开云体育

CAN Bus in a car has term resistor and each end.

On 7/21/22 6:12 AM, Magicbean wrote:

Hello - Totally new to VNA here...
I want to measure the characteristic impedance of 120 ohm shielded twisted pair wire. I know that you can measure it for 50 ohm coax using the NanoVNA but I wondered if there was any reason why it wouldn't work for non-50 ohm twisted-pair? The purpose of measurement is to make sure I fit the best termination resistors for long runs (100m ish) of cable for CAN bus cable.

It should work fine. That STP has a "reasonably well controlled" impedance, but i wouldn't be surprised by 5% changes. It might be most useful to do the TDR type display, rather than a S11. Then you can see the dips and peaks along the cable. (Fun experiment, tie a knot or kink it in the middle of the cable)

I do have a Siglent SSA3032X plus spectrum analyser for which there is a return loss bridge (RB3X25) but at around ?400. I'm not sure whether that would be useable for that kind of measurement, especially as the cable isn't 50 ohms.

No, that's not the right tool. The NanoVNA is what you want.

I assume the NanoVNA probably is suitable for measuring the 120 ohm pair and can probably offer more functionality even if some of the specs of the Siglent are rather better!? If so, what's the shortest length of twisted pair I could sensibly make measurements on? I have some 1m samples of cable which I would like to evaluate. The final cable runs will be in the order of 100m.

An interesting question. The "resolution" in TDR type displays is determined by the frequency span. So 1 GHz gives you 1 ns (20-30cm) resolution. The unambiguous range is set by the lowest frequency.

If you want to distinguish the "cable" from the "ends" (both at the VNA end and the termination end) you probably want the cable to be long enough that it's several resolution points. The NanoVNA can go up to 900 MHz, so a 3-4 m piece should be long enough.

Before you get into fixturing and all that, try it with a 2-3 m piece of 50 ohm coax, and see if you can distinguish the termination (i.e. do the cal with no cable, put the cable on, and then look at TDR with the short, open, and load at the far end)

BTW I know I can also fire a pulse down the cable and match with a variable resistor when the cable is easily accessible (and I might do that too) but I am interested to know if the NanoVNA would be useful for this and also, in some circumstances, I won't be able to get to the far end of some cables. Would it be necessary to normalise to something other than 50 ohms? Would I need any other components or equipment to do the test (apart form connectors to physically connect)?

NanoVNA does that nicely.

Whether you normalize or not depends on whether you need to have actual values, or if you're just doing a qualitative test. If you get a reflection from the end vs no reflection, you don't really care what the value is.

I have some other applications for the NanoVNA too, so I will probably get one (I was thinking NanoVNA H4 although there seems to be a bewlidering number of variants). I'd just like to know if it is suitable for this measurement as well. Or... if I could do everything with a Siglent RF bridge, maybe that would be a better focus as I already have the SA.
Thanks.

Have you used it with nanoVNA tests?

I haven't, Ed, but N6LF has. What I like about this particular transformer is that the passband is very flat. That reduces any calibration error that may arise from interpolation between frequency samples. Check the datasheet for a version with even wider frequency response that is not quite as flat.

Brian

On Thu, Jul 21, 2022 at 04:43 AM, Brian Beezley wrote:

Coilcraft SWB2010-PCL. 50 kHz to 100 MHz, 1:1 with secondary CT, 6-pin DIP.,
$4.54 each from Mouser. Note link to datasheet.




Brian

Looks useful.

Have you used it with nanoVNA tests?

Kind regards

Ed, G8FAX

On Wed, Jul 20, 2022 at 05:42 PM, Dragan Milivojevic wrote:

On Tue, 19 Jul 2022 at 07:48, Ed G8FAX <ed@...> wrote:

Jim Lux wrote: “google for "woodward balun balance quality 1983" “

Thanks, I no longer have academic access, so will need to buy or get site
of the paper some other way. However, there might be alternatives/better
solutions to pursue.

Thanks, but sc-hub.se is blocked in UK, so no access.

Hello - Totally new to VNA here...

I want to measure the characteristic impedance of 120 ohm shielded twisted pair wire. I know that you can measure it for 50 ohm coax using the NanoVNA but I wondered if there was any reason why it wouldn't work for non-50 ohm twisted-pair? The purpose of measurement is to make sure I fit the best termination resistors for long runs (100m ish) of cable for CAN bus cable.

I do have a Siglent SSA3032X plus spectrum analyser for which there is a return loss bridge (RB3X25) but at around ?400. I'm not sure whether that would be useable for that kind of measurement, especially as the cable isn't 50 ohms.

I assume the NanoVNA probably is suitable for measuring the 120 ohm pair and can probably offer more functionality even if some of the specs of the Siglent are rather better!? If so, what's the shortest length of twisted pair I could sensibly make measurements on? I have some 1m samples of cable which I would like to evaluate. The final cable runs will be in the order of 100m.

