开云体育

On 7/16/22 10:32 AM, Douglas Butler wrote:

I have 4" x 10" sheets of brass and aluminum I can hold up to the antenna and see the VNA plots move a bit, and steel moves them a bit more, but nothing on the scale of the difference between my plots and the manufacturer's plots.

Then I'd go for "what's the mfr's test fixture/reference design look like" (if there is one..)

In my case, this will be a transmitter with the antenna driven by a MOSFET H-bridge with less than 1 Ohm source impedance and only a few inches of wire, so a low antenna impedance means more power can go into the antenna. But I am still getting used to using my NanoVNA and want to make sure I am not doing something stupid.

I have 4" x 10" sheets of brass and aluminum I can hold up to the antenna and see the VNA plots move a bit, and steel moves them a bit more, but nothing on the scale of the difference between my plots and the manufacturer's plots.

Doug

On 7/16/22 9:20 AM, KENT BRITAIN wrote:

Hi Doug, having done quite a few RFID antennas, that 1 to 1 is rather optimistic.
One thing to consider, do you have any other metal within a foot or so of the antenna?? ?You can easily see a pair of pliers passing a foot or so from the loop.
Good luck with your project, Kent WA5VJB

Indeed..
at 125 kHz and low power for embedded systems, 50 ohms matched circuits would be pretty unusual. Consider the AM radio in a car - not 50 ohms.

At low frequencies (consider 60 kHz WWVB receivers or 100kHz LORAN) receivers are usually just a high input Z, and the antenna resonance is used for front end selectivity. Environmental noise is huge, so it's not like you're trying to match for optimum NF.

And for transmitting, unmatched systems are also common - It's not like you're running a feedline 100s of feet - a lot of transmitters in this band are basically a switch to the supply rail that turns on and off, so effectively a zero ohm source impedance. You don't care about the match - if the goal is to radiate a milliwatt or two, Thevenin maximum power transfer theorem isn't super important.

"Badge reader" and RFID systems at low frequencies (and even the 13.56 MHz flavor) are interesting - the two antennas are within each other's (very) near field. It's often better to look at it as a poorly coupled transformer (since the magnetic field is often what's being used), if you want to try and model it.

Does the mfr provide a "how we measure it" document?

Test fixturing, sample circuits, etc. are important.

You'll see this kind of thing with devices designed for use in embedded systems - For instance, there are a lot of "monolithic WiFi" antennas, which are designed for 2.45 GHz, but only on a specific board layout in a specific environment. It's a tiny SMT "chip" with no connector, and if you put it into a SMT clip fixture, it probably doesn't read 50 ohm resonance.

If your measurements are showing more inductive, it might be that the mfr is assuming some capacitance to surroundings in the circuit. I notice in the picture that there's a bunch of capacitors and spots for capactiors. It might be that the mfr intends for you to add components for a specific application.

Hi Doug, having done quite a few RFID antennas, that 1 to 1 is rather optimistic.
One thing to consider, do you have any other metal within a foot or so of the antenna?? ?You can easily see a pair of pliers passing a foot or so from the loop.
Good luck with your project, Kent WA5VJB

I have a 125 kHz RFID antenna consisting of a coil and series tuning capacitor. The manufacturer supplied plots of the SWR and Z of the antenna showing a nice 50 Ohm 1:1 SWR at resonance. I have tried to replicate these plots using my newly purchased NanoVNA and I get radically different plots showing a load closer to 25 Ohms and a 2:1 SWR. Measuring the L and C separately gives about the same readings I get with my HP4801 vector impedance meter at 125kHz, but it is near the upper frequency limit of the HP4801 so results are shaky. I calibrated the NanoVNA over the same range as shown in the manufacturer's plots. My Smith plots are centered on the horizontal axis like I would expect them to be. The manufacturer's plots are shifted up to the inductive half of the Smith chart. Both show a nice resonant dip at 125 kHz.

Am I doing something wrong? Is the antenna manufacturer lying?

SherpaDoug WA1UWP

exactly !

Hey Dave,

I agree with you, my TI99/4A is still my favorite. (just kidding but I did use it for years) I now use win 7 and buy malware to keep it clean.

Mike C.

On Fri, Jul 15, 2022 at 11:44 AM, Diane BONKOUNGOU wrote:

Hello,
Thanks for your response I've got -462ps for the delay correction is it
good?
Best regards.

Your display looks OK but 462 ps would be the value for a very short transmission line - about 4.5 cm long. What is the length of your transmission line and what kind of coax is it?

Here is a video you might find interesting. It shows how to measure an antenna located inside a product...



Roger

I would solder the coax shield to the area labeled GND in the antenna design, not to the ground plane nearby.

Hi Dianne,

What model/version of the nanoVNA do you have? many do not cover the 2.4GHz bluetooth frequency band.

