Hi

The key bit of missing info:

You likely have a BT cut crystal in that OCXO. If it was an AT, you wold look for a minimum frequency. With
the BT, you look for a maximum.

Bob

If you replaced the thermistor with an exact replacement,
then you shouldn't need to change the pot.  If you didn't
replace the thermistor with an exact replacement, then you
should carefully measure the resistance of the pot as you
found it.  (I hope you did not fool with it already.)  From
the pot resistance, and the thermistor curve specs of
the original thermistor, you should be able to calculate
the factory oven temperature set point.  You can then
calculate the new pot resistance based on the specs
for the replacement thermistor.

Trying to determine the turn over temperature as you
described (which was done with an automated system
in the E1938A) would be extremely cumbersome when
done by hand.  It would be a last resort if you fooled
with the pot.

Rick

On 2/27/2016 6:48 PM, Nigel Vander Houwen wrote:

Bob,

I’m not sure what type is in it. It is a 10544A, not one of the 10811 devices. The references I’ve found indicate the 10544A being an AT type. If it’s of any use, the date codes on the parts in the oven control circuit indicate it was built in the latter part of ’77.

I can’t say I’ve got a whole lot of experience looking at this, but should I expect to see a curve along the values, of which there would either be a minimum or maximum? That would be fairly indicative.

Rick,

The thermistor was not an exact replacement, and I couldn’t find any specifications beyond 9.93KOhm @ 80C referenced on the schematic. I don’t know that the pot wasn’t turned during the original debugging of why the crystal wasn’t on frequency, so my best bet now unfortunately is to find the set point myself.

Nigel

Hi

The whole “BT Cut” issue was a big top secret in HP. They spent a lot of time obscuring
the fact that they used BT’s. The belief was that if any of the other outfits figured out that
was what they were doing, the competition would have better OCXO’s.

Watch the frequency as it turns on. An AT will go down a bunch in frequency as it warms up.
An SC or BT will go up.

Bob

On 2/28/16 6:23 AM, Bob Camp wrote:

Can you tell what the cut is if you have the blank in front of you?

Wouldn't the competition just buy an instrument with the oscillator, saw
it open, and measure it?

Jim,

Why cut it open? It is relatively easy to measure just by power the
oscillator up. Just in this thread is an echo of a discussion we had
many years ago that explains the method.

The temperature curve for AT, BT and SC-cut is known and quite
different, so just measure the heat-up curve from room temperature to
oven temperature will be distinct for these types. I think the IT-cut
can also be identified uniquely this way.

If you have the crystal blank, you can use X-ray to see the
crystal-orientation.

Cheers,
Magnus

On 02/28/2016 05:48 PM, jimlux wrote:

Jim wrote:

You can do it directly with x-ray crystallography.  Otherwise, you
need to characterize multiple crystal parameters and infer the cut
from those, which may not be particularly accurate.

Best regards,

Charles

Hi

Ok a bit more of the story.

It’s easy to simply turn on the device and see how it warms up. Back when it was made, the
SC did not yet exist. The only thing it could be was a BT. With an X-ray setup you can absolutely
tell it’s a BT. With the blank and a pair of calipers you can make a darn good guess it’s a BT.

Since HP did not make their own blanks, the “competition” was the source of their blanks. No need
for those guys to guess about anything.

Despite all of this logical and rational thinking, the BT remained a “top secret” sort of thing as far
as (at least certain people at) HP were concerned. Those who were concerned also had the route
to the HP PR machine so that’s the story that went out to the world.

Those involved left HP long ago. The whole thing became a non-issue once the 10811 came out.
What is the most amazing thing to me is that 30 years after it became a non-issue there still is
confusion about the topic.

Bob

Howdy All,

This thread has split into a couple, but I’ll try to respond here to the various things.

Based on how this crystal warms up, it does appear that it is a BT type crystal, as it warms up frequency goes up, and as it gets hotter eventually turns around and heads downhill again.

I’ve managed to tune the oven to where the peak seems to be. It’s a many turn potentiometer (something like 21 turns), that plays a small portion on the resistance, so +/- about a quarter turn at the peak didn’t seem to really impact the frequency. I left it in the middle of that range.

My replacement thermistor has a lower beta than what it seems the original had. The original thermistor being specified at 9.93K @ 80C, and the calculator shows mine at about 12K @ 80C. Since the circuit for this crystal has a 10K + 9K & the 2K POT on one side of an op-amp comparator, and a 10K + the thermistor on the other side, I added a 2.2K resistor to the POT leg of that voltage divider to bring the tuning range about to where it was stock, as I found that unaltered, with the new thermistor I couldn’t set the oven temperature low enough, every frequency was on the downhill slope past the peak.

I appreciate Frank’s offer for a close to original thermistor, however I’m not in the netherlands (despite my name), and I’ve already got a reasonably close replacement epoxied into the oven like the original was.

Now I’ll leave it run for a while and see about tuning the frequency. I’ve got it pretty close at the moment, but the coming days/weeks will show their own colors.

I would like to thank everyone for the insight around getting this thing tuned up. It’s sincerely appreciated.

I’ll probably start designing a DAC + Phase comparator board to GPS discipline this old crystal. Have to see how that ends up comparing to my thunderbolt.

Nigel