[time-nuts] Accurate 1 pps signals

Mark Sims holrum at hotmail.com