Crystal oscillator and Schmitt Trigger Inverter ?
Alexandre Beaulieu
I'm building a 32 MHz oscillator using a Pierce Configuration
but in my configuration, my inverters have to be unbuffered
inverters.Do you have any idea why I'm not able to use an Schmitt
Trigger Inverter ?It does not seem to work with a crystal in parallel
mode. I would like to work with Schmitt Trigger Inverters as they are
less affected by noise and change in temperature but my circuit is not
working with these kind of inverters. Thanks
Jim Thompson
alexandre...@space.gc.ca (Alexandre Beaulieu),
In Newsgroup: sci.electronics.design,
Article: <691b126e.02041...@posting.google.com>,
Entitled: "Crystal oscillator and Schmitt Trigger Inverter ?",
Wrote the following:
As I've harped for months... you *must* use unbuffered inverters for
crystal oscillators. Everyone seems to forget, crystal oscillators
are *analog* !!
George R. Gonzalez
run as a crystal oscillator. The "noise rejecting" feature is what's
killing you-- the crystal at startup just has random noise as the only
signal--
the schmitt trigger is ignoring all of this.
Maybe if you gave the circuit a good ZAP it *might* start up, or maybe not.
"Alexandre Beaulieu" <alexandre...@space.gc.ca> wrote in message
news:691b126e.02041...@posting.google.com...
Sparky
"Alexandre Beaulieu" <alexandre...@space.gc.ca> wrote in message
news:691b126e.02041...@posting.google.com...
To get the oscillator to work, you are using the unbuffered invertors as a
(pseudo) 'linear' amplifier by biasing it into its linear part of the
transfer characteristic.
The Schmitt trigger inverter by its very nature has hysteresis, and by
definition does not have a linear region into which you can bias it. (crap
sentence but you get the drift)
Sparky
Spehro Pefhany
> You're going to have a heck of a time getting a Schmitt trigger to
> run as a crystal oscillator. The "noise rejecting" feature is what's
> killing you-- the crystal at startup just has random noise as the only
> signal--
> the schmitt trigger is ignoring all of this.
>
> Maybe if you gave the circuit a good ZAP it *might* start up, or maybe not.
There are crystal oscillators that use a ST, such as the one in the 8049,
which works okay at ~11MHz with HC-49 crystals. It starts up in a
relaxation oscillator mode at ~50kHz and then goes to crystal oscillation.
Since yours isn't working, I guess this method won't work with your
particular crystal and chip and capacitors and ambient temperature. ;-)
Since most modern CMOS Pierce oscillators use a linear amplifier
configuration, you might want to take that as a hint.
Best regards,
--
Spehro Pefhany --"it's the network..." "The Journey is the reward"
sp...@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com
9-11 United we Stand
Joseph Legris
For the sake of the OP:
Schmitt triggers usually behave much like inverters in digital circuits,
but they have a dead zone in the middle where too small an input voltage
variation yields no output. That's why they are so immune to noise.
Unfortunately, the feedback signal in a crystal oscillator can be very
small, especially at start-up. Hence, no output. As Jim says, you must
consider the analog characteristics of the device when you stray from
standard digital operation.
Regards,
Joseph Legris
Mike Monett
[...]
> There are crystal oscillators that use a ST, such as the one in the 8049,
> which works okay at ~11MHz with HC-49 crystals. It starts up in a
> relaxation oscillator mode at ~50kHz and then goes to crystal oscillation.
Speff, can you give more information on this startup mode? Do you have
any links or product info that describes it in more detail?
What is a "ST"? and are you referring to the Intel 8049 microprocessor?
Best Regards, Mike
Spehro Pefhany
> Spehro Pefhany wrote:
>
> [...]
>
>> There are crystal oscillators that use a ST, such as the one in the 8049,
>> which works okay at ~11MHz with HC-49 crystals. It starts up in a
>> relaxation oscillator mode at ~50kHz and then goes to crystal oscillation.
>
> Speff, can you give more information on this startup mode? Do you have
> any links or product info that describes it in more detail?
Intel AP-155. One of the subject they discuss is making sure that the
relaxation oscillations don't dominate, by choice of asymmetrical
capacitor values or by tacking on a pullup on the input. The feedback
resistor value must have been relatively low in those early processors.
> What is a "ST"? and are you referring to the Intel 8049 microprocessor?
Schmitt Trigger, and yes, the MCS-48 (NMOS) series of microcontrollers.
Bill Sloman
You can use Schmitt Trigger inverters, but you will wish you hadn't -
I've done it,and posted appeals for help in this user group when the
oscillator took to starting up at frequencies other than the crystal
frequency.
The problem is that you have to set up your oscillator with an even
number of inverters (more than zero) so that it has to oscillate, and
it is then tricky to get it to oscillate at the crystal frequency, and
only the crystal frequency.
I got a single proof-of-principle circuit working by carefull choice
of capacitors, but I wouldn't dream of trying it for a production
circuit.
---
Bill Sloman, Nijmegen