Newsgroups: comp.robotics
Path: brunix!news.Brown.EDU!noc.near.net!MathWorks.Com!europa.eng.gtefsd.com!howland.reston.ans.net!torn!newshub.ccs.yorku.ca!newshub.ariel.cs.yorku.ca!cs911225
From: cs911225@ariel.cs.yorku.ca (KEN E WILLMOTT)
Subject: Re: A very STABLE 40khz generator (was IR receivers)
Message-ID: <CJMysw.LDv@ariel.cs.yorku.ca>
Sender: news@ariel.cs.yorku.ca
Organization: York University, Dept. of Computer Science
References: <sasrer.758213811@cinnamon> <CJFxE5.A22@ariel.cs.yorku.ca> <8hB5UsS00YUo8ACpNQ@andrew.cmu.edu>
Date: Fri, 14 Jan 1994 20:12:32 GMT
Lines: 69

>From: "Richard M. Moore" <rm52+@andrew.cmu.edu>

>>You're like me and like this little gem, right?
>>Okay, then, how about a VERY, VERY stable circuit? :-)
>>You really should include a resistor as shown in the diagram
>>below, don't ask why, just do it!

>>              12 >----------------------+
>>                                 740pf  |    (you may be able to obtain
>>              11 >-------+----+---|(----+     a resonator with builtin
>>                         |    |         |     capacitors and three leads)
>>                640khz  ---   \         |
>>                 res.    O    / 1Mohm   |
>>                        ---   \         |
>>                         |    |         |
>>              10 >-\/\/\-+----+---|(----+
>>                   330ohm        740pf  |
>>             GND >----------------------+

>Why is so that you don't overdrive the resonator or crystal.


No. It is there for three primary reasons:

1) Adds some additional phase shift in conjunction with the
   capacitor to ground.

2) Eliminates higher harmonics of the drive signal, in order
   to prevent spurious oscillations.

3) Isolates the reactive load from the gate output.


>The
>extra resistor acts as a voltage divider with the effective series
>resistance of the crystal to reduce the drive voltage.

No. It acts as a voltage divider with the input reactance of
the entire feedback network.


>This is more of a problem with CMOS than with TTL circuitry since
>CMOS swings rail-to-rail.

If so, its not by more than a factor of 2. What is the source
of this hypothesis?


>For a reference on this, read the Motorola "High-Speed CMOS Logic Data"
>data book, section 4 on design considerations (page 25 in my edition).
>There is a section that talks about Pierce crystal oscillators, which
>is the name of the circuit shown below.

Well, I don't have this one, so I can't comment on it, except to say
that it must be skimpy and misleading, or else you haven't read it
in any great detail. I'm using National Semiconductor's application
note AN-340, "HCMOS Crystal Oscillators". No mention of maximum
drive levels for crystals is made anywhere in it.

The rule of thumb for selecting the resistor, is to make its
resistance equal or a little greater than the reactance of the
capacitor to ground at the frequency of operation. If you are
involved in a design which is expected to be permanant and work
reliably, I strongly recommend the above document. Details of
component selection are well covered there.

	-Ken Willmott


