开云体育

Maybe you are assuming that the characteristics of your RG58A/U are as
published. There is no guarantee that coax from lesser vendors actually
meets the published specification defined for RG58A/U. I am just saying you
must be careful of these things.

*Clyde K. Spencer*

Dave - I'd first like to explain why the Smith Chart method is showing 51.067 ohms for the RG-58A/U in the SimSmith simulation. As you can see in the attachments of my previous post, no adapters are being used. All we have is the source, coax, and load (100M ohms or 0 ohms).

If I followed your instructions correctly, the capacitance figures don't translate directly to the capacitance per foot in the coax. However, that would not matter if the method is still calculating the characteristic impedance correctly. The significance of the complex impedance at the two measurement points is not entirely clear.

I agree that additional measurements would help clarify the actual hardware being used.

Thanks, Kent
AA6P

Why not measure the capacitance of the adaptors, alone, without the cable.
You may have to borrow or buy those connectors identical to those on your
cabled adaptors.

Yes, a longer cable would also tend to swamp the capacitive affects of the
adaptors.

When I measured the coax cable as I described in a previous post, I didn't
use adaptors. I inserted the stripped coax (with absolutely minimal
conductor length) into the calibrated measurement plane of the 8753C. Of
course, the other end of the coax was treated in like manner for the open
and short measurement of capacitance and inductance, respectively.

Dave - W?LEV

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

Dave (W?LEV) - I looked at the method you described to determine the characteristic impedance of coax cable. See the attached plots from SimSmith using RG-58A/U.

/g/nanovna-users/message/23853

If I did this correctly, the capacitance measured 593.6 pf and the inductance measured 1.548 uH. The square root of L/C indicates a characteristic impedance of 51.067 ohms.

The resistance of the conductors may be a factor. The results are better with RG-213/U and a theoretical cable with zero resistance.

Jim - Note that the capacitance value in SimSmith is even higher than the 565 pf I measured on the NanoVNA. That value may not translate to the capacitance per foot in the cable.

I came across a third method to measure characteristic impedance from John Gord.

/g/nanovna-users/message/19735

This approach may be the best of the three methods I have tried so far.

73, Kent
AA6P

Jim Lux and I have both written 1 MegOhm (if Jim had such a resistor). Not
that this an optimum value by any means. It's just what we had. The sole
purpose of the resistor is to keep voltage from building up on the coaxial
(or parallel conductor ) transmission line. A dipole offers an open
circuit for DC voltages. With differential voltages from where ever in
nature (rain, dust, snow, wind, lightning,.....) applied across the two
conductors of the dipole or any other open circuit radiating structure,
voltages will build up on the transmission line. The resistor across the
feedline in the shack serves the same function as the bleeder resistors
across the filter capacitors of amateur power amplifiers DC supplies I
remember my Novice transmitter, the Heathkit DX-40. It had a large-valued
RF choke from the back of the PL-259 to chassis. The sole purpose of that
RF choke was to keep voltages from building up on the feedline - same as
the resistor in question.

As such, the resistor can be inductive. It doesn't matter as its' dealing
with DC, so the complex portion of the impedance is zero. The only
consideration I'd throw out is that it have adequate power dissipation to
not involve itself in loading the source impedance of the shack end of the
feedline. Also, that it be high enough not to potentially dissipate the
power applied to the shack end of the feedline. Both those statements
imply the same consideration. 100 k to a couple of megohms is good.

Dave - W?LEV

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

On 8/14/21 9:14 AM, David Feldman via groups.io wrote:

I need to measure S11 going into an RF power amplifier (1-15 MHz range of interest.)

While the amplifier has proven stable (so far) when driven through a 10 dB attenuator (from RF source), I am concerned that the impedance of the Nanovna's S11 port (I am using a first-generation unit purchased mid 2019) which is likely not 50 ohms could create a risk of oscillation (which could put the nanovna's circuitry at risk.) The nanovna does not have separable testset (like "option 011" on HP 8753 series), so I don't think I can apply methods in HP app notes pertaining to measurements of RF power amplifier circuits.

