[time-nuts] This may be my new favorite old oscillator
skip.withrow at gmail.com
Tue Nov 8 00:16:48 EST 2016
I recently acquired an HP 5065A rubidium oscillator (with 10811 10MHz
OCXO). I think I pretty much have it running now and have been letting it
cook for the last couple of weeks. I offset the C-field + and - and
measured the frequency to calculate the C-field sensitivity. My unit came
out to 1.96x10E-12 per dial unit, which agrees with the manual stated
2x10E-12. So, calculated the on frequency C-field value and dialed it in.
Attached is a Lady Heather plot of the frequency over the last 3 days. The
purple line is the 1pps plot with the vertical scale being 20ns per
division. So, the unit is off about 125ns over the last 72 hours (running
about 4.92x10E-13 slow). So my C-field setting is off about 1/4 of a
division, but I think I'm going to leave well enough alone.
The yellow line is the NTBW50AA temperature sensor, and you can clearly see
when the furnace cycles. I was away for the weekend, and you can also
clearly see when I came home this evening and turned up the heat. At the
very end of the plot is the spike when I turned on the lights in the shop.
I love using a Thunderbolt/NTBW50AA for making frequency measurements this
way. I remove the OCXO, and insert the 10MHz from the DUT. Then disable
disciplining so the DAC voltage doesn't try to chase the open loop
oscillator. Of course the short-term performance looks worse than it
actually is because of measuring against GPS, but the long-term
measurements are very good.
I want to log this unit at regular intervals to see what the aging looks
like. Also need to do some measurements against the cesium to see what the
short-term performance might be. But, I think this oscillator will be a
good reference in many cases in lieu of using the cesium.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 39037 bytes
Desc: not available
More information about the time-nuts