[ntp:questions] Re: drifting on crystal
Marc Brett
marc at fordson.demon.co.uk
Thu Jan 13 11:53:40 UTC 2005
On Thu, 13 Jan 2005 10:24:25 -0000, "David J Taylor" <david-taylor at invalid.com>
wrote:
>Brad Knowles wrote:
>> At 6:33 AM +0000 2005-01-13, Marc Brett wrote:
>>
>>> That's the hardware cost. How many watts does it consume?
>>
>> No more than the low-quality oscillator that is actually included
>> by the hardware vendor.
>>
>>> why would an engineer include a component with
>>> "oven" in its name if he didn't have to?
>>
>> There is no actual oven included with the crystal. The issue is
>> that the OXCO has gone through a more rigorous testing process which
>> includes testing in an oven, and has been certified to work to
>> certain tighter performance standards than your average everyday
>> crystal, and either actively compensates for temperature variations
>> or is relatively immune to them.
>
>An OCXO is a crystal oscillator assembly contained inside a temperature
>regulated oven to keep it at a constant temperature. The crystal itself
>will be chosen to have a low temperature coefficient, but over a more
>limited range than a non-ovened crystal. I.e. a normal crystal might be
>specified to be +/- 10ppm over 0C..45C, but one intended for oven
>operation might be +/- 1ppm over 60-65C. (I mage the figures up, by the
>way). Such characteristics may be achieved by a different cut of crystal.
The power consumption of an OCXO is typically 1 - 6 Watts, whereas a standard
quartz oscillator consumes microWatts. They are physically quite large.
Since NTP gives more than enough accuracy for most people (even if the synch
interval is hours or days), it's a fair engineering tradeoff to leave out the
OCXO. Most consumers would prefer lower power, thus quieter operation and a
smaller package, than hyper-accurate time.
Instead of OCXOs everywhere, I'd prefer to see a standard connector so consumers
can easily replace the crystal with a TCXO, OCXO, or one of those new tiny
atomic clocks if their application requires it.
More information about the questions
mailing list