开云体育

On most firmware versions, the averaging or bandwidth settings can be manipulated to give longer settling time at each frequency step. This helps for narrowband devices.
--John Gord

On 7/8/23 3:57 PM, Ma?l H?rz wrote:

Hello,
I have read various sources on the Internet and it seem that the classic
NanoVNA is best suited for testing filters (below 300 MHz) and measuring
crystals.
It also seems that the DiLord firmware for NanoVNA-H (not H4) can be
used on the original NanoVNA. Does that mean NanoVNA and NanoVNA-H have
similar performance.
What about NanoVNA-H4? I read it does have worse performance for this
application (<=300MHz) and measuring filters/crystals. But I would
prefer its larger screen.
I thought about additionally getting the LiteVNA for its larger
bandwidth, but apparently the noise floor for lower frequencies is worse
than the original NanoVNA (what about NanoVNA H-4)?
I could not find any info about the RBW of NanoVNA (original),
NanoVNA-H, NanoVNA-H2, or LiteVNA.
My main interest is really testing lower frequencies, as mentioned above.
Any clarification would greatly help. What I find on the Internet is
kind of confusing between various firmware versions and upgraded version
of the various models.
Best regards,
Ma?l H?rz

Without getting into the details of the various firmware versions, if you're measuring high Q devices (like crystals), then using one of the NanoVNAs with a computer running software that lets you have a lot more points is quite useful.

The problem with high Q is that the NanoVNA only samples for a millisecond - that is, it switches the stimulus and receiver frequencies to the next step, then delays a *short* time to let the synthesizers lock, then measures for a millisecond.

If the thing you're measuring has a time delay, or takes a while to settle, then the measurement won't be as accurate. The "time constant" for a resonator (i.e. the time for the energy to decay to 37%, 1/e) is Q radians. Since one cycle is 2pi radians, the number of radians is 2*pi greater. So Q/(2*pi) is the number of cycles, each of which is 1/f, so after Q/(2*pi*f) seconds, the energy will be 37%. (down 4.3 dB)

If your resonator has a Q of 5000, and resonates at 10 MHz, we can run some numbers: 5000/(2*pi*10E6) = about 0.08 milliseconds - so probably not a big deal.

I've not thought about what phase error one would get from the delay, though.

Hello,

I have read various sources on the Internet and it seem that the classic
NanoVNA is best suited for testing filters (below 300 MHz) and measuring
crystals.

It also seems that the DiLord firmware for NanoVNA-H (not H4) can be
used on the original NanoVNA. Does that mean NanoVNA and NanoVNA-H have
similar performance.

What about NanoVNA-H4? I read it does have worse performance for this
application (<=300MHz) and measuring filters/crystals. But I would
prefer its larger screen.

I thought about additionally getting the LiteVNA for its larger
bandwidth, but apparently the noise floor for lower frequencies is worse
than the original NanoVNA (what about NanoVNA H-4)?

I could not find any info about the RBW of NanoVNA (original),
NanoVNA-H, NanoVNA-H2, or LiteVNA.

My main interest is really testing lower frequencies, as mentioned above.

Any clarification would greatly help. What I find on the Internet is
kind of confusing between various firmware versions and upgraded version
of the various models.

Best regards,
Ma?l H?rz

On 7/7/23 2:08 PM, Andrea I2UEA wrote:

Did it, but unfortunately it didn't work: same condition at 300MHz.
Is really the SI5351 generator overheating? The problem appears even when the device is just switched on.
Additional question: I live close to the centre of Milan and I receive very strong signals, especially from FM broadcast stations, receiving levels around + 6dBm (yes: 0,44 V !!!) or so. Can they affect the masurements?

Yes, very much so
There are two effects to be aware of:
1) if the signal is close to the test frequency, it makes it through the receiver. Broadcast FM is pretty wideband - the VNA's receiver is about 1 kHz wide, so if you're sweeping across that band, you'll likely get interference

2) saturation of the receiver. I can't remember what the max input before saturation is, but considering that the VNA puts out around 0 dBm, and with a short, all that comes back to the receiver, 6dBm probably won't saturate the mixer.

You reset calibration after change threshold? Old calibration contain bad data

Did it, but unfortunately it didn't work: same condition at 300MHz.

Is really the SI5351 generator overheating? The problem appears even when the device is just switched on.

Additional question: I live close to the centre of Milan and I receive very strong signals, especially from FM broadcast stations, receiving levels around + 6dBm (yes: 0,44 V !!!) or so. Can they affect the masurements?

si5351 generator overheating and not stable work near top limit
Need reduce threshold to 295MHz, save config, recalibrate

This solve problems in 290-310MHz

Others threholds - only fixed inside firmware
Try add radiator on SI5351 generator

Hi, I am sure that almost 3 months ago, when I measured my Log Periodic antenna Create-5130-1N, its SWR was almost perfectly flat in all the 50-1300 MHz range. However I am now seeing strong and variable spikes in a whide band around 300 MHz which, if I properly understood, might be probably fixed by changing the setting CONFIG - EXPERT SETTINGS - THRESHOLD. However I am also seeing the same problem around 500 MHz as per attached images. Please note:
1) I didn't change anything in my NanoVNA-H4, apart updating the firmware
2) After updating the firmware I re-calibrated my unit
3) I am seeing the same spikes even by measuring my 2m/6m/70cm vertical antenna (Diamond V2000A)
4) No problem if I measure a 2m/70cm portable antenna

What can I do? Thank you for your comments and help.