BTW I know I can also fire a pulse down the cable and match with a variable resistor when the cable is easily accessible (and I might do that too) but I am interested to know if the NanoVNA would be useful for this and also, in some circumstances, I won't be able to get to the far end of some cables. Would it be necessary to normalise to something other than 50 ohms? Would I need any other components or equipment to do the test (apart form connectors to physically connect)?

I have some other applications for the NanoVNA too, so I will probably get one (I was thinking NanoVNA H4 although there seems to be a bewlidering number of variants). I'd just like to know if it is suitable for this measurement as well. Or... if I could do everything with a Siglent RF bridge, maybe that would be a better focus as I already have the SA.

Thanks.

On 7/21/22 3:56 AM, DougVL wrote:

On Tue, Jul 19, 2022 at 09:00 AM, Jim Lux wrote:

Not actually pulses, it's a CW measurement

Sorry, my mistake.
How long does the CW signal last, at 101 measurement points?

Long enough for the synthesizers to stabilize, then capture ADC samples of the 5kHz output from the mixer for the receivers. A few milliseconds.

I thought that length would be called a pulse.
(In my Air Force AC&W radar technician days, a 6 millisecond transmission was called a pulse.)

Sure, it's pulsed, but the reason I describe it as CW is that the measurement is not made with looking at the reflection of a transmitted pulse as in a radar.

It's a pulsed CW measurement, and the measurement is made at the same time as the source is on.

The entire sweep is a few hundred milliseconds.

Coilcraft SWB2010-PCL. 50 kHz to 100 MHz, 1:1 with secondary CT, 6-pin DIP., $4.54 each from Mouser. Note link to datasheet.



Brian

On Tue, Jul 19, 2022 at 09:00 AM, Jim Lux wrote:

Not actually pulses, it's a CW measurement

Sorry, my mistake.
How long does the CW signal last, at 101 measurement points?
I thought that length would be called a pulse.
(In my Air Force AC&W radar technician days, a 6 millisecond transmission was called a pulse.)
--
Doug, K8RFT

THANKS!!!

Rather than cutting the cable, you should calibrate with the cable attached
to the nanovna, attaching the calibration standards at the other end of the
cable, using connector adapters if necessary. This way ensures that all
aspects of the cable (length, loss, etc.) are compensated for by the
calibration, perfectly removing all those circles from the smith chart.
Stan

For those looking for papers:

R. L. Jesch, "Repeatability of SMA coaxial connectors," in IEEE Transactions on Instrumentation and Measurement, vol. IM-25, no. 4, pp. 314-320, Dec. 1976, doi: 10.1109/TIM.1976.6312234.

At least 50 mate/demate cycles

D. Bergfried and H. Fischer, "Insertion-Loss Repeatability versus Life of Some Coaxial Connectors," in IEEE Transactions on Instrumentation and Measurement, vol. 19, no. 4, pp. 349-353, Nov. 1970, doi: 10.1109/TIM.1970.4313926.

10,000 cycles for 14mm GPC, 7mm GPC, N, SMA (gold plated ss, and plain SS)


There's also a post by Rich NE1EE last may giving some data from Keysight

/g/nanovna-users/message/28062

Yes, a LARGE POWER attenuator (or several seriesed)

Dave - W ?LEV

On Wed, Jul 20, 2022 at 9:52 PM Siegfried Jackstien <
siegfried.jackstien@...> wrote:

with a 60 db attenuator inline you could measure a 60 db amplifier ...
but in priciple you are right ... not really built for used inline after
a rig ... but inline after an amp (with proper attenuation!!) yes that
works

to say it short

RTFM!!!!!!

dg9bfc sigi

Am 20.07.2022 um 19:50 schrieb W0LEV:

Any amateur radio operator using a NANOVNA (or even the TinySA) should
absolutely know it's a low-power device and not fit for inline or
terminating installation!!!!!!! It ain't no Bird Wattmeter!!!!!!! If

one

is not familiar with these sorts of test equipment, read the "welcome

mat"

or "quick start" guide that is included in the box with all of them I've
bought (too many) well before even charging the unit(s)!!

Please.......please.........where did common sense go???? I know. Common
sense is not common and it is no longer taught.

Dave - W?LEV

On Wed, Jul 20, 2022 at 12:18 PM DougVL <K8RFTradio@...> wrote:

On Thu, Jul 14, 2022 at 07:56 PM, Ken wrote:

Be careful about suggesting using nanoVNA as an SWR Meter. Most SWR

Meters are

used between a transmitter and an antenna to show the SWR in the

transmission

line at an operating frequency. The nanoVNA is connected to a

transmission

line and antenna to show what the SWR will be when fed from a

transmitter at

various frequencies. It would lead to disaster if someone connected a

nanoVNA

between a transmitter and antenna.

That's another good reason or feature of the Nano's price! If you blow

it

up (by transmitting into it), it probably won't break the bank. BUT if

the

transmitter blows up too, well then you'll wish you looked (and learned)
before you jumped.

