开云体育

ctrl + shift + ? for shortcuts
© 2025 Groups.io
  1. Nanovna-Users
  2. Topics

Test fixture question

 

I am building a test fixture for use with my nano-VNA. The primary purpose is to test/characterize a group of crystals for use in a crystal ladder filter in a couple of receiver projects I am using to stay busy through this crazy time we are experiencing. It is based on a simplified version of the IEC444 fixture as implemented by AE1C.
One of the suggested fixtures is found on github, in the project by Battosai42. In the supporting hardware for his amcp program, the impedance matching is done with a pi attenuator as shown in the (crude) text schematic below. For simplicity, only the input side is shown. The output is a mirror image.

O----+-----( 66.2 ohms )-----+---------------------+------[DUT]--
| | |
| | |
(159 ohms) (14.2 ohms) (33 nF)
| | |
| | |
O---+-----------GND-----------+---------------------+----------------

Now to the point of this. The capacitor value (33nF/0.033uF) seems very large for use at all but very low frequencies. Virtually all of the similar designs lack this capacitor entirely. So am I missing something? I have seen nothing of this in other designs discussed on the web, in QEX, or other discussions. Perhaps someone here can enlighten me, as well as the group.
Thanks for your consideration.
73,
Bill
P.S.
As soon as the design is complete, I plan to post the files here and make the boards available through OshPark.
 

Yeah, that cap seems awfully big to me, too.? I recently built a similar fixture for crystal testing, and I didn't use any capacitor in it.
 

Hi Bill
Sound as an interesting project. I visited the gitbug and found not much description about the text fixture. I suppose you found it via KiCad. It must be a bug it will not work up 150MHz as claimed.
I did such an test fixture years ago
Test Fixture for DG8SAQ VNWA and N2PK VNA.pdf
And calibrated at the DUT position not and not as I think the gitbug application does in the input/output side which introduces error as the two attenuators are included and not without amplitude variation and phaseshift variation over the frequency range. However is the application (as said there is not much info how to use the application) is measuring S21 only and calibration done by a shorting wire substituting the DUT then attenuators are waved and part of the calibration. Then watch out as isolation should not be done neither SOL.
That is what I can conclude but let me hear what you experience or have you found a usage guide for the gitbug application.
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af Bill Buoy
Sendt: 9. maj 2020 20:35
Til: [email protected]
Emne: [nanovna-users] Test fixture question

I am building a test fixture for use with my nano-VNA. The primary purpose is to test/characterize a group of crystals for use in a crystal ladder filter in a couple of receiver projects I am using to stay busy through this crazy time we are experiencing. It is based on a simplified version of the IEC444 fixture as implemented by AE1C.
One of the suggested fixtures is found on github, in the project by Battosai42. In the supporting hardware for his amcp program, the impedance matching is done with a pi attenuator as shown in the (crude) text schematic below. For simplicity, only the input side is shown. The output is a mirror image.

O----+-----( 66.2 ohms )-----+---------------------+------[DUT]--
| | |
| | |
(159 ohms) (14.2 ohms) (33 nF)
| | |
| | |
O---+-----------GND-----------+---------------------+----------------

Now to the point of this. The capacitor value (33nF/0.033uF) seems very large for use at all but very low frequencies. Virtually all of the similar designs lack this capacitor entirely. So am I missing something? I have seen nothing of this in other designs discussed on the web, in QEX, or other discussions. Perhaps someone here can enlighten me, as well as the group.
Thanks for your consideration.
73,
Bill
P.S.
As soon as the design is complete, I plan to post the files here and make the boards available through OshPark.
 

On Sat, May 9, 2020 at 11:34 AM, Bill Buoy wrote:

... "Now to the point of this. The capacitor value (33nF/0.033uF) seems very large for use at all but very low frequencies. Virtually all of the similar designs lack this capacitor entirely. So am I missing something?" ....
=============================================
I posted about the ACMP Xtal testing project using the NanoVNA at Github ( ) a while back. Last change was about 4 month's ago so not much progress since then.
According to the schematic posted at the Github (see attached) the fixture presents a capacitive load of 19 pf with the values used.

