[Lowfer] TCXOs, free to good homes

Stewart Nelson [email protected]
Sat, 8 Mar 2003 14:03:14 -0800


Hi all,

I recently acquired some temperature compensated oscillators that
might be useful for improving the frequency stability of your beacon
or receiver.  There are about seven left over; if one would help your
rig, let me know and I'll mail it out.

Obviously, if your present oscillator is GPS-locked, Rb, or oven
controlled, this is not for you.  But if you are using a cheap
clock module or a homebrew crystal circuit, and would like better
stability, then read on.

These units are KTXO-18 DIP modules by Kyocera, intended for use in
cellular handsets.  Comments below are based on the manufacturer's
spec, as well as my tests on one part from the lot.

Nominal frequency is 12.8 MHz.  Output is AC coupled, but can drive
HCMOS or LSTTL by just connecting two resistors to the output.

There is a manual trimmer but no electrical frequency control.
Kyocera spec is +/- 3 ppm minimum adjustment, but the actual range is
about +/- 20 ppm, in the usual 180 degrees, so it takes a while to set
it better than 0.1 ppm.  Before adjustment, the test unit measured
12.8000121 MHz at 20 deg. C.

Frequency versus voltage is pretty good.  Spec is +/- 0.3 ppm over a
+/- 5% change in supply.  Actual shift was only 1 Hz for 200 mV
change.  I would guess that most systems would exhibit a supply
voltage change of less than 10 mV during a half hour transmission;
this would be less than 1 mHz at LF.

Temperature stability, however, is not so great.  Claimed is +/- 3 ppm
from -20 to +60 C.  But the curve has at least two inflections, and
the slope is about 0.05 ppm per degree at say, 15 C.  If you consider
decent lowfer stability to be 0.01 Hz over a half hour transmission,
then the temperature would need to change no more than 2 C per hour.
Most indoor living spaces conform, but many shacks are in locations
that don't.

Long term drift spec is 1 ppm per year.  The test unit has been
connected to a counter for the past month.  During the first week,
frequency fell about 0.1 ppm, and then started to rise.  It's
presently increasing about 0.01 ppm per week.

I found an interesting PIC-based timer that uses this oscillator at
http://www.picfun.com/equipj39.html .  The above page has a spec,
picture, and schematic showing how to drive a PIC.  If you have Win 2k
or XP, there should be no trouble viewing the page, otherwise you may
need Japanese fonts or language pack.

I was unable to find the official spec for this part on the web.  It
is obsolete, since current handsets use surface mount parts.  However,
I believe that the KT18 series is electrically almost identical.  At
http://global.kyocera.com/frame/product/electro/ec/english/tmn/index.html
you can download a TCXO catalog in English.

Let me know if you would like one of these parts.

73,

Stewart KK7KA