Talk:Crystal oscillator/Archive 1

Untitled
I think these could all be merged: Timing crystal, Quartz oscillator, Crystal oscillator. If no one objects, I will put them all under Crystal oscillator. And maybe Quartz clock, also. - Omegatron 20:15, May 4, 2004 (UTC)

I guess quartz clock and crystal oscillator should be separate articles. One is a specific application of the other. I will move some material from each to the other, though. - Omegatron 15:16, May 7, 2004 (UTC)

mechanical steel filter
From the article: '''For example, steel is very elastic and has a high speed of sound. It was often used in mechanical filters before quartz.'''

What is a mechanical filter, how was that one used, what has it to do with oscillators? Thanks, --Abdull 09:32, 19 August 2005 (UTC)


 * Any resonant system can be used as a component of an oscillator; simply add an amplifier, a feedback loop, and bit of noise to get things rolling.


 * Atlant 22:09, 27 April 2007 (UTC)

PhysicistQuery 17:24, 20 July 2007 (UTC)The most common mechanical filters were typically cylindrical rods of steel with wide sections and narrow sections along their length. Electroacoustic transducers were used to excite torsional motion, and the wide sections acted as the mass loads and the narrow sections served as the springs. Such filters were analogues of LC filters with all the capacitors grounded.

Only the quartz crystal or more
Is a crystal oscillator made up only by the single crystal and the feedback/amplifier, or does one still need additional components such as inductors and capacitors? --Abdull 10:14, 19 August 2005 (UTC)


 * I think all the other components would be considered part of the amplifier. - Omegatron 20:03, August 21, 2005 (UTC)


 * I agree -- the other components are considered part of the feedback/amplifier.
 * All the crystal oscillators I've designed (32 KHz to 20 MHz) are made of a single crystal, a single inverter, and 2 capacitors -- a total 4 components. To actually oscillate, some external power source needs to be connected to the inverter -- perhaps a linear regulator connected to a 9 V battery, with 2 more capacitors. No inductors are necessary. --76.209.30.239 06:53, 30 December 2006 (UTC)


 * It depends on whether you want the purest sine wave or not and whether you want to oscillate on an "overtone". While the crystals are obviously a high Q factor device, the simple crystal+two caps+amp structure doesn't necessarily put out a clean sine wave, free of harmonics. A nice LC tank in the output will help filter out those harmonics. And it can also be tuned to resonate at an overtone of the ordinary frequency that the crystal would otherwise produce.


 * Atlant 01:14, 31 December 2006 (UTC)

On a similar note, does anyone know what is the exact difference (if any) between a crystal oscillator and a crystal resonator. Are they synonyms, or does one imply additional circuitry or some other difference? Mattopia 09:24, 27 April 2007 (UTC)


 * To me, a "resonator" is more analagous to "filter" than an "oscillator". A tank circuit, for example, is a resonator. You can certainly use a resonator to make an oscillator, though. (Others' opinions may, of course, vary.)


 * Atlant 11:47, 27 April 2007 (UTC)


 * I agree. Crystals are used in electronics for two purposes; generating a signal, and filtering a signal.  To generate a signal, the crystal is used in a 'crystal oscillator' circuit.  The terminology for a crystal used as a filter is more unsettled.  Often the term 'crystal resonator' is used to distinguish a crystal used as a filter.  Unfortunately in physics the term 'crystal resonator' sometimes refers to a third use, as a transducer to receive or transmit mechanical vibrations. -- Chetvorno TALK 20:25, 10 May 2008 (UTC)

What does it do exactly?
If a 1MHz crystal oscillator is put in series with a 3v battery will the voltage in the circuit jump from 3v to 0v at a frequency of 1MHz??? Wolfmankurd 17:25, 28 April 2007 (UTC)


 * Basically, yes, if by "oscillator" you mean the entire circuit with crystal, amplifier, etc. See the above section. — Omegatron 17:49, 28 April 2007 (UTC)
 * Well it looks like [[Image:Crystal_oscillator_4MHz.jpg|100px]]. not one of those 4 pin ones. Wolfmankurd 00:56, 29 April 2007 (UTC)
 * That's a crystal, not an oscillator. The oscillator is the entire circuit containing a crystal, amplifier, and some other components. — Omegatron 01:29, 29 April 2007 (UTC)
 * Okay Thank you. Wolfmankurd 11:06, 29 April 2007 (UTC)
 * This is what I thought as well. Isn't the article title misleading then, in that the component is a crystal but not a crystal oscillator? A crystal oscillator would be the entire circuit that oscillates the signal??
 * The problem is that the terms crystal and quartz crystal have a more general meaning (in geology and chemistry), so the definition of the term 'crystal' in electronics was put in this article. Also, the term 'crystal' originated because historically they were made of quartz crystal, but modern resonators called 'crystals' can be made of other materials, and may not even be crystalline.  So 'piezoelectric resonator' is a more accurate term, and suggestions have been made (see above discussion) that the content on the component 'crystal' be split off into an article of that name.  -- Chetvorno TALK 10:25, 2 March 2009 (UTC)

series resonant oscillator
The Pierce oscillator article says that crystals tuned to a particular "parallel resonance frequency" are tuned using a Pierce oscillator circuit.

So ... what about crystals tuned to a particular "series resonance frequency"? What kind of oscillator circuits ( Category:Oscillators ) resonates at the series resonance frequency? And which one(s) in particular is(are) the one(s) used by crystal manufacturers? --76.209.28.72 05:43, 13 June 2007 (UTC)

PhysicistQuery 17:50, 20 July 2007 (UTC)If you look at the load presented by the oscillator to the crystal under steady-state conditions, it will consist of a negative resistance in series or in parallel with a reactance. The negative resistance will provide the energy needed to overcome the losses in the crystal; the reactance shifts the resonant frequency.