- Herb

testfixture.png

 

It appears that this fixture is specific to the crystal under investigation. The fixture is providing a total load R of 25 ohms series from each side. Is possible if you transform the shunt C to its series equivalent from each side of the fixture, the total C load is 19 pF. A check on this requires knowing the test frequency. Seems like a stretch but you need to check if that is the case by transforming the network from its parallel to series form.

If you desire to build a general purpose test set for xtals, and obtain their parameters, this fixture is inappropriate. You can stay in the 50 ohm system or to improve measurement accuracy, transform 50 ohms to a lower Z. Either 12.5 or 6.25 ohms is done with a transformer. Hence loss, S21 will get you Rs from a voltage divider and loaded Q is obtained from a 3 dB BW measure, hence unloaded Q is obtained from the 20 log ratio of Q's and loss. The motional L and C parameters as well the holder C are obtained from some additional freq response measures.

Alan
 

My guess is the 33 nF is 33 pF, a typo. Transfer to a series may give 39 pF and hence the double loaded C is 19 pF. Again the fixture is specific and not generic test set.
 

I use a crystal test fixture that I built for the Poor Ham Scalar Network Analyzer (PHSNA) and I used it very much with good success with the nanoVNA also. It consist on two 50 to 12,5 Ohm broadband transformers with resistive attenuators at both ends to insure correct loading and adequate drive level for the crystal. It could be useful to have the capability of replacing the crystal with a variable resistor, either with a switch or a plug in resistor, to find the equivalent series resistance, but it could be calculated with the appropriate functions included in the the software / firmware without that resistor. I have been talking with the nanoVNA Saver author about this and he plans to include specific crystal measurement functions in a future version.

The PHSNA group is plenty of these test fixtures. An example is here: /g/PHSNA/files/Crystal%20Test%20Fixture/Crystal%20Test%20Fixture%20Overview.pdf

I built mine without the optional amplifier and without the variable resistor also, so I did it without a PCB, just the input and output connectors, a socket for the crystal, the transformers and the two Tee pads. If anyone is interested I can upload info and pictures to the group.

Regards,
Ignacio? EB4APL

El 10/05/2020 a las 2:20, alan victor escribió:
It appears that this fixture is specific to the crystal under investigation. The fixture is providing a total load R of 25 ohms series from each side. Is possible if you transform the shunt C to its series equivalent from each side of the fixture, the total C load is 19 pF. A check on this requires knowing the test frequency. Seems like a stretch but you need to check if that is the case by transforming the network from its parallel to series form.

If you desire to build a general purpose test set for xtals, and obtain their parameters, this fixture is inappropriate. You can stay in the 50 ohm system or to improve measurement accuracy, transform 50 ohms to a lower Z. Either 12.5 or 6.25 ohms is done with a transformer. Hence loss, S21 will get you Rs from a voltage divider and loaded Q is obtained from a 3 dB BW measure, hence unloaded Q is obtained from the 20 log ratio of Q's and loss. The motional L and C parameters as well the holder C are obtained from some additional freq response measures.

Alan

--
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.
 

Wouldn?t it rather be 33 pF, the standard parallell load for FT243 crystals?

73 de Jose, CO2JA

PS: In russian 33 пФ is 33 pF... just in case...