You could calibrate, with the 10 dB pad in place.? The SNR of the measurement is worse, but it's still valid.


The other thing is that the output Z of the NanoVNA is probably pretty close to 50 ohms. It's a resistive bridge.

If I insert (example) 6 dB pad between nanovna S11 port and amplifier input port, and calibrate the nanovna (open-short-load) at the "far" end of the pad, what impact would this have on S11 measurements?

They'll be correct - just that you have 12dB worse SNR - which means that the "floor" is higher - that reduces the accuracy of measurements into a "good match" because you're looking for a tinier signal.


Consider this.. the S11 measurement is looking at the reflected power.? Say the VNA puts out 0dBm, your DUT reflects -20dB, and the VNA sees -20dBm coming back.? If you put a 6dB pad in there, the actual power arriving back at the VNA is now -32dB down.

At those kinds of match, you are fine.? But say you've got a super duper amplifier with a -40dB S11.? Now you're measuring -56dBm, and that's harder, against the noise floor of the analyzer.

Are there other methods that would reduce risk when measuring S11 going into a live amplifier?

Thanks for any suggestions/pointers.

On 8/13/21 3:35 PM, Jim Lux wrote:

On 8/13/21 1:03 PM, Fred Moore wrote:

Just a quick question.? How many ohms would you suggest for that bleeder?

Megohms.

I used to use scrap black vacuum hose - it was UV resistant and happened to be conductive, and best, it was cheap.

Also low inductance, low capacitance.


Kent, AC1HJ

Over and above the NANOVNA issue: I presume you are dealing with a BJT
amplifier? Even with a FET PA, the following should be considered. The
input and output impedances of a power amplifier (PA) change noticeably
with drive power and corresponding out power, especially in the low-power
region before it reaches the linear region. The P(in) vs P(out) curve lof
a typical PA looks a bit like a forward biased diode VI characteristic.
Even if you connect the NANO directly to the input of the PA, the level of
RF from the NANO is too low to make a proper measurement of the input Z
which is representative of the PA at full power.

However, you're not out of luck. Insert an appropriate adjustable matching
network between your input to the amp (from a sig gen or transceiver - we
don't know the target power level of your PA or its rated input) and your
RF power source. Insert a GOOD SWR measurement device between the input to
the adjustable matching network and your RF source. Your RF source must be
from a good 50 ± j0 source - use good 50 Ω attenuators if needed to
stabilize the impedance. Adjust the variable matching network for 1:1
SWR. Then disconnect your PA and your RF source. In place of the RF
source, place a good 50 ± j0 resistor. Then connect the VNA to what was
the input from the matching network to your PA and measure the complex
impedeance. You're looking 'backward' into the required match for your
PA. The measured impedance made in this manner will be the complex
conjugate of the input to your PA. The resistance portion of the complex
measurement is real. The complex portion of the measurement must be
complemented with a reactive component of opposite sign to represent the
actual input impedance of your PA. This is the method the "big boys" use
to measure impedances of solid state power devices at RF frequencies.

Dave - W?LEV

On Sat, Aug 14, 2021 at 4:14 PM David Feldman via groups.io <wb0gaz=
[email protected]> wrote:

I need to measure S11 going into an RF power amplifier (1-15 MHz range of
interest.)

While the amplifier has proven stable (so far) when driven through a 10 dB
attenuator (from RF source), I am concerned that the impedance of the
Nanovna's S11 port (I am using a first-generation unit purchased mid 2019)
which is likely not 50 ohms could create a risk of oscillation (which could
put the nanovna's circuitry at risk.) The nanovna does not have separable
testset (like "option 011" on HP 8753 series), so I don't think I can apply
methods in HP app notes pertaining to measurements of RF power amplifier
circuits.

If I insert (example) 6 dB pad between nanovna S11 port and amplifier
input port, and calibrate the nanovna (open-short-load) at the "far" end of
the pad, what impact would this have on S11 measurements?

Are there other methods that would reduce risk when measuring S11 going
into a live amplifier?

Thanks for any suggestions/pointers.

