VSWR 50
|S11| (75) |S11| (50)
Load 43.3333 86.6666








0
43.3 28.9
-1 -0.267606 #DIV/0! 1.7
0.00 -11.45
5
48.3 31.0
-0.875 -0.2341293 15.0 1.6
-1.16 -12.61
10
53.3 33.0
-0.7647059 -0.2045892 7.5 1.5
-2.33 -13.78
15
58.3 34.9
-0.6666667 -0.1783299 5.0 1.4
-3.52 -14.98
20
63.3 36.6
-0.5789474 -0.1548335 3.8 1.4
-4.75 -16.20
25
68.3 38.2
-0.5 -0.1336858 3.0 1.3
-6.02 -17.48
30
73.3 39.7
-0.4285714 -0.1145514 2.5 1.3
-7.36 -18.82
35
78.3 41.1
-0.3636364 -0.0971558 2.1 1.2
-8.79 -20.25
40
83.3 42.5
-0.3043478 -0.0812724 1.9 1.2
-10.33 -21.80
45
88.3 43.7
-0.25 -0.0667121 1.7 1.1
-12.04 -23.52
50
93.3 44.9
-0.2 -0.0533163 1.5 1.1
-13.98 -25.46
55
98.3 46.1
-0.1538462 -0.0409506 1.4 1.1
-16.26 -27.75
60
103.3 47.1
-0.1111111 -0.0295006 1.3 1.1
-19.08 -30.60
65
108.3 48.1
-0.0714286 -0.0188682 1.2 1.0
-22.92 -34.49
70
113.3 49.1
-0.0344828 -0.0089689 1.1 1.0
-29.25 -40.95
75
118.3 50.0
0 0.00027065 1.0 1.0
#NUM! -71.35
80
123.3 50.9
0.03225806 0.00891424 1.1 1.0
-29.83 -41.00
85
128.3 51.7
0.0625 0.01701775 1.1 1.0
-24.08 -35.38
90
133.3 52.5
0.09090909 0.02463026 1.2 1.1
-20.83 -32.17
95
138.3 53.3
0.11764706 0.03179508 1.3 1.1
-18.59 -29.95
100
143.3 54.0
0.14285714 0.03855058 1.3 1.1
-16.90 -28.28
105
148.3 54.7
0.16666667 0.04493085 1.4 1.1
-15.56 -26.95
110
153.3 55.4
0.18918919 0.05096632 1.5 1.1
-14.46 -25.85
115
158.3 56.0
0.21052632 0.05668421 1.5 1.1
-13.53 -24.93
120
163.3 56.6
0.23076923 0.06210893 1.6 1.1
-12.74 -24.14
125
168.3 57.2
0.25 0.06726247 1.7 1.1
-12.04 -23.44
130
173.3 57.8
0.26829268 0.07216466 1.7 1.2
-11.43 -22.83
135
178.3 58.3
0.28571429 0.07683346 1.8 1.2
-10.88 -22.29

Simple 75 to 50 and 50 to 75 Ohm minimum loss matching pads.
These are lossy by nature, about 5.7dB.

Best!

Steve Mahrer

Gentlemen - if you look in the forum's file section, you will find NanoVNA & H (not H4) firmware for 75 ohm by forum member Xenomorph.