On 5/9/20, Kurt Poulsen <kurt@...> wrote:
Hi Bill
Sound as an interesting project. I visited the gitbug and found not much
description about the text fixture. I suppose you found it via KiCad. It
must be a bug it will not work up 150MHz as claimed.
I did such an test fixture years ago
Test Fixture for DG8SAQ VNWA and N2PK
VNA.pdf
And calibrated at the DUT position not and not as I think the gitbug
application does in the input/output side which introduces error as the two
attenuators are included and not without amplitude variation and phaseshift
variation over the frequency range. However is the application (as said
there is not much info how to use the application) is measuring S21 only and
calibration done by a shorting wire substituting the DUT then attenuators
are waved and part of the calibration. Then watch out as isolation should
not be done neither SOL.
That is what I can conclude but let me hear what you experience or have you
found a usage guide for the gitbug application.
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af Bill
Buoy
Sendt: 9. maj 2020 20:35
Til: [email protected]
Emne: [nanovna-users] Test fixture question

I am building a test fixture for use with my nano-VNA. The primary purpose
is to test/characterize a group of crystals for use in a crystal ladder
filter in a couple of receiver projects I am using to stay busy through this
crazy time we are experiencing. It is based on a simplified version of the
IEC444 fixture as implemented by AE1C.
One of the suggested fixtures is found on github, in the project by
Battosai42. In the supporting hardware for his amcp program, the impedance
matching is done with a pi attenuator as shown in the (crude) text schematic
below. For simplicity, only the input side is shown. The output is a mirror
image.

O----+-----( 66.2 ohms )-----+---------------------+------[DUT]--
| | |
| | |
(159 ohms) (14.2 ohms) (33 nF)
| | |
| | |
O---+-----------GND-----------+---------------------+----------------

Now to the point of this. The capacitor value (33nF/0.033uF) seems very
large for use at all but very low frequencies. Virtually all of the similar
designs lack this capacitor entirely. So am I missing something? I have seen
nothing of this in other designs discussed on the web, in QEX, or other
discussions. Perhaps someone here can enlighten me, as well as the group.
Thanks for your consideration.
73,
Bill
P.S.
As soon as the design is complete, I plan to post the files here and make
the boards available through OshPark.
 

Hi Herb
Thank you for the schematic
As alan victor mentioned in the next comment here, the test fixture is in not to recommend as a general test fixture designed for one specific frequency as the Shunt C to the crystal will be frequency dependent. What this fixture is trying to accomplish is to measure the antiresonans frequency (the high one) and the series resonance and derive the crystal parameters thereof knowing the circuit components and including these in the calculation.
I have done a lot of crystal measurement with the VNWA which have an function to measure the Rs, Ls Cs and Cp and deliverer the result in a spreadsheet for as many XTAL you like. No need for a 15 ohm test fixture at all just plain 50 ohm input output sources. For the VNWA you can use S11 or S21 as you wish. In general for S21 it is a good idea to insert 2x 10dB attenuator attached directly to the input and output of the test adaptor. I would be interesting to see how this application could be modified to cope with this setup.
For high precision the VNWA has a feature to measure the crystal using S21 12 term error correction but that is only if you own a VNWA
I have made a 50 ohm environment test adaptor Version 5 as seen on the images allowing to add shunt (e.g. 30pF) and series components L or C to shift the series resonance slightly.
I also made an even more universal version 5. If anyone interested I have Eagle Layout and gerber files so PM me to kurt@... and I can send it
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af hwalker
Sendt: 10. maj 2020 00:46
Til: [email protected]
Emne: Re: [nanovna-users] Test fixture question

On Sat, May 9, 2020 at 11:34 AM, Bill Buoy wrote:

... "Now to the point of this. The capacitor value (33nF/0.033uF) seems very large for use at all but very low frequencies. Virtually all of the similar designs lack this capacitor entirely. So am I missing something?" ....
=============================================
I posted about the ACMP Xtal testing project using the NanoVNA at Github ( ) a while back. Last change was about 4 month's ago so not much progress since then.
According to the schematic posted at the Github (see attached) the fixture presents a capacitive load of 19 pf with the values used.

- Herb

PCB side.jpg
Comp_side.jpg
X-Tal Adaptor_V5_Schematic.pdf
X-Tal Adaptor_V5_Schematic.pdf
X-Tal Adaptor_V5_layout.pdf
X-Tal Adaptor_V5_layout.pdf

 

Your explanation agrees with my logic... There is a mistake in the
drawing attached, the 33 pF capacitors appear marked 33 nF, not 33 pF.
That?s it.

