[time-nuts] Fundamental limits on performance

Magnus Danielson magnus at rubidium.dyndns.org
Sat Sep 19 15:36:21 UTC 2009


Lux, Jim (337C) wrote:
> 
> 
> On 9/13/09 3:54 PM, "Magnus Danielson" <magnus at rubidium.dyndns.org> wrote:
> 
>> Jim,
>>
>> Magnus Danielson wrote:
>>> Jim,
>>
>> One aspect of why synchronous data and carrier may be worth pointing
>> out. If the carrier and modulation is asynchronous, then the carrier
>> tracking and data reception needs to recover their respective clocks
>> independently. However, by transmitting them in a synchronous fashion
>> and making use of this fact at the receiver, then the carrier tracking
>> can aid the code tracking in which case the code tracking only need to
>> retain the phase, which leaves more margin to propperly decoding the
>> message. Thus, a better BER is achieved for the same S/N or for that
>> matter, a worse S/N can be tolerated for the same achived BER compared
>> to the asynchronous modulation technique.
> 
> 
> As in GPS, for instance, where chip rate is related to carrier frequency.

Exactly. Very strict integer locked.

Just realized that the gravity potential detuning needs to be adapted to 
the mars orbit and gravity field. Which also makes an interesting 
side-project to develop gravity maps for mars.

Is there an established coordinate system of suitable precission for 
mars? Essentially the equalent of WGS 84.

>> Anyway, the questions you are asking have been covered before. It should
>> come as no suprise that Dr. Simon was (is?) with JPL.
> 
> Marv Simon is still at JPL..

As expected. You just don't know what people do today.

If rolling back to the original question, how would you like the 
fundamental limit to be expressed? There are several parameters relating 
to the mission which can vary. I think orbit solution, transmitt power 
of suitable band(s) and signal structure interconnect. As secondary 
issues is long term stability of frequency sources and stability of 
orbit prediction and hence stability of position. Receiver noise can be 
fairly well estimated.

So start from receiver levels, coding gains to suitable correlation 
signal strength, then calculate typical receiver path losses. In the 
other end, calculate typical space losses from orbit, and expected gain 
structure at the transmitter. I think the critical parameters for the 
mission would resolve themselfs fairly quickly. MMO gives long 
observation times. An alternative would be going back to TIMEX style and 
just use doppler observations, but that worked best for navy and land 
apps with restricted dynamics (no dropping down from the sky dynamics).

Bi-directional ranging could be used, but it does not really change the 
fundamental limits, it just change the mechanisms for orbit predictions 
somewhat and convert pseudo-ranges into ranges for improved 
time-transfer and orbit prediction before full constellation exists. 
With a single node (sat or surface) can bi-directional ranging resolve 
time difference completely and place the node in a sphear of known 
radius from the node. A second node would reduce the position to be 
placed on the intersecting circle between those nodes and a third would 
provide the point estimate, which infact is a small space due to 
uncertainties. The GPS requires 4 observable nodes to achive the same, 
as it's pseudoranges is unidirectional and thus can not by themselfs be 
reduced to real ranges. Another aspect is that the range equations 
should be much quicker and easier to resolve as the time-error would be 
much less than initial for a GPS equalent case. But it doesn't shift the 
receiver limits by much.

Assuming a decent IMU, receiver aiding can be used. But the IMU aspect 
should not be very new to you guys. :)

Cheers,
Magnus



More information about the time-nuts mailing list