NanoVNA V2 / Forum

Note: this page is a mirror of https://groups.io/g/NanoVNAV2.
Click here to join and see most recent posts.

Problems in bandpass crystal filter freq response


fabio.morsani 2020/12/07 08:43

I have a 8-pole crystal filter from a Drake RTX, 50 Ohm in/out, 5645 kHz center freq, 1.8 kHz bw.

This is his frequency response seen with a HPxxxx tracking/spectrum analyzer, -10 dBm tracking
signal output level:

This is how I have with nanoVNA V2plus4, same center freq same span:

If I look at the nanoVNA output signal I see this, CW, freq 5645kHz, no matter what output level 0...3 I set:

Zooming out the span to 100 kHz:

Is it normal? A circuit failure?

Thanks for help
Fabio

OwO 2020/12/08 01:01

All V2 versions switch the test signal rapidly and can not measure
crystals. I've added a note about this after the specifications table:
https://nanorfe.com/nanovna-v2.html#specifications

amirb 2020/12/07 09:01

very interesting information.
i think your output signal is an RF pulse train (i mean pulse modulated CW) but i dont know why it should be. can you look at it on scope?

however this by itself should not be a problem and the firmware should be able to extract the correct information from that
but your freq response on nano shows that firmware has a bug.

Fabio Morsani 2020/12/07 18:26

Please try to explain why a crystal 1.8 kHz bandpass filter cannot be
measured.
Performing a frequency sweep with a given step should be a loop from start
freq to end freq of:
- set the new freq
- perform the measurement

Isn't it?

"switch signal rapidly"?
I see the sweep is performed in several seconds, 2000 points.
Please try to explain.
Thanks
F.


Il giorno lun 7 dic 2020 alle ore 18:01 OwO <OwOwOwOwO123@outlook.com> ha
scritto:

OwO 2020/12/08 01:30

The port 1 is switched to a short or open at about 1kHz to obtain a
reference measurement. The DUT needs to have an impulse response much
shorter than 1ms. This corresponds to a physical size of 300 kilometers,
so no physical device can ever have a response this long, unless it's an
electro-acoustic device. That's why crystals can not be measured, and
only crystals. You won't encounter the issue measuring any resonant
filter that is not electro-acoustic.

Fabio Morsani 2020/12/07 18:58

Why cannot be added a sweep mode without switching?
In the past I've designed some similar crystal filter measuring the single
xtal
resonance freq with a DDS sweeping at 1 Hz step @ 10 MHz
and the results have been excellent!

Here a gaussian-to-12dB 300 Hz bw 6-pole 10 MHz crystal filter:

[image: image.png]

and this 10-crystals 2.6 kHz bw filter:

[image: image.png]

So, again, what about to add a simple sweep with no switching?
F.

Il giorno lun 7 dic 2020 alle ore 18:31 OwO <OwOwOwOwO123@outlook.com> ha
scritto:

John Gord 2020/12/07 14:45

The Ojisan firmware for the V2 can deal with crystals if set for narrow bandwidth and/or high averaging.  When set for 10 Hz RBW, it gates the RF on for about 400ms at a time.  At 1 kHz RBW, 50 AVG it gates the RF on for about 200ms at a time.  The latest Ojisan V2 firmware includes built-in crystal parameter calculation.
--John Gord

On Mon, Dec 7, 2020 at 09:31 AM, OwO wrote:

Fabio Morsani 2020/12/08 07:53

Oh good!
What is the firmware bin file considered to be stable and tested?

[image: image.png]
LC and crystal calculations seems to be implemented here: ·
Binary20200906test.Bin
<https://www.dropbox.com/s/jzgai1pmp3va3ku/binary20200906test.bin?dl=1>

Do you confirm?
F:


Il giorno lun 7 dic 2020 alle ore 23:45 John Gord via groups.io <johngord=
verizon.net@groups.io> ha scritto:

Fabio Morsani 2020/12/08 08:32

I've right now tried all "4 inch" releases but all gives all-white display.
Any idea?
---

Il giorno mar 8 dic 2020 alle ore 07:52 Fabio Morsani via groups.io
<fabio.morsani=gmail.com@groups.io> ha scritto:

Fabio Morsani 2020/12/08 08:46

Ah, I have a nanoVNA V2plus4.
---

Il giorno mar 8 dic 2020 alle ore 08:31 Fabio Morsani via groups.io
<fabio.morsani=gmail.com@groups.io> ha scritto:

OwO 2020/12/08 17:16

There isn't yet a firmware that allows measuring crystals but I'm
working on it.

Fabio Morsani 2020/12/08 10:21

Great!!!