On 5/9/20, hwalker <herbwalker2476@...> wrote:
On Sat, May 9, 2020 at 11:34 AM, Bill Buoy wrote:

... "Now to the point of this. The capacitor value (33nF/0.033uF) seems very
large for use at all but very low frequencies. Virtually all of the similar
designs lack this capacitor entirely. So am I missing something?" ....
=============================================
I posted about the ACMP Xtal testing project using the NanoVNA at Github (

) a while back. Last change was about 4 month's ago so not much progress
since then.
According to the schematic posted at the Github (see attached) the fixture
presents a capacitive load of 19 pf with the values used.

- Herb
 

The original circuit from the service manual of the HP device.

Кварцы.箩辫驳

 

Now I wonder: How do I extract the crystal equivalent model parameters
with this fixture? Does anyone have a link to the measurement method
explanation and perhaps associated software?

Jose, CO2JA

On 5/11/20, Aleksander Shalygin <shalygin@...> wrote:
The original circuit from the service manual of the HP device.
 

Hola José,

The crystal parameters can be calculated as explained in the PHSNA User's Guide Appendix I, by Nick Kennedy WA5BDU. All formulas are there.

It can be downloaded from here: /g/PHSNA/files/Windows%20PHSNA/PHSNA%20User%27s%20Guide%201r42.pdf

Here is an extract of a message that Nick sent to me when we were talking about crystal measurements with the nanoVNA:

QUOTE

I use this formula for the motional capacitance:

Cm = (f1 - f2)/(2*PI*f1*f2*Rt)

Where f1 and f2 are the upper and lower 3 dB points, so you could just say
BW for f1 - f2.

In the denominator, f1*f2 is not far from fc^2 where fc is the resonant
frequency of the crystal so you can use fc^2 if you like.

Rt is the total resistance of the circuit. So assuming you have a 50 ohm
generator plus 50 ohm detector and you measured the loss resistance of the
crystal at resonance at 55 ohms, your value is 155 ohms.

Having found Cm, Lm is the value of inductance that has the same reactance
as Cm at the resonant frequency. But a formula for that is:

Lm = 1/(4 * PI^2 *f^2 * Cm)

Where f is just the resonant frequency of the crystal.

You'll probably wind up with Cm in femtofarads and Lm in millihenries.

All of this depends of the resolution of the VNA. Being able to get down to
1 Hz resolution can help. I tried one crystal and didn't think I got good
results. But in the nanoVNA notes by Wes Hayward linked here recently, he
seemed to get some good crystal measurements. And Hayward is the guru of
crystal measurements although he wasn't deriving motional parameters in his
paper.

In doing crystal measurements, people often use transformers or resistive
pads to put the crystal in a 12.5 ohm environment, meaning the crystal sees
12.5 ohms looking in either direction. I'm not sure how advantageous this
is, but it does lower the BW and so would seem to make fine resolution even
more important.

For parallel capacitance, generally using a low frequency capacitance meter
like the AADE or eBay clones works.

UNQUOTE

The resistive pad approach is easier, but it could attenuate too much, it is something to be calculated and tested.

Saludos,

Ignacio, EB4APL

El 11/05/2020 a las 18:53, Jose Amador escribió:
Now I wonder: How do I extract the crystal equivalent model parameters
with this fixture? Does anyone have a link to the measurement method
explanation and perhaps associated software?

Jose, CO2JA

On 5/11/20, Aleksander Shalygin <shalygin@...> wrote:
The original circuit from the service manual of the HP device.




--
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.
 

On Mon, May 11, 2020 at 09:07 PM, EB4APL wrote:

The resistive pad approach is easier, but it could attenuate too much, it is
something to be calculated and tested.
I calculated a minimum loss resistor network, reducing the 2 x 14.8 dB from the HP fixture
to 2 x 11.4 dB.
Please see at

Attached are 2 photos of my test fixture.

