Does anybody know of a nano VNA that they recommend that is a decent unit that won't break the bank and has a dynamic range of about 110 db. I will strictly use it to tune band pass cavities, duplexers, etc.

Thanks for your help.

Joe

On 13/11/2025 19:23, Joseph Giardina via groups.io wrote:
> Does anybody know of a nano VNA that they recommend that is a decent unit that
> won't break the bank and has a dynamic range of about 110 db. I will strictly
> use it to tune band pass cavities, duplexers, etc.
> Thanks for your help.
> Joe

To tune just the notch you can add an amplifier at the input to port 2, and only
sweep the range around the notch frequency. Measure the cavity’s notch and pass
separately. Be careful not to feed too much RF into the VNA's port, though.

Cheers,
David
--
SatSignal Software
Web: https://www.satsignal.eu
Email: davidtaylor@writeme.com
BlueSky: @gm8arv.bsky.social, Twitter: @gm8arv

KC 901 ( different models available)

dg9bfc sigi



Am 13.11.2025 20:23 schrieb "Joseph Giardina via groups.io"
<Josephgiardinasr=gmail.com@groups.io>:

> Does anybody know of a nano VNA that they recommend that is a decent unit
that won't break the bank and has a dynamic range of about 110 db. I will
strictly use it to tune band pass cavities, duplexers, etc.

>

> Thanks for your help.

>

> Joe



_._,_._,_

* * *

I doubt any of the current crop of inexpensive VNAs has a dynamic range of
110dB. It’s mostly limited by the ADC following the receivers. To get 110 dB,
you’d need *at least* 18-19 bits. You do pick up some dynamic range because
you’re essentially summing multiple samples in the digital filter to get I/Q,
so you get a sqrt(N) improvement.



Sure there are plenty of audio 24 bit ADCs around (although in practice you
probably would only get 21-22 bits - Effective Number of Bits - ENOB in the
data sheet). I believe the original NanoVNA has 12-13 ENOB out of a 16 bit
ADC.



But all the rest of the circuitry has to be clean to that level, i.e. you need
a much cleaner sampling clock to make clock errors less than 1 LSB - you also
need a lower system noise (or larger stimulus). And the larger stimulus needs
to be clean.







> On Nov 13, 2025, at 19:42, Joseph Giardina via groups.io
<Josephgiardinasr=gmail.com@groups.io> wrote:
>
>

> 

>

> Does anybody know of a nano VNA that they recommend that is a decent unit
that won't break the bank and has a dynamic range of about 110 db. I will
strictly use it to tune band pass cavities, duplexers, etc.

>

>
>

> Thanks for your help.

>

>
>

> Joe

_._,_._,_

* * *

The only units I know of that can and do attain 110 dB of dynamic range are
"professional" and cost an arm and a leg. You might find a used and good
HP 8753C somewhere, but it will still cost in the several kilobucks. The
NanoVnas typically bottom out around 85 dB dynamic range *at best*.

Dave - WØLEV

On Fri, Nov 14, 2025 at 3:42 AM Joseph Giardina via groups.io
<Josephgiardinasr=gmail.com@groups.io> wrote:

> Does anybody know of a nano VNA that they recommend that is a decent unit
> that won't break the bank and has a dynamic range of about 110 db. I will
> strictly use it to tune band pass cavities, duplexers, etc.
>
> Thanks for your help.
>
> Joe
>
>
>

--

*Dave - WØLEV*

Hi Joe,

I don't believe you will find any nano VNA that will meet your spec. -90, maybe -95 is the best you should expect to find and you must be careful as many of the knock offs are far from it.

I have been successful in using my nano VNA Plus 5 which I bought from TINDIE web site directly to tune many cavities but it take some really careful tuning and I use a signal generator and a RF volt meter to check my final adjustments.

If you really plan to tune lot of cavities, I would look for a used HP Spectrum analyzer with a tracking filter option. They are out there if you watch closely and you should be able to find them for under $1K.

