[time-nuts] Measurement of frequency of HP 8720D option 1D5 oscillator after switch on

Magnus Danielson magnus at rubidium.dyndns.org
Sat Oct 4 02:57:11 EDT 2014



On 10/02/2014 09:21 AM, Dr. David Kirkby (Kirkby Microwave Ltd) wrote:
> On 2 Oct 2014 07:10, "Magnus Danielson" <magnus at rubidium.dyndns.org> wrote:
>>
>> David,
>>
>> The character of starting high/low and then stabilize some 5-30 min later
> is typical of oven oscillators. Underdamped ovens have been seen before, I
> have even seen one on the brink of oscillation.
>
> Thank you. Do you know the likely cause of the somewhat odd behaviour
> between about 150 and 220 s?

Could be that it overshoots in this non-linear way due to a little low 
cooling, so that as the heater runs full current, it doesn't cool of 
quick enough for the transition to be nice and linear. Rather it 
overshoots, turn heater off and then cools off and the heater smoothly 
goes on again and stabilizes. It's not a big problem.

You can have too little cooling, and then the heater chirps the heater 
and then the crystal decays in temperature to get a heater chirp again. 
This is really bad and leaves the crystal unstabilized for most of the 
time and you get a modulation on frequency from the saw-tooth 
temperature. You don't look to be close to that, but I've seen it happen.

> This eBay auction,  which someone posted
>
> http://m.ebay.com/itm/151256172424
>
> says it's the high stability oscillator for an HP 8753D or 8753ES. I
> checked the part number at parts.keysight.comand it would appear it is the
> same as used in my VNA and several other microwave VNAs.
>
> That said,  I have noticed a few errors on parts.agilent.com, one of which
> resulted in me buying the wrong part. I was later warned by Agilent not to
> trust the accuracy too much, especially on older equipment. It is better to
> check with them before making purchases.
>
> But they have a very helpful parts service that does make every effort to
> sort out what parts are. They spent quite some time finding out what
> connectors were on am obsolete cable for me.
>
>> TCXO will not have the same wide range, as it compensate for the
> temperature.
>
> This answers my original curiosity now - did I have an OCXO  or TCXO.

With that curve it is definitely an OCXO, and my guess is that it is an 
SC-cut, which matches starting +40 ppm high (if I recall things 
correctly). It is also typical for 2" OCXO oscillators and the expected 
Q and thus noise will come at handy in a VNA.

> Although I am not going to bother, as it will be easier and more accurate
> to feed the VNA from a rubidium or GPS locked TCXO/OCXO, it would probably
> be possible to buy one of those off of eBay and replace the OCXO with a
> better one. Then stick it in my VNA - I would not want to modify the
> original one.
>
> There have some rather small double oven OCXOs on eBay recently for very
> little money.  From the earlier comments about this oscillator, it would
> appear its specification is quite poor for an OCXO.

On time-nuts, things can get a little exaggerated in the "poor" aspect. 
What is more relevant is will it greatly affect your VNA performance?
Precision on frequency is trimmable once the oven has been well heated.
Noise-wise I would say it suffice.

What is your VNA frequency resolution anyway? 1 Hz?

Cheers,
Magnus


More information about the time-nuts mailing list