73, Rudi DL5FA

nanoVNA_crystal_fixture_Top_DSC08223.jpg
nanoVNA_crystal_fixture_Bottom_DSC08224.jpg

 

"It can be downloaded from here: /g/PHSNA/files/Windows%20PHSNA/PHSNA%20User%27s%20Guide%201r42.pdf"

The group does not accept new logins, so the referenced file is not available.
73, Gyula HA3HZ

--
*** nothing is permanent only change ( ) ***

ScreenHunter_04 May. 12 09.59.png

 

I didn't know that, probably it is because the project is some years old. If anybody is interested I can send it via personal message.

73 de Ignacio EB4APL

El 12/05/2020 a las 10:07, Gyula Molnar escribió:
"It can be downloaded from here: /g/PHSNA/files/Windows%20PHSNA/PHSNA%20User%27s%20Guide%201r42.pdf"

The group does not accept new logins, so the referenced file is not available.
73, Gyula HA3HZ
--
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.
 

I'm not sure if the group does not accept new logins, new members should be approved by the moderator since the group was started. Nick Kennedy, which is the software author of the PHSNA, is also a member of this nanoVNA group and he is more informed than me.

73 de Ignacio EB4APL

El 12/05/2020 a las 13:18, EB4APL via groups.io escribió:
I didn't know that, probably it is because the project is some years old. If anybody is interested I can send it via personal message.

73 de Ignacio EB4APL


El 12/05/2020 a las 10:07, Gyula Molnar escribió:
"It can be downloaded from here: /g/PHSNA/files/Windows%20PHSNA/PHSNA%20User%27s%20Guide%201r42.pdf"

The group does not accept new logins, so the referenced file is not available.
73, Gyula HA3HZ
--
El software de antivirus Avast ha analizado este correo electrónico en busca de virus.
 

I've put the PHSNA User's Guide in a shared folder (see link). I think the
groups.io group still accepts members. The moderator is N5IB.

The User's Guide wouldn't be of much interest to most nanoVNA users, but I
did put in an appendix discussing measurement of crystal parameters in a
general way and that might be of interest to the current discussion.



I also compiled some information on the subject for a presentation I did
once. The slide show and an accompanying text document are linked here:



I hope some of this information might be useful.

73-

Nick, WA5BDU

On Tue, May 12, 2020 at 6:26 AM EB4APL <eb4apl@...> wrote:

I'm not sure if the group does not accept new logins, new members should
be approved by the moderator since the group was started. Nick Kennedy,
which is the software author of the PHSNA, is also a member of this
nanoVNA group and he is more informed than me.

73 de Ignacio EB4APL


El 12/05/2020 a las 13:18, EB4APL via groups.io escribió:
I didn't know that, probably it is because the project is some years
old. If anybody is interested I can send it via personal message.

73 de Ignacio EB4APL


El 12/05/2020 a las 10:07, Gyula Molnar escribió:
"It can be downloaded from here:
/g/PHSNA/files/Windows%20PHSNA/PHSNA%20User%27s%20Guide%201r42.pdf
"

The group does not accept new logins, so the referenced file is not
available.
73, Gyula HA3HZ
--
El software de antivirus Avast ha analizado este correo electrónico en
busca de virus.
 

Hi Nick,

thank you for sharing.
The last message from the PHSNA group was on December 28, 2019.
In the "Group Milestone" topic, Jim, N5IB, wrote that he wanted someone to take over the group, which they said didn't happen. (or just not visible to me) I tried to log in but no confirmation.
Thanks again for sharing the referenced material.

73, Gyula HA3HZ

--
*** nothing is permanent only change ( ) ***
 

I was thinking 33pf also. I have finished the layout and sent the boards off for prototype production. I should have boards back around the 23rd.
I’m finishing adding notes to the schematics. I’ll post those as well as the layout this week.

1 - 20 of 30