What is the significance of frequency tolerance after matching for a Quartz crystal

**dyunasing** · 16th May 2019, 08:04

hi, all,

I am a bit confused by the concept of frequency tolerance of a quartz crystal.
I can always carry out a load capacitance matching exercise to make sure the crystal is working at its center frequency, right?
After matching, the crystal is working at 0 ppm?
Then how does the frequency tolerance concept come in?
Lets say I have 2 quartz crystals, one has 20ppm frequency tolerance specified in the datasheet and the other 50ppm.
After both crystals go through the matching exercise at room temp, both crystal operates at their center frequency.
Are both parts equal in performance under room temp?

Dyu

**kripacharya** · 16th May 2019, 08:13

The keyword is 'temperature' and 'time'.
Tolerance is usually specified over a range of temperatures - not just a single one - like 25°.
Also you may adjust for zero today, but what about a month from now? A year on?

**dyunasing** · 16th May 2019, 08:17

hi,
My understanding is that 'temperature' drift is classified as frequency stability and 'time' drift is as aging.
Frequency tolerance is the initial tolerance.

**kripacharya** · 16th May 2019, 08:28

You may be right. Sometimes terminology is used differently, so you have to be careful in it's consistent use and interpretation.

**dyunasing** · 16th May 2019, 09:15

After both crystals go through the matching exercise at room temp, both crystal operates at their center frequency.
Are both parts equal in performance under room temp?
And i can just focus on the frequency stability and aging spec in selecting the suitable crystal?

**kripacharya** · 16th May 2019, 09:28

I'm no expert, but i would agree. However it came to mind - how do you adjust for both +ppm and -ppm. Inductor in one case, cap in the other?

**dyunasing** · 16th May 2019, 09:37

https://www5.epsondevice.com/en/info...es_e201302.pdf

**kripacharya** · 16th May 2019, 09:42

Also to consider is the stability etc etc of the passive element ( cap? ) used to tune it.

- - - Updated - - -

Originally Posted by dyunasing

https://www5.epsondevice.com/en/info...es_e201302.pdf

So that explains it all then :)

**dyunasing** · 16th May 2019, 10:12

My thinking is that since matching is a norm to make every part starting with 0 ppm at room temp, why the supplier bothers to put the frequency tolerance into the spec then.

**kripacharya** · 16th May 2019, 10:28

I can think of many reasons, including that I do not believe matching is a norm

**dyunasing** · 17th May 2019, 00:19

Any expert in related field would like to share?

**c_mitra** · 17th May 2019, 04:45

Lets say I have 2 quartz crystals, one has 20ppm frequency tolerance specified in the datasheet and the other 50ppm.

It is not meant that way. Say you have a crystal with a X MHz printed on it. In actual operation, it may be X+/- y with a 99% probability. The y is the commonly meant tolerance. In a datasheet you may find three values: typical, max and min. Usually you will consider the typical value while designing some item.

Quartz crystals are less sensitive to temperature and the centre frequency will change with temp because of the temp effect on mechanical dimensions and elastic properties.

Far more important is the loading; that may affect the actual frequency much more than the tolerances or temp coeff.

You can tune a crystal (to a small extent only) by changing the loading. If the crystal is operated in a higher harmonic (say 2nd harmonic) these effects become less.

After both crystals go through the matching exercise at room temp, both crystal operates at their center frequency....

If you are trying to match the frequencies, obviously one of them (perhaps both) are not running at their centre frequency (because of the tolerances).

Near the centre frequency, the crystal may act like an inductor or a capacitor (depending on which side of the peak it is currently working).

Both parts may be considered equal for routine applications (but certainly not for critical applications).

Consider the crystals used in clock ICs.

**dyunasing** · 17th May 2019, 08:43

I understand the issue of tuning tolerance that may not give me an exact center frequency.
But lets say this is being taken care of. They are both nicely matched.
Are they considered identical at the starting temperature where the tuning takes place?

I have 2 crystals on hand now.
Crystal 1: FT at 25 degree: +/-20ppm Frequency stability: +/-30ppm aging: 5ppm/yr Price: $1.5
Crystal 2: FT at 25 degree: +/-100ppm Frequency stability: +/-30ppm aging: 5ppm/yr Price: $1

I can tune both crystals to centre at 25 degrees. Now my question is can i safely use crystal 2 since i can save that 50 cents?
Thanks.

**kripacharya** · 17th May 2019, 12:49

Originally Posted by dyunasing

...
I can tune both crystals to centre at 25 degrees. .

Can you, really ?

**c_mitra** · 17th May 2019, 12:59

Now my question is can i safely use crystal 2 since i can save that 50 cents?...

That depends. What is the specific application? You must load the crystal as little as possible to get the max benefits from the frequency stability specs.

But I am bot an expert in this specific area.

What is the significance of frequency tolerance after matching for a Quartz crystal

Thread Tools

Search Thread