- working update methods for me:

I have original devices, - I always like to use the STM32CubeProgrammer


One device (I think not important which one)
by the tray icon (same like usb drive) shows me: STM32 BOOTLOADER
- at the same time in Device manager: STM Device in DFU Mode (Yes, - I didn't mix up the description!)
- at the same time in STM32CubeProgrammer at port: No DFU detected...
but with the aplication; the FW upgrade working.
Here is the link:


On the same computer one other device (I think not important which one)
by the tray icon shows me: STM32 BOOTLOADER
- in Device manager also: STM32 BOOTLOADER
the STM32CubeProgrammer see at port: USB1
- then possible to connect, and making a FW upgrade

Istvan

Got the firmware update done this afternoon and it's working fine! Thanks again for the help!

Tom AE5I

I have check Core Isolation and it is OFF.
Many thanks. There are other driver that not permit activation of this protection.
Thank you for help.

--
73 de Salvo IZ7VLL.

Thanks Alan, and thanks to everyone else for sharing your knowledge and experience.

But now I must move on. The sun is shining and the coil is up in the air. I have optimised the antenna wire lengths for resonance and I'm looking forward to making some good contacts.

I have another similar coil and will continue the measurements on that when I get a moment.

Mike

--
Mike

Maybe this will help...

/g/tinysa/topic/97866810

Istvan

I have the same problem. Now I'm waiting for new one.
I able to upgrade firmware but not to use vna with app on Windows 10 or Debian 12.

--
73 de Salvo IZ7VLL.

I should add, strap a bar or plate where the LC is located to get a sense of the fixture R. Back out that mesured loss to get Rs fixture. You
can use that mesure to provide some correction.

In the same fixture, check the inserted 15 ohm only. The uniform loss should be ~ 8.5 dB.
It is possible that not all the current with the LC inserted is passing via the inserted R.
Since this is a distributed loss,

Hi Brian,

Thank you for your explanations.
That is what I love about this forum, I learn something every day.

The coil length explanation is really interesting and makes sense given your explanation.
As for the resistance issue, I felt a niggle as I wrote it. I would have put it down to skin effect. But I was not aware of the proximity effect. Makes sense.
Again thanks for the excellent explanations.

Cheers...Bob VK2ZRE

On Wed, Jul 5, 2023 at 09:26 PM, Bob Ecclestone VK2ZRE wrote:

But in your 2nd iteration, I notice the coil length is significantly longer
(156.8mm vs 80mm) with the same coil diameter (37.5mm) yet the inductance
remains the same at 97.8uH.
Surely the inductance should increase.

Bob, for a given number of turns, inductance decreases as the coil is lengthened due to lower coupling between turns. The optimizer maximizes Q while keeping inductance constant. It increased the number of turns just enough to compensate for the smaller coupling in the longer coil.

And finally, the Resistance (of the winding?) for the 2nd iteration is lower
than the 1st despite using a longer length of wire (6 vs 7.39).

Resistance is the RF resistance of the coil plus leads due to all causes. It is R in the R + jX model. Mike close-wound his coil with no space between turns. This minimizes coil length but maximizes the proximity effect. This magnetic effect bunches current inside a turn due to the proximity of adjacent turns. It increases RF resistance in the same way as the skin effect, by reducing available copper. The proximity effect is strong. I've never seen a close-wound coil where spacing the turns didn't increase Q even if it required more wire. The optimizer increases coil length to the point where any additional drop in the proximity effect is offset by the increase in wire resistance.

Brian

On Wed, Jul 5, 2023 at 02:55 PM, Mike wrote:

It's solid, not foam.

Actually foam, being mostly air, would be better. Another approach would be to drill many holes in the material, leaving a skeletal frame which, again, is mostly air. This is essentially what B&W does with their Miniductor stock, plastic rods are melted into the wire coils, stabilizing them physically while ensuring that the electric fields mostly encounter air.
73, Don N2VGU

On Wed, Jul 5, 2023 at 08:30 PM, alan victor wrote:

Hi Mike.

I suspect that the dielectric loss of your coil is secondary. A Qul of nearly
350 implies to me that the primary losses are likely copper wire R and
fixture.

A couple of useful things to do. Place a known R in series with the coil.
Deliberate de-Q and return to a measurement on Rs.

The number calculated should make sense. Example, suspect 7 ohms is the "real"
number, place a 10 ohm (low inductance R) not a WW, you should
see say 17 ohms. This is simply a sanity check.

Second, in looking at your solenoid coil, the L/D ratio looks wrong for
optimum Q. Something like L/D of 0.5 to 1 per Terman is better served for Q.
There is quite a few papers on this relation. But worth investigating if Q is
the big deal.

Hi Alan

I tried the experiment with a 15 ohm carbon resistor but it didn't produce the expected result. The calculated resistance went up from 7.5 ohm to 12.6 ohm.

I'm aware of the effect of L/D on Q. On a previous antenna I used a coil diameter of 48mm but I want to reduce the visual impact because our garden backs on to a footpath around a lake.

--
Mike