Lpc2000

While looking at the schematic for the Keil MCB2130, I noticed that they 
put a 560K resistor in series with the 32KHz crystal.  This seems odd, 
the only reason I can see for a series resistance is to lower the 
current through a crystal circuit.

Anyone see this?   From what I can tell from the LCP2138 manual, only 
load capacitors are called for to pull the crystal on frequency.  
AFAICT, my 32KHz crystal is running fine on my board without the series 
resistor.


Look at: http://www.keil.com/mcb2130/mcb2130-schematics.pdf  to see what 
I mean.

TomW

-- 
Tom Walsh - WN3L - Embedded Systems Consultant
http://openhardware.net, http://cyberiansoftware.com
"Windows? No thanks, I have work to do..."
----------------------------------------------------

From: "Tom Walsh" <tom@...>
> While looking at the schematic for the Keil MCB2130, I noticed that they
> put a 560K resistor in series with the 32KHz crystal.  This seems odd,
> the only reason I can see for a series resistance is to lower the
> current through a crystal circuit.
>
> Anyone see this?   From what I can tell from the LCP2138 manual, only
> load capacitors are called for to pull the crystal on frequency.
> AFAICT, my 32KHz crystal is running fine on my board without the series
> resistor.

A deja vu, see the archives of this group...

The max power into a 32kHz Xtal is very limited, the resistor lowers the
actual power. I've had Xtals literallly break in a 5V CMOS oscillator when
that resistor was missing. And they were $10 in those days...

BTW: the series resistor must be between the output of the oscillator buffer
in the chip and the C-Xtal-C network, not in series with the XTAL itself! A
drawing error in the diagram, I'm sure.

The extra heating caused may cause the Xtal to drift, too.

Arie de Muynck

The 32KHz "tuning fork" crystals used in watches and microcontroller
sleep circuits have a maximum drive rating.  IIRC it is on the order
of 5 microwatts or less.  The resistor is not intended to have any
direct effect on the frequency of oscillation, although it does tend
to suppress any overtone oscillation tendancy. The purpose of the
resistor is to prevent overdriving the crystal.  If the crystal is
overdriven it's frequency accuracy will degrade over time at a rate
much faster than it is specified for. In extreme cases it could
actually crack, but that is not likely with a low voltage CMOS
oscillator.  In cell phone ASICs we always designed the crystal power
limiting resistance internal to the ASIC itself.  I wonder what Keil
found out concerning the crystal drive circuit in the 2130?

--Dave




--- In lpc2000@yahoogroups.com, Tom Walsh <tom@o...> wrote:
>
> While looking at the schematic for the Keil MCB2130, I noticed that
they 
> put a 560K resistor in series with the 32KHz crystal.  This seems odd, 
> the only reason I can see for a series resistance is to lower the 
> current through a crystal circuit.
> 
> Anyone see this?   From what I can tell from the LCP2138 manual, only 
> load capacitors are called for to pull the crystal on frequency.  
> AFAICT, my 32KHz crystal is running fine on my board without the series 
> resistor.
> 
> 
> Look at: http://www.keil.com/mcb2130/mcb2130-schematics.pdf  to see
what 

> I mean.
> 
> TomW
> 
> -- 
> Tom Walsh - WN3L - Embedded Systems Consultant
> http://openhardware.net, http://cyberiansoftware.com
> "Windows? No thanks, I have work to do..."
> ----------------------------------------------------
>

derbaier wrote:

>The 32KHz "tuning fork" crystals used in watches and microcontroller
>sleep circuits have a maximum drive rating.  IIRC it is on the order
>of 5 microwatts or less.  The resistor is not intended to have any
>direct effect on the frequency of oscillation, although it does tend
>to suppress any overtone oscillation tendancy. The purpose of the
>resistor is to prevent overdriving the crystal.  If the crystal is
>overdriven it's frequency accuracy will degrade over time at a rate
>much faster than it is specified for. In extreme cases it could
>actually crack, but that is not likely with a low voltage CMOS
>oscillator.  In cell phone ASICs we always designed the crystal power
>limiting resistance internal to the ASIC itself.  I wonder what Keil
>found out concerning the crystal drive circuit in the 2130?
>
>  
>
Either that, or their designer has had problems with excessive heating 
of the crystals in past designs and simply carried over a tried-n-true 
solution.

I did not consider that the colpitts circuit would do that, overheat the 
crystal, but I guess it is possible.  You always learn something new.  ;-)

I'll have to look at the specs to see if over-driving the crystal may 
indeed be a potential problem...


TomW


-- 
Tom Walsh - WN3L - Embedded Systems Consultant
http://openhardware.net, http://cyberiansoftware.com
"Windows? No thanks, I have work to do..."
----------------------------------------------------