What diameter/type of coaxial cable are you currently using? Type RG316 has diameter 2.5mm and RG178 diameter 1.8mm.

My advice is to test your antenna with a length of cable that will be used in the final product with a connector that will be used in that design.

I hope the above is of some help to you

Kind regards

Ed

Thanks for your reply.
I managed to correct the delay
I was asking the same question to my tutor about the size of the cable in
relation to the PCB. I am on an internship and I have to design a PCB trace
antenna. The company wants the antenna size to be as small as possible to
integrate into their IoT device for low-energy Bluetooth communication. I
don't think I can get a smaller SMA cable than this for soldering, I will
look if you have a proposal, I'm a taker. Attached is the antenna design
and the solder image.
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.

Hello,
Thank you for your answer. My DUT is a PCB trace antenna. Attached is a
picture of the antenna. I need to perform the measurement in the feed line
of the antenna. I am in an internship and the company is designing an
antenna for their Iot device(Bluetooth lower energy communication). They
want the antenna to be as small as possible. I am wondering if the size of
the antenna feed line can skew the antenna measurement.
I succeed to correct the delay I think.
Best regards.

Le ven. 15 juil. 2022 à 16:20, Douglas Butler <sherpadoug@...> a
écrit :

Hello,
Thanks for your response I've got -462ps for the delay correction is it
good?
Best regards.

Le ven. 15 juil. 2022 à 18:22, Roger Need via groups.io <sailtamarack=
[email protected]> a écrit :

On Fri, Jul 15, 2022 at 05:58 AM, <dianebonk2@...> wrote:

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.

The most accurate way is to calibrate at the end of the cable using a
short, open and a 50 ohm SMD or very small resistor with absolute minimum
leads.

The other way is to use the electrical delay to compensate for the length
of the cable With a 33 cm (12 inch) cable I had to set the edelay to 3.18
ns to get the open at the far right of the smith chart. Try 0.1
nanoseconds (100 picoseconds) for each cm of cable and then keep adjusting
until you get close to the far right on the Smith chart. The velocity
factor of the cable you use will have an effect so the 100 picoseconds is
just an estimate to get you in the ballpark.

Roger

On Fri, Jul 15, 2022 at 05:58 AM, <dianebonk2@...> wrote:

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.

The most accurate way is to calibrate at the end of the cable using a short, open and a 50 ohm SMD or very small resistor with absolute minimum leads.

The other way is to use the electrical delay to compensate for the length of the cable With a 33 cm (12 inch) cable I had to set the edelay to 3.18 ns to get the open at the far right of the smith chart. Try 0.1 nanoseconds (100 picoseconds) for each cm of cable and then keep adjusting until you get close to the far right on the Smith chart. The velocity factor of the cable you use will have an effect so the 100 picoseconds is just an estimate to get you in the ballpark.

Roger

What exactly do you want to measure? What is the Device Under test, DUT? Is it the PCB? Is it the cable? Is it the combination of the two? If you just want to measure the PCB you need to do the calibration through the cable.

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@...:

I bought my NanoVNA in large part to have a solution looking for a problem to solve. My main use for this device was the SWR sweep capability for my antennas, but I soon discovered previously unknown problems that now needed a VNA solution. How pleasantly convenient.

Having a real time Smith Chart display is absolutely invaluable for getting a good feel on impedance for one. More, I have some SDR receivers that can be overloaded by a 50KW AM broadcast station a few miles away. The solution was to build custom filters. My first one, using an online passive filter design tool, was a 500 KHz LPF. It's quite satisfying whipping up a design and then see it perform. I found that even using 10% tolerance parts and tiny axial inductors, I could still produce an effective filter. The VNA proved to me that specifying a cutoff frequency with a good gap to the required cutoff frequency meant I could get by without strict tolerances and hi-Q torroid inductors.

My latest filter design was for a 22nd order 3.5 - 30 MHz BPF. I never would have ever tried that without my NanoVNA. And it works, knocking down the BCB and giving an excellent, although not as deep [nor needed], upper cutoff. This exercise was just as much about learning how to build and test the filter as its actual use.

Testing 455 KHz IF filters made for another fun experience, especially since I had to compensate for the wildly different 2K Ohm I/O impedances.

Yes the NanoVNA for me was a solution looking for a problem to solve. Kinda like going to school and learning how to solve a problem and then look for a problem to solve, lab exercises so to speak, but with immediate real world applications. Chicken or egg? Which comes first is not important. Just pick one and work towards the other.

Theory is always incomplete. It is said that the field of thermodynamics learned much more from the operations of steam locomotives than steam locomotives learned from thermodynamics. Often electrical building codes lag advances in electrical technology by decades. In my specialty of ship hull inspection there are always the ship design drawings and the "As Built" drawings. "We ran out of 10# plate so we used 12# plate until we could get more from the foundry."