Not quite as simple as the PTC, an alternative may be:

http://shop.kuhne-electronic.de/kuhne/en/shop/accessoires/crystal-heater/Precision+crystal+heater+40%C2%B0+QH40A/?card=724

No it probably doesn't hold the crystal at it's optimum turn over
temperature, but it will keep the temperature of a crystal approximately
constant especially on a windswept hilltop.

On 05/06/2017 09:35, Ulf Kylenfall via time-nuts wrote:

--
Stephen Tompsett

There's a few 'OCXO' designs out there, I'm not qualified to comment on the
timenutty quality of them but someone else mentioned Hans Summers offerings
and I would offer Roman Black's simple design (if it's not been mentioned
already):

http://www.romanblack.com/xoven.htm

I've no idea if it's useful but it's ridiculously simple to implement.

Both Hans' and Roman's designs are 'on the list' of things to try for
myself at some point.

On 5 June 2017 at 09:56, Stephen Tompsett stephen@tompsett.net wrote:

--
Clint.

No trees were harmed in the sending of this mail. However, a large number
of electrons were greatly inconvenienced.

Hi

The posistor approach to heating a crystal was originally pioneered in Russia. Morion was
building vacuum insulated / PTC controlled OCXO’s long before the PTC parts started showing
up more generally in the 1970’s.

Bob

Here is a national new-technology of the art crystal oven from 1956:

http://tf.boulder.nist.gov/general/pdf/1642.pdf

Using the phase change properties of p-dibromobenzene it keeps temperature constant to 0.01C. It notes other organic compounds can be used for different temperature ranges.

Did this oven technology ever get beyond lab use and into the real crystal oven world?

I know in the past decade "thermowax" has been used in Honda lawnmower auto-chokes. I don't think it's ever supposed to be anything but solid, but it undergoes a phase change that causes it to expand by a large fraction.

Tim N3QE

Sent from my VAX-11/780

Wax is also used for thermostatic valves in engine cooling systems and
domestic heating systems.

On Mon, Jun 5, 2017 at 1:13 PM, Tim Shoppa tshoppa@gmail.com wrote:

Look at the dates on the uses of these exotic temperature controllers.
Organic compounds and positive temp coefficients and so on.  All these were
used before the current era.  Today all you need is a reliable way to
measure the error between the crystals' current temperature and the set
point.

Those exotic methods were good back when controllers were analog devices
build with op-amps and precision resisters and every math operation cost
you one more op-amp.  Today we can do a million floating point operations
per second for the cost of one kind good op-amp.

The trick with using a uP and it's built-in A/D converters is scale.  You
want the limited 10-bits of revolution to fall over the operating range
which is very narrow, like 1C.  Anything outside of that is either 0000 or
1111 and only seen at start-up.,  So at start up the the controller is on
"bang-bang" mode then later you have milli-degree resolution over your 1C
range.  Basically you are measuring noise. but your $2 uP can take 100,000
measurments per second and putt tour a digital filter.

Today we can do things the old-time designers even as recent as the 1980's
would not even dream of doing and it cost under $10.

On Mon, Jun 5, 2017 at 1:35 AM, Ulf Kylenfall via time-nuts <
time-nuts@febo.com> wrote:

--

Chris Albertson
Redondo Beach, California

Chris wrote:

That's all that's ever been needed.  But it is devilishly difficult to
measure the actual quartz temperature, or even to find a good proxy that
is easier to measure.

There is a fair body of published research on these topics, including
Rick's (et al.) work on zero-gradient ovens.

Keep this in mind when someone says they are controlling the "oven
temperature" to 0.001C (or even 0.1C).  They are measuring something,
and may even be holding whatever it is "constant" within fairly tight
tolerances.  But they have no idea what the quartz temperature is, and
no way to know with precision the relationship between the measured
temperature and the actual quartz temperature.

Some years ago, I consulted for a research group that was using a number
of non-contact technologies to measure the temperature of oscillating
quartz crystals.  The results were promising, but there were some issues
with measuring the temperature (which is, essentially, quantifying tiny
random molecular motions within the crystal lattice) against the
background of the hugely greater macro motion of the vibating quartz.  I
never knew the final conclusions, nor am I aware of any systems designed
using these principles or methods.  But it is something to think about
if you really want a temperature-stable oscillator.

Best regards,

Charles

On Mon, June 5, 2017 5:38 pm, Charles Steinmetz wrote:

In most cases what you really care about is the stability of the
frequency, and the temperature of the crystal is just a proxy for that,
correct?
I thought there was some effect where different modes of oscillation
shifted by different amounts with temperature, and if you had two
oscillation circuits running from the same crystal but different modes,
you could use the shift in difference frequency between the two modes to
infer the temperature change.

Found a reference in the Vig tutorial:
S. Schodowski, "Resonator Self-Temperature-Sensing Using a
Dual-Harmonic-Mode Crystal Oscillator," Proc. 43rd Annual Symposium on
Frequency Control, pp. 2-7, 1989, IEEE Catalog No. 89CH2690-6.

As is shown in chapter 4, see “Effects of Harmonics on f vs. T,” the f
vs. T of the fundamental mode of a resonator is different from that of
the third and higher overtones.  This fact is exploited for
“self-temperature sensing” in the microcomputer compensated crystal
oscillator (MCXO). The fundamental (f1) and third overtone (f3)
frequencies are excited simultaneously (“dual mode” excitation) and a
beat frequency fb is generated such that fb = 3f1 - f3 (or fb = f1 -
f3/3). The fb is a monotonic and nearly linear function of
temperature, as is shown above for a 10 MHz 3rd overtone (3.3. MHz
fundamental mode) SC-cut resonator.

The graph shows a line with slope of around 80ppm/deg C.  Not sure what
that translates to in terms of what you could realistically measure and
use for frequency compensation.  I guess you could use that information to
either control an oven or just let the crystal run free and control a
synthesizer for the used output.

--
Chris Caudle

On Mon, Jun 5, 2017 at 3:38 PM, Charles Steinmetz csteinmetz@yandex.com
wrote:

How much does this matter?  What we measure is the ambient temperature
inside the insulated box that contains the crystal.  The assumption is
that given some time the temperature will be uniform. OK, so the assumption
is not 100% correct but lets say the crystal is held to a range of 0.1C
This is a reasonable goal for a home shop made controller.

Here is another question:
Lets assume we place the operating point on the flat part of the curve with
say a 1.0 C absolute error and can hold the relative temperature to 0.1C
What does this mean in terms of frequency.?

This is a "Poor Man's" oven.  So really the question is this:  I have a
$10 budget, shouldI blow half my budget on a better sensor or is the 75
cent part good enough.  Or is it worth buying a second 75 cent sensor so I
can detect a temperature gradient  With a poor man's budget, I think the
trick is to use a good size thermal mass, chunks of scrap meter are cheap.

--

Chris Albertson
Redondo Beach, California