--
Doug, K8RFT

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


--
Dave - W?LEV

and if you want to keep the coax off then use sma to bnc and bnc to banana direct

dg9bfc sigi

ps nicely build with those pcb as open short load :-) i made a similar thingy for my capacitance/ inductance meter (even smallish smd caps can be measured that way)

Am 20.07.2022 um 17:49 schrieb Roger Need via groups.io:

with a 60 db attenuator inline you could measure a 60 db amplifier ... but in priciple you are right ... not really built for used inline after a rig ... but inline after an amp (with proper attenuation!!) yes that works

to say it short

RTFM!!!!!!

dg9bfc sigi

Am 20.07.2022 um 19:50 schrieb W0LEV:

On 7/20/22 13:53, Donald S Brant Jr wrote:

Your filter looks fine, no funny stuff in the response. You might try expanding the scale of the S21 trace downwards to see how far down you can see the harmonic attenuation.
73, Don N2VGU

Cool.

Here's a screenshot with only S21 on the lower left graph.

73

Stan
KM4HQE

Hello Siegfried,
Thanks for your response.
So If I cut more of my cable could I reduce the number of circles (the
phase shift)?. I struggle to correct the delay with a lot of circles.
Best regards.

Le ven. 15 juil. 2022 à 15:35, Siegfried Jackstien <
siegfried.jackstien@...> a écrit :

You have to calibrate at the cable end with open short load
And with such big cable soldered to such a tiny pcb your results will be
at least questionable...
On so high frequencies even an sma to n adaptor will shift your smith 90
degrees around... A long cable if not calibrated out will show a handfull
of circles (as you can see)..
Dg9bfc sigi

Am 15.07.2022 14:20 schrieb dianebonk2@...:

Hello everyone,
I am a beginner in using the NanoVNA, I have a calibration problem after
connecting a transmission line into the VNA. You can see some pictures of
the result in the attached document after connecting the cable to the VNA
in the attached file.
Could someone tell me how to correct the impedance shift introduced by

the

cable? Thanks
I went to the "electrical delay" menu to correct the problem by adding a
delay but it got worse. I don't know what to do.

I want to solder the cable afterwards into a PCB where we have antenna
traces.

On 7/20/22 10:12 AM, W0LEV wrote:

SMA connectors are for good two reasons: 1) Small size and 2)
Frequency response.
SMA connectors are bad for one reasonable consideration: Limited service
lifetime for mates/demates.

500 mate/demate cycles is typical spec sheet number, but I'd say that there are plenty of SMAs that have seen a lot more cycles and are just fine, as long as they're not bent or damaged.

There's a paper out there by some folks at Maury Microwave, where they mated/demated an SMA thousands of times with the intent of understanding the degradation in repeatability.

The reason a spec sheet calls out a number like 50 or 500 cycles is because that's a "testing" number that is bigger than the customer is likely to use, but which is small enough that the price is reasonable.

Typically it would derived from some MIL-STD and is basically a statement that the connector will meet the requirements after that many cycles.

Another thing to watch out for is a loss or mismatch spec. Most connectors do NOT have anywhere near the loss in the data sheet, it's more driven by the lowest number that is practical to measure in a manufacturing environment.

So an SMA (Amphenol M39012) at 1 GHz is 0.06 dB max loss - that's pretty close to the minimum measurable loss at 1GHz without going to a lot of trouble.
Likewise, the mismatch (VSWR) is given as 1.05 to 1.20 plus some frequency dependent factor. It would be hard to measure a VSWR of 1.01 (46 dB RL), even if that is what the connector usually is.

Banana jacks/plugs work fine at HF but not at or above 50 MHz. BNCs and
Type-N connectors are appropriate at and above 50 MHz. Top "careless" use
of BNCs is around 1 GHz. From there on upward in frequency, either SMA or
Type-N connectors should be used.

Well, I'd say that banana jacks and similar single pin connectors are something that you need be aware of the non-ideal nature. Fortunately, with a NanoVNA, one can actually measure that non-ideal.

If you must, SO-239 / PL-259 connectors are *only* for HF use. Their
impedance is NOT controlled as are the other referenced RF connectors.

True, the UHF connector isn't constant impedance, however, if you have a set of calibration standards in UHF, then that is managed in the calibration.

My understanding is that SMA connectors were originally designed for interconnections inside equipment which are usually installed and then not touched, so that the limited durability was not a factor; low cost was. It was only when they began to be used in test equipment that its durability shortcomings became apparent.
An improved design, intermateable with SMA, but with superior perfomance and durability and intended for test equipment, is the 3.5mm connector (and its derivatives, 2.9mm, 2.4mm, ...1.0mm, etc.) However it is also a precision connector so quite costly; but I am often able to convince my clients to install them on front panels of new test equipment we design.
Much of its improved performance is its use of air dielectric at the interface, which has more stable performance than the SMA's PTFE (Teflon?) interface with its variable air gap, and its beefier (thicker-walled) outer conductor construction.
I have purged my shack of UHF connectors; all cables are type N, BNC, SMA/3.5mm, or type F for RX-only. High-quality adapters or (preferably) replacement of equipment connectors when practical were used.
73, Don N2VGU