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

This firmware conatain measure cable option (MARKER->MEASURE->CABLE):
Use 1/4 wavelength (on image this marker 1 point)
For detect cable impedance used 1/8 wavelength (on image this marker 2 point)
For detect cable loss used 1/2 logmag at current marker point

User need select correct frequency range (more long cable need select more long wavelength or less frequency), also used velocity factor set in DISPLAY->TRANSFORM->VELOCITY FACTOR
In this case length measured very accurate

You can also iterate the VF. Measure the cable length exactly. Use Transform and set VF to a guessed value for the cable. Set the marker at the cable end on the curve (the top) and see what length the Nano says. Now change the VF until the Nano reports the same length as the cable was measured for. Start with higher steps say 5 % to get a feeling of the right direction and change necessary.

73/Torbj?rn

I need to measure S11 going into an RF power amplifier (1-15 MHz range of interest.)

While the amplifier has proven stable (so far) when driven through a 10 dB attenuator (from RF source), I am concerned that the impedance of the Nanovna's S11 port (I am using a first-generation unit purchased mid 2019) which is likely not 50 ohms could create a risk of oscillation (which could put the nanovna's circuitry at risk.) The nanovna does not have separable testset (like "option 011" on HP 8753 series), so I don't think I can apply methods in HP app notes pertaining to measurements of RF power amplifier circuits.

If I insert (example) 6 dB pad between nanovna S11 port and amplifier input port, and calibrate the nanovna (open-short-load) at the "far" end of the pad, what impact would this have on S11 measurements?

Are there other methods that would reduce risk when measuring S11 going into a live amplifier?

Thanks for any suggestions/pointers.

Fred: You would need a resistor which would handle the power AND the voltage it would see; for legal limit transmit power, many otherwise-suitable resistors have insufficient voltage rating. Phil AD5X uses one or more 3MΩ high-voltage resistors:

You could also use a choke, Phil shows the construction of a legal-limit HF bias tee, the choke in it should also do the job:
For less than legal limit power much less robust (and cheaper) components could be used. Of course chokes are frequency-sensitive unlike resistors.

Bleeders are only required for antennas which do not have a DC return between center and outer conductors; many dipoles or verticals with baluns or transformers, folded dipoles, Yagis with folded dipole feeds or hairpin match, etc. already have a DC return inherent in their designs; a quick measurement with an ohmmeter will tell. If it shows a low resistance a bleeder is unnecessary.

I also use gas-discharge tube arrestors on ALL of my antenna feedlines and rotor control lines, with the gas tubes sized for the power levels to be used; in my station, receive only, rotor(56V), 100W (150V), or 500W (1000V). My arrestors, which have replaceable/swappable gas tubes, are mounted directly to one of my 4 station ground rods using DX engineering mounting clamps: Replacement gas tubes are available inexpensively from distributors like Digi-Key. I mark the arrestor with the installed gas tube's voltage for easy reference.
Here is a QEX article with more information:

My station has already withstood a nearby lightning strike, unlike my fiber internet modem/router and everything on my wired network, including some of the wires!

The arrestors are for lightning protection only, they will NOT bleed down to a low-enough voltage to protect a network analyzer input!
Even with bleeders installed it is good practice to short out any cable before attaching it to an instrument.

73, Don N2VGU

No, not necessarily. Some types of these do not short both conductors together and even if they do, there's no guarantee that the potential of the feedline will be the same as that of your device until they have been equalized. The feedline and antenna may all be at ground potential, but you and your device may have accumulated a static charge, which should be dissipated before making the connection.

Yeah, longer would be better, and there's no need to cut a specific length to make the measurement. Any reasonable length you have lying around can be used as long as you can measure its physical length with adequate relative accuracy. 1% is good enough. I've used arbitrary lengths I had on hand ranging from around 10' to over 50'. (I have a 100' tape measure). Set the max frequency of the measurement scan range appropriate to the length being measured. Higher for shorter lengths and lower for longer lengths.

On 8/14/21 6:41 AM, S Johnson wrote:

I have an idea for finding the velocity factor of any cable. Please let me know if this could work.