Il giorno mar 8 dic 2020 alle ore 10:16 OwO <OwOwOwOwO123@outlook.com> ha
scritto:

kyle.husmann 2020/12/30 13:31

>
> There isn't yet a firmware that allows measuring crystals but I'm working
> on it.

Good to hear! So just to confirm, this is something that we are confident can be addressed in software? I read the recommendation is to get a nanovna v1 if we want to measure crystal filters, but if we hold out this will be possible some day with v2? (I'm trying to decide between a v1 and v2 and the crystal limitation is what's holding me back) thanks!

Fabio Morsani 2020/12/31 10:31

I think adding this feature to V2 will open a great opportunity for HAMs
still playing with crystals, I'm sure I'm not alone in the world.
---


Il giorno gio 31 dic 2020 alle ore 10:19 <kyle.husmann@gmail.com> ha
scritto:

David Eckhardt 2020/12/31 23:14

In testing crystal filters, remember the fixed impedance of the VNAs is
50-ohms, non-reactive (ideally). Many filters I've dealt with are not
designed to work properly in a 50-ohm system. Nor can they be properly
characterized in a 50-ohm system unless they were specifically designed for
that source/load impedance.

Dave - WØLEV

On Thu, Dec 31, 2020 at 9:19 AM <kyle.husmann@gmail.com> wrote:

> There isn't yet a firmware that allows measuring crystals but I'm working
> on it.
>
> Good to hear! So just to confirm, this is something that we are confident
> can be addressed in software? I read the recommendation is to get a nanovna
> v1 if we want to measure crystal filters, but if we hold out this will be
> possible some day with v2? (I'm trying to decide between a v1 and v2 and
> the crystal limitation is what's holding me back) thanks!
>
>
>

--
*Dave - WØLEV*
*Just Let Darwin Work*

Fabio Morsani 2021/01/01 08:39

It is possible to use well done step-up and step-down transformers.
Any input or output reactance introduced by the circuit can/must be tuned
to be eliminated: put an inductance and further variable capacitance to
tune the parallel LC circuit and see the filter ripple to return to the
designed values.

In the past I've selected crystals for SSB and CW 300Hz filters, designed
for 800 Ohm
in/out impedance, all with 50 Ohm instrumentation, here two example as seen
by
the spectrunm analyzer with tracking generator:

10-crystal SSB filter

[image: immagine.png]


Here 8-crystal gaussian-to-12dB CW 270 Hz filter:

[image: immagine.png]

It is possible to measure the single crystal if we have a true CW generator
with user selectable
sweep step/time, I did it with a DDS, on 2013:

[image: immagine.png]

[image: immagine.png]


*Fabio - I5LHY*

Il giorno ven 1 gen 2021 alle ore 00:14 David Eckhardt <
davearea51a@gmail.com> ha scritto:

Stephen Laurence 2021/01/10 09:00

I am a bit late in replying to this, but......
Possible solutions....

1.  Get a v2 or v2N and install the Obijan software on it. It is relatively easy to change/update the software for when you need it for crystals and reinstall the most recent software for general use when you are done on the crystals.

2. Get a v1 nanovna (eg h4 with a decent screen) as a second device. A mere £45 or so, but as most crystals and crystal filters are below 150mhz, the performance of a v1 on its fundamental will probably be satisfactory. The dynamic range above 300mhz is much worse than the v2, especially if you tweak the v2 a bit (a few extra capacitors, a few ferrite plates around the circuit board to suppress leakage) and you then (well, I do) have 60db dynamic range up to 3ghz.

You will never get down to 110db below passband (as in the previous post) with either solution, however. The nanovnas can however be put in your pocket. The gear for displaying 110db filter skirts will give you a hernia moving it about and if you get modern pro kit, you may have to mortgage the house.

The cost of a v2N plus a v1H4 is no more than two tanks of petrol for most cars.

Steve L

Fabio Morsani 2021/01/12 22:46

Thanks Steve for your info.

I agree with your comments, but to measure the crystal response with a
frequency sweeping
of a continous source and manually entering the sweep parameters it seems
to me to be
a simple task because I did it without effort (a DDS and a microcontroller,
a SWR resistive bridge
with diodes and linearization circuit for them).
To measure the crystal parameters is also easy in 50 Ohms systems, as you
seen in my post (formulas found somewhere in the radioamateurs literature)
if you can see the -3 dB points in the displayed curve.

110-120 dB of dynamic is really challenging also in the filter fixture,
cables
and additional shielding is often necessary.
I will use for that my old HP analyzer.

F.


Il giorno dom 10 gen 2021 alle ore 18:00 Stephen Laurence <
Gaslaurence@gmail.com> ha scritto:

To reply to this topic, join https://groups.io/g/NanoVNAV2

View this thread on groups.io