| Home | Terms of Use | Site Map | Contact Us |
IndustryCommunity.com > Electrical and Electronic Community > Analog Circuit Design Forum > Message
Main Menu

[ List Subjects ][ Main Page ]
[ View Followups ][ Post Followup ]

Subject: Re: Crystal circuit

Date: 08/17/00 at 9:09 PM
Posted by: John Dunn - Consultant
E-mail: ambertec@ieee.org
Message Posted:

In Reply to: Re: Crystal circuit posted by Alvonzo on 08/16/00 at 7:15 AM:

Hi, Alvonzo.

Robert's answer is essentially correct. Let me go a little further into a description and caveats.

The circuit uses an inverting gate, a device normally intended to run with digital inputs and outputs, as a linear
gain stage instead. The 10 Meg resistor tied from the gate's output back to its input sets up negative DC feedback
to accomplish this. When in its linear region, the gate has enough small signal gain to function as a linear (sort of)
amplifier which can be used to make an oscillator. In this case, AC feedback for setting up that oscillation consists
of the crystal and the two shunt capacitors to ground. Those capacitors are required for establishing the proper
phasing for oscillation.

Let me say that I truly dislike this circuit and I mean that personally!.

It is the same circuit Xilinx for example, has used in their gate arrays. Instructions read something like "set Rx as
required" to make the oscillator work, but I've made some singularly unsuccessful attempts at this.

The resistor Rx tempers the AC feedback with the intention that oscillation should only take place at the intended
frequency as determined by the crystal. Trying to establish a correct Rx is the part which has turned many heads
gray, including my own!

Therefore, to save my clients from the cost of hair dye, I always recommend that this circuit be avoided and that
a separate oscillator module be purchased instead.

Good luck.

John Dunn - President
Ambertec, Inc.

Follow Ups:

Post a Follow-up:


Message to Post:


1999-2001 Sunlit Technology Co., Ltd. All rights reserved.