[time-nuts] Trimble Tbolt temperature

[Bob kb8tq]

Hi

Since the DAC / reference is the main source of error, placing the sensor
a bit away from the OCXO makes sense. You also don’t want to watch the 
OCXO heat the board. You want data on the outside temperature. 

Since it’s all buried away in a Khalman filter, even the people who designed
it were unable to be 100% sure of what exactly it was doing. ( I paid for 
lunch to find that out ….it was back in the late 1990’s so lunch was cheap) 

Bob

> On Dec 4, 2019, at 10:58 PM, Didier Juges <shalimr9 at gmail.com> wrote:
> 
> Maybe the Khalman  filter does not want to know the crystal temperature and
> is more interested in the ambient temperature which is better acquired away
> from the internals, i.e. in the corner of the board. The oven regulates the
> crystal temperature so if it does its job, there should not be too much
> variation to measure and the drift of the DAC and reference which I assume
> are outside the oven would not be compensated.
> 
> On Wed, Dec 4, 2019, 9:23 PM Charles Steinmetz <csteinmetz at yandex.com>
> wrote:
> 
>> Tom wrote:
>> 
>>> Section 5.1.2 (Kalman Filtering) on page 5-4 (PDF page 48) says:
>>> 
>>> Oscillator performance is subject to two basic effects. First, changes
>>> in environmental temperature can cause the oscillator to speed up to
>>> slow down. Second, the oscillator has a natural tendency to drift over
>>> time. This is called aging.
>>> 
>>> Both temperature and aging can be mathematically predicted. However, the
>>> characteristics vary from crystal to crystal. The Kalman filtering
>>> monitors the unique oscillator performance over time and temperature and
>>> records this behavior.
>> 
>> The DS1620, perched way over in the corner of the board, cannot provide
>> useful crystal temperature data, even by inference.  That would need to
>> come from within the oven at the very least, more likely from inside the
>> crystal envelope itself using, e.g., an IR thermometer.
>> 
>> Instead, Kalman filters in GPSDOs track the oscillator frequency
>> continuously and do their best to separate the drift effects from the
>> temperature effects using software algorithms (note "mathematically
>> predicted" in the paragraph above).  You can see a hint of this by
>> monitoring the statistics from one of the older HP units like a Z3801.
>> 
>> Best regards,
>> 
>> Charles
>> 
>> 
>> 
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
>> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.