richard@karlquist.com said:

That still leaves the temperature quirks of the DAC and amplifiers.

Has anybody put the DAC and all of the analog stuff inside the oven?  Seems
like an obvious idea so somebody has probably patented it.

--
These are my opinions.  I hate spam.

On Fri, July 22, 2016 2:15 pm, Hal Murray wrote:

I ran across some OCXO's with a DAC inside the oven a few years back.
They were 5MHz instead of the 10MHz I was looking for so I didn't buy them
at the time.  I have forgotten the vendor, but I seem to recall they had a
Maxim DAC inside the oven.

--
Chris Caudle

Hal Murray <hmurray@...> writes:

Seems

I am using a 5 MHz custom built design which has everything inside the
outer oven in a double oven.

This includes:

Precision regulators for DC supply.
24 bit DAC
Reference for DAC
All analog devices, resistors and capacitors after DAC.
All buffer amplifiers from oscillator.

Unit only has raw DC and I2C lines for inputs and 5MHz sine wave out at
3 volts p-p.

It also seemed the obvious way to go in the design.

Performance is difficult to measure with equipment on hand.  Waiting on
funding for experiment which includes equipment traceable to NIST.

On 7/24/2016 6:44 AM, James Flynn wrote:

IMHO, it isn't obvious that this is the way to go.
Nor is it a "free lunch" as it seems to be envisioned.
The last thing you want in an oven is a lot of
added thermal overhead, especially in a double oven
where you already fighting against running out of
temperature range.  If you want to ovenize everything
but the kitchen sink, put it in it's own oven that is
separate from the crystal oven.

I also don't like intermingling digital signals with the
analog oscillator signal.

Have you measured the thermal gain of your outer oven?
I suspect it's not much.  You could use inner oven
current draw as a proxy.

Rick

Richard (Rick) Karlquist <richard@...> writes:

The thermal overhead is quite small compared to the power required for
the outer oven to hold temperature.  80mW maximum as opposed to about 1
watt.

I did a comparison of the temperature rise of running the unit
essentially as a single, inner oven with the rest of the electronics
mentioned in the outer "box", and then turning on the outer heater.  The
Q from the inner oven was far and away the dominant factor.

Putting the DAC inside the outer oven was the obvious solution for me,
as opposed to putting them outside in ambient.  I did consider thermal
compensation, but again the outer oven would be able to keep things
within a fraction of a degree over the range of normal ambient in the
lab.

Separate ovens seems to be inviting noise getting into the signal lines
between the two ovens.

Not sure what you mean by "intermingling". There are separate ground
returns for the oscillator output (transformer isolated) and the digital
signals. There is a low pass filter between DAC and oscillator control
to minimize noise getting across. Also digital circuits and oscillator
have their own individual precision regulators.

I did a while ago and remember it was on the order of 50 - 100. But I
have changed the design somewhat and should do it again when the
experiment gets going.

Hi

A typical double oven design runs about 10 mw or so in the circuitry
in the inner oven. The “stuff” inside the outer oven likely doubles that number.
The rest of the circuit is put outside the outer oven to reduce heat rise. There
could be another 40 mw in that part of the circuit. Everything else is controlled
power to heat things and will go to zero at the highest temperature.

If you run the inner oven at an offset of 3 to 5 C from the outer, the inner
heater will likely not pull more than 40 mw and may pull quite a bit less.

Bob