Regards, mark

Hi, Joe!
Yeah, you are not alone!
OK, given that the silly cavities are soooo picky about adjustments...
And the better they are, the more _dynamic_range_ you need, across the _entire_ swept frequency range.
[Of course, you could borrow nice professional $pectrum analyzer$, with tracking generator$, to do the job.]
OR another way of approaching the problem is to "hack" the _relative_noise_floor_ on your VNA rig & display...!
Here's what I have been doing successfully:
1- Obtain a nice low noise small signal Amplifier (and extra SMA-M to SMA-M cable)
- Gain about 20dB, low NF <~ 0.4dB
- Preferably with SMA connectors
- and simple power, like 5V USB, but preferred with internal battery for isolation, etc.
- Should of course cover the frequency range of interest, and a bit beyond (many "wideband" units work fine).
2- Key Feature- must include (or be rigged for) switching (internal or external) to
a) allow the test signal to _bypass_ the amplifier when OFF, and
b) allow the test signal be _amplified_ when ON.
c) And easily switchable _during_tests_ without disturbing elements or connections.
3- There are lots of devices on Ebay, but review specs carefully.
- A Good Example, and claimed 20 dB & NF 0.4 dB:
https://www.ebay.com/itm/396214171357
- Note- My current Amplifier is the similar "green" one, but unavailable in 2024, and without battery, and with uncertain NF.
4- Configure the normal test setup, _after_ normal VNA calibration
- VNA Source -> Cavities -> VNA Input
- Perform initial testing and tuning to align and seek the correct notch (even though you can't see below the noise floor)
- Ensure you optimize the best feasible tuning adjustments and approximate settings.
>> Specifically, the VNA display should show the pesky noise floor limits,
- With the cavities curves arising out of this (hidden) notch, both below and above the deep noise floor hole.
>> Then DO NOT change the frequency limits of the VNA, to ensure that your testing is _limited_ to _deep_ into the notch range!
- And thus prevent potential to overpower the Amplifier!!
5- >>> Then insert the Amplifier (OFF!!) inline between Cavities and VNA Input, for Gain On Demand!
- VNA Source -> Cavities -> Amplifier -> VNA Input
6- You can skip VNA recalibration, because at this point, the only meaningful criteria is the _relative_ depth of the rejection notches!
7- With amplifier OFF (bypass mode!), reconfirm essentially equivalent tuning performance, as above without the Amplifier.
8- Then turn the Amplifier ON, and,,,
>>> The hidden notch should lift up about 20 dB, and rise out of the noise floor, and then be very easy to see and tune!
- My VNA provides a floor about -95 dB, for a combination of about -115 dB, enough to tune most cavities.
9- When demobilizing the setup, be sure to turn OFF the Amplifier and VNA.
Soooo, that's my Short response to your Long Question!!  ;-)
IF anybody has better clues, Please let us all know!
73!
Gene :)
N3EV@arrl.net

You do not need a high resolution ADC to get a high dynamic range.
Signal processing gain will increase the dynamic range. Averaging,
oversampling and (bandpass) filtering can yield many extra bits of
effective resolution. Basically, there is a trade-off of resolution
versus time. The more time you have, the higher the resolution can be.
The analog spectrum analyzers increased dynamic range by lowering the
detection filter bandwidth, making it slower.

Op 14-11-2025 om 17:10 schreef Jim Lux via groups.io:

> I doubt any of the current crop of inexpensive VNAs has a dynamic
> range of 110dB. It’s mostly limited by the ADC following the
> receivers. To get 110 dB, you’d need *at least*  18-19 bits.  You do
> pick up some dynamic range because you’re essentially summing multiple
> samples in the digital filter to get I/Q, so you get a sqrt(N)
> improvement.
>
> Sure there are plenty of audio 24 bit ADCs around (although in
> practice you probably would only get 21-22 bits - Effective Number of
> Bits - ENOB in the data sheet).  I believe the original NanoVNA has
> 12-13 ENOB out of a 16 bit ADC.
>
> But all the rest of the circuitry has to be clean to that level, i.e.
> you need a much cleaner sampling clock to make clock errors less than
> 1 LSB - you also need a lower system noise (or larger stimulus). And
> the larger stimulus needs to be clean.
>
>
>
>> On Nov 13, 2025, at 19:42, Joseph Giardina via groups.io
>> <Josephgiardinasr=gmail.com@groups.io> wrote:
>>
>> 
>> Does anybody know of a nano VNA that they recommend that is a decent
>> unit that won't break the bank and has a dynamic range of about 110
>> db. I will strictly use it to tune band pass cavities, duplexers, etc.
>> Thanks for your help.
>> Joe
>

https://nanorfe.com/vna6000.html

[image: image.png]

I have the "A" version ( 95 dB max dynamic range ) for my needs, that's
adequate. The "B" version costs $ 710 more.

If you go for the "B" version, it's $1,499.00

for you, this may or may not 'break the bank'.