At the present time, the most common configuration for "simple" oscillators is the Pierce configuration, largely because all its elements can be integrated, and it naturally provides an output that is approximately symmetrical between the rails. This makes it convenient for a simple logic buffer to provide a rail-to-rail output suitable for driving digital systems. The reactance of the Pierce oscillator is naturally capacitive, and this is why the crystal is specified with a capacitive load (a pedantic point, but the circuit configuration of a Pierce circuit essentially provides negative impedance and capacitance in series - the "parallel" here refers only to a historic measurement method for the crystal that used a parallel capacitor). The Pierce oscillator can be modified using an external inductor in series with the crystal so that it is operated at series resonance, but this is seldom desirable except when it is necessary to tune the crystal.

An 'ideal' cross-coupled oscillator would ideally operate at series resonance, but in practice stray capacitances are inverted such that the majority of integrated circuit designs provide an inductive load to the crystal. If these oscillators are allowed to limit, the effective series inductance can become quite large.

xtal interaction
"Each new distinct crystal source needs to be rigorously justified, since each one introduces new, difficult to debug probabilistic failure modes, due to multiple crystal interactions, into equipment."

Thats a new one on me. What is the failure mode, in what way do the xtals interact? Tabby (talk) 23:39, 22 December 2007 (UTC)

20.000 Mhz Use?
Does anyone know what 20MHz crystal oscillators are used for? If so, please add to the table. Thanks. Pi.1415926535 (talk) 15:05, 20 April 2008 (UTC)

20MHz crystals are used for 10Mbit/s ethernet, which uses a 20MHz clock and Manchester encoding. I've added it to the table. Glenn Anderson (talk) 00:50, 7 January 2009 (UTC)

why are 33.33 MHz and 40 MHz mentioned but not on the list? what are they for? 09:26, 2 March 2009 (UTC)

Repeat of references
The link, as well as the associated title, in reference 4 is the same as that of reference 2. The only difference is that reference 4 provides more title detail. This artificially inflates the number of references. MVD (talk) 17:03, 23 March 2009 (UTC)

Stale link in references
The hyperlink for the article The Evolution of the Quartz Crystal Clock by Warren A. Marrison has gone stale. The corrected link is as follows:

http://www.ieee-uffc.org/main/history.asp?file=marrison

I don't know how to edit the list of references, hopefully some enterprising admin will pick up this torch and run with it. —Preceding unsigned comment added by 72.13.210.165 (talk) 03:08, 17 April 2009 (UTC)

13.248 Mhz
Used in VCXO for DSL applications. —Preceding unsigned comment added by 24.205.237.171 (talk) 21:41, 22 October 2009 (UTC)

Enough frequencies
You can get a quartz crystal ground for any frequency. The Digi Key on-line catalog lists more than 400 frequencies. The listing isn't very useful - there's a number of reasons to pick any particular freqeuncy and there's no general significance in any particular numeric value, let alone a list of hundreds. It's of no encyclopediac value to say the Binford 6100 used a 16.123 MHZ crystal while the sucessor model used a 16.321 MHZ clock unless this detail was actually significant. --Wtshymanski (talk) 21:02, 7 December 2009 (UTC)
 * Instead of robotically listing even more frequencies, which has about zero value in my opinion, how about a discussion of the different cuts of crystal and their varying temperature characteristics? AT cut, BT cut, etc. --Wtshymanski (talk) 03:45, 23 December 2009 (UTC)

Agree that the frequency list is not interesting as it stands. A list of "standard application" frequencies that gave a bit of background might be - but should probably be a separate page, and would include frequencies that are not directly related to crystals. BTW, the references to 32 kHz are perhaps less irritating than including the irrelevant precision - but maybe they should merely be described as "watch crystals" except in a single place where 32768Hz is given as a commonly used frequency for these? PhysicistQuery (talk) 12:19, 22 December 2010 (UTC)

Circuit notations and abbreviations
What does it mean: SPXO? Regards --Elcap (talk) 17:04, 14 September 2010 (UTC)
 * It doesn't mean anything, it's a brand name or product identifier. --Wtshymanski (talk) 14:11, 15 September 2010 (UTC)
 * It means "Simple Packaged Crystal Oscillator" PhysicistQuery (talk) 11:41, 22 December 2010 (UTC)

Crystal Cuts
AT cut crystal: The description of frequency-temperature shape of the AT cut as "sine" is misleading. Third order would be correct. This seems to be a case where a picture would be worth a thousand words. PhysicistQuery (talk) 11:56, 22 December 2010 (UTC)

SC cut crystal: The statement "coupled modes are worse than the AT" does not seem to be borne out in practice. Indeed, the performance of the MCXO relies on the excursions due to coupled modes being very small. Would the author perhaps be referring to the requirement to suppress the (usually more active) B-mode vibration at (about) 10% lower frequency than the wanted frequency? This is not at all the same thing as a coupled mode. BTW, coupled modes are not described/defined anywhere in the article, so this reference is anyway meaningless to general readers PhysicistQuery (talk) 11:56, 22 December 2010 (UTC) Similarly, the temperature sensitivity (third-order coefficient) of the SC is quite similar to the AT. Because of the high inflection temperature, however, the temperature sensitivity of an SC cut crystal can be made relatively small for use in an ovened oscillator. On the other hand, the frequency-temperature sensitivity for temperature ranges that are centred near "normal" temperature is considerably greater for an SC cut than for an AT. PhysicistQuery (talk) 12:04, 22 December 2010 (UTC)