The NanoVNA can measure cable lenth accurately if the cable’s VF is known and can be input into the calculations. Most people either just guess at the VF or look up the cable’s manufacturer specs. But here is the idea for either unknown cable or for those wanting to confirm the manufacturers number. Using the cable of interest, make up a one meter length of cable, with the connectors you plan to use in the actual antenna system, and then use the NanoVNA to measure that, inputting different VF’s until the Nano shows 1 meter long. Then you’ll know the VF factor exactly.

Exactly.? Or you can put in 100 for the VF, measure the length, and compare the measured length to the actual physical length.? VF = physical length/VNA length@100 VF

I'd use more than a meter long piece, so the "connectors and fixtures" are a smaller fraction of the cable length.

I have an idea for finding the velocity factor of any cable. Please let me know if this could work.

The NanoVNA can measure cable lenth accurately if the cable’s VF is known and can be input into the calculations. Most people either just guess at the VF or look up the cable’s manufacturer specs. But here is the idea for either unknown cable or for those wanting to confirm the manufacturers number. Using the cable of interest, make up a one meter length of cable, with the connectors you plan to use in the actual antenna system, and then use the NanoVNA to measure that, inputting different VF’s until the Nano shows 1 meter long. Then you’ll know the VF factor exactly.

I would always discharge an antenna coax before measuring, just to be on the safe side. However, I would assume that if one has an unun, balun, or matching transformer between the antenna an receiver end of the coax the antenna already has a path to ground and is in all likelihood self-discharging. Please correct me if I'm wrong, but this is something to consider. Safest practice would be to short the coax regardless.? 73
Ken --? WB?OCV

I always connect my coax from all antennas to ground when not in use. No problems with static then and I can measure them any time I need/want to. Well, as long as I've just disconnected them from their normal grounded resting place.

--
Phil KE3FL

On 8/13/21 1:27 PM, Donald Hellen wrote:

Andrew . . .

On Thu, Aug 12, 2021 at 08:15 PM, Andrew Kurtz wrote:

Doesn’t that phrase “1 port device” suggest a single input? My antenna
has a single wire going outside; where would I find some other signal to
attach to the outer conductor of the small SMA on the VNA? Please don’t
reply if you want to tell me I am an idiot and still not help me out.

The single port needs two wires to function.



It does seem perhaps a little confusing, but the article above shows a port with two wires into a "black box" circuit. Think of it like this. A single port device might have only an input. A two port device might have an input and an output.

Donald KX8K

This is true, and one of the wires connecting one port to the other might be "inside the VNA" (i.e. the chassis ground)

On 8/13/21 1:03 PM, Fred Moore wrote:

Just a quick question. How many ohms would you suggest for that bleeder?

Megohms.

I used to use scrap black vacuum hose - it was UV resistant and happened to be conductive, and best, it was cheap.


Now, I'd probably just use a 1 Meg resistor if I had one.

Fred - N4CLA

On Fri, Aug 13, 2021 at 1:42 PM John Nightingale via groups.io <if455kc=
[email protected]> wrote:

As a matter of routine, consideration should be given to installing a

bleeder across the antenna's lead in the station to drain static
accumulation and never mind if it's a "grounded" antenna. A permanent
bleeder like that does not affect reception or transmission. For the
present purpose it would not affect v.n.a. observations. In urban areas
particularly, diode and gas discharge devices can cause problems; old
fashioned "R" is all that's required.

John
at radio station VE7AOV.

+++++


On 2021-08-13 8:27 a.m., Jim Lux wrote:

On 8/13/21 8:12 AM, David Eckhardt wrote:

When connecting any length of coax to any sensitive measurement
instrument,
I ALWAYS first grab the end of the coax and effectively short the end
with
a finger or hand. If there has been a recent lightning storm, you
might be
surprised and the discharge might surprise you in an undesirable manner.
The cylindrical capacitor that makes up the coaxial cable can hold a
charge
for a very long time.

Blowing dust, snow, or rain, are even more of a charging hazard than
the field from a passing thunderstorm. Continuous charging. It
charges until something breaks down, so you get a buzz or continuous
clicking. A sort of relaxation oscillator.