GL,
Mike
K6THZ



On Fri, Nov 14, 2025 at 10:32 PM a a via groups.io <N3EV=arrl.net@groups.io>
wrote:

> Hi, Joe!
> Yeah, you are not alone!
> OK, given that the silly cavities are soooo picky about adjustments...
> And the better they are, the more _dynamic_range_ you need, across the
> _entire_ swept frequency range.
> [Of course, you could borrow nice professional $pectrum analyzer$, with
> tracking generator$, to do the job.]
> OR another way of approaching the problem is to "hack" the
> _relative_noise_floor_ on your VNA rig & display...!
> Here's what I have been doing successfully:
> 1- Obtain a nice low noise small signal Amplifier (and extra SMA-M to
> SMA-M cable)
> - Gain about 20dB, low NF <~ 0.4dB
> - Preferably with SMA connectors
> - and simple power, like 5V USB, but preferred with internal battery for
> isolation, etc.
> - Should of course cover the frequency range of interest, and a bit beyond
> (many "wideband" units work fine).
> 2- Key Feature- must include (or be rigged for) switching (internal or
> external) to
> a) allow the test signal to _bypass_ the amplifier when OFF, and
> b) allow the test signal be _amplified_ when ON.
> c) And easily switchable _during_tests_ without disturbing elements or
> connections.
> 3- There are lots of devices on Ebay, but review specs carefully.
> - A Good Example, and claimed 20 dB & NF 0.4 dB:
> https://www.ebay.com/itm/396214171357
> - Note- My current Amplifier is the similar "green" one, but unavailable
> in 2024, and without battery, and with uncertain NF.
> 4- Configure the normal test setup, _after_ normal VNA calibration
> - VNA Source -> Cavities -> VNA Input
> - Perform initial testing and tuning to align and seek the correct notch
> (even though you can't see below the noise floor)
> - Ensure you optimize the best feasible tuning adjustments and approximate
> settings.
> >> Specifically, the VNA display should show the pesky noise floor limits,
> - With the cavities curves arising out of this (hidden) notch, both below
> and above the deep noise floor hole.
> >> Then DO NOT change the frequency limits of the VNA, to ensure that your
> testing is _limited_ to _deep_ into the notch range!
> - And thus prevent potential to overpower the Amplifier!!
> 5- >>> Then insert the Amplifier (OFF!!) inline between Cavities and VNA
> Input, for Gain On Demand!
> - VNA Source -> Cavities -> Amplifier -> VNA Input
> 6- You can skip VNA recalibration, because at this point, the only
> meaningful criteria is the _relative_ depth of the rejection notches!
> 7- With amplifier OFF (bypass mode!), reconfirm essentially equivalent
> tuning performance, as above without the Amplifier.
> 8- Then turn the Amplifier ON, and,,,
> >>> The hidden notch should lift up about 20 dB, and rise out of the noise
> floor, and then be very easy to see and tune!
> - My VNA provides a floor about -95 dB, for a combination of about -115
> dB, enough to tune most cavities.
> 9- When demobilizing the setup, be sure to turn OFF the Amplifier and VNA.
> Soooo, that's my Short response to your Long Question!! ;-)
> IF anybody has better clues, Please let us all know!
> 73!
> Gene :)
> N3EV@arrl.net
>
>
>
>
>

yeah, but there is a limit to what you can get (and what’s practical in a low
cost device like the NanoVNA)

You can get a sqrt(N) improvement in SNR, so to pick up 10 dB, you need 100
samples. to get 20 dB, you need 10,000 samples. The stock NanoVNA collected
48 samples in a millisecond, so, picks up sqrt(48) -> 8dB SNR. Newer versions
raise the sample rate from 48 kHz to 192 (a factor of 4, so 3dB improvement in
SNR).

14 bit ENOB + the 48 samples gets you to around 1 part in 100,000 for the
voltage measurement then there’s some effects from the turning I/Q into S21.
But call it 100 dB dynamic range, at best - probably more like 80 - which is
what we observe as a apparent noise floor of S21 measurements.



But to get more, you need to have a stable enough system that you can average.

For instance, if you integrate for 1 second instead of a millisecond you
getabout 15 dB improvement in SNR. But that also implies the measurement is
made with 1 Hz detection BW. Is the Allan deviation of the oscillator and
synthesizer chains stable to that level? Is the ADC clock clean enough? I
had problems with an SDR using a ~49 MHz clock that had some leakage from the
66 MHz processor clock on the power to the clock buffer) and that caused odd
results on measuring narrow band signals being swept across the receiver
bandwidth (it’s whether the intermods happened to be “in band” with the
measurement).



Making >100 dB dynamic range measurements is FULL of traps. You can do it
with remarkably primitive equipment, but you DO have to be careful you aren’t
fooling yourself.

(e.g. teaching new engineers that you need to change the input attenuator
setting on the spectrum analyzer and see if the “spurs” from the Unit Under
Test change relative amplitudes, indicating you’re seeing IM in the SA front
end)



And there’s plenty of stories about making IMD measurements and being careful
about the two sources not interacting with each other.





> On Nov 16, 2025, at 00:59, Reinier Gerritsen <reinier123@idcircuits.com>
wrote:
>
>

>  You do not need a high resolution ADC to get a high dynamic range. Signal
processing gain will increase the dynamic range. Averaging, oversampling and
(bandpass) filtering can yield many extra bits of effective resolution.
Basically, there is a trade-off of resolution versus time. The more time you
have, the higher the resolution can be. The analog spectrum analyzers
increased dynamic range by lowering the detection filter bandwidth, making it
slower.
>
>

>

> Op 14-11-2025 om 17:10 schreef Jim Lux via groups.io:
>

>

>> I doubt any of the current crop of inexpensive VNAs has a dynamic range of
110dB. It’s mostly limited by the ADC following the receivers. To get 110 dB,
you’d need *at least* 18-19 bits. You do pick up some dynamic range because
you’re essentially summing multiple samples in the digital filter to get I/Q,
so you get a sqrt(N) improvement.

>>

>>
>

>>

>> Sure there are plenty of audio 24 bit ADCs around (although in practice you
probably would only get 21-22 bits - Effective Number of Bits - ENOB in the
data sheet). I believe the original NanoVNA has 12-13 ENOB out of a 16 bit
ADC.

>>

>>
>

>>

>> But all the rest of the circuitry has to be clean to that level, i.e. you
need a much cleaner sampling clock to make clock errors less than 1 LSB - you
also need a lower system noise (or larger stimulus). And the larger stimulus
needs to be clean.

>>

>>
>

>>

>>
>

>>

>>
>

>>

>>> On Nov 13, 2025, at 19:42, Joseph Giardina via groups.io
[<Josephgiardinasr=gmail.com@groups.io>](mailto:Josephgiardinasr=gmail.com@groups.io)
wrote:
>
>

>>

>>> 

>>>

>>> Does anybody know of a nano VNA that they recommend that is a decent unit
that won't break the bank and has a dynamic range of about 110 db. I will
strictly use it to tune band pass cavities, duplexers, etc.

>>>

>>>
>>>

>>> Thanks for your help.

>>>

>>>
>>>

>>> Joe

>

>
>

_._,_._,_

* * *

A big limitation in high-dynamic S12 measurement like the notch depth of a duplexer is the port isolation within the VNA.  Isolation is hard to get in a small package, and there is a limit to what the "Isolation" part of an SOLT calibration can cancel out.
--John Gord

*This is also my question about the NanoVNA; I want to buy a low-cost VNA, if possible, to adjust BPBR cavities on the 2-meter amateur band. Therefore, I'd like to know if an original NanoVNA can be used satisfactorily for occasional cavity adjustments.*

*PU5AOM*

Just a thought. Do you really need a vector network analyzer to tune a
cavity notch? That is a scalar measurement. Can’t that be done with a
signal generator and a spectrum analyzer? Lots of dynamic range available
between the sig gen’s output level control and the SA’s sensitivity
adjustment.

Bob

On Thu, Nov 20, 2025 at 9:29 AM ADILSON MATOS via groups.io <pu5aom.t=
gmail.com@groups.io> wrote:

> *This is also my question about the NanoVNA; I want to buy a low-cost VNA,
> if possible, to adjust BPBR cavities on the 2-meter amateur band.
> Therefore, I'd like to know if an original NanoVNA can be used
> satisfactorily for occasional cavity adjustments.*
>
>
> *PU5AOM*
>
>
>

problem is that the cheap analysers do not have enough dynamic range

for a repeater you need 100.. better 120db

...

and the very cheap Nano V2 can do 90 dB at best

dg9bfc Sigi



Am 20.11.2025 15:47 schrieb "Bob W0EG via groups.io"
<morrisnc7=gmail.com@groups.io>:

> Just a thought. Do you really need a vector network analyzer to tune a
cavity notch? That is a scalar measurement. Can’t that be done with a signal
generator and a spectrum analyzer? Lots of dynamic range available between the
sig gen’s output level control and the SA’s sensitivity adjustment.

>

>
>

>

> Bob

>

>
>

>

> On Thu, Nov 20, 2025 at 9:29 AM ADILSON MATOS via
[groups.io](http://groups.io)
<pu5aom.t=[gmail.com@groups.io](mailto:gmail.com@groups.io)> wrote:
>

>

>> **This is also my question about the NanoVNA; I want to buy a low-cost
VNA, if possible, to adjust BPBR cavities on the 2-meter amateur band.
Therefore, I 'd like to know if an original NanoVNA can be used satisfactorily
for occasional cavity adjustments.**

>>

>> **PU5AOM**



_._,_._,_

* * *

Not sure it has to be a super duper SA. A Tiny SA doesn't have huge dynamic range, but you don't care if you overload it outside the passband, if you're adjusting the rejection. And certainly a tracking generator is nice, but "max hold" on the SA might work. And any signal generator and perhaps some suitable amplifiers (either to get more drive into the filter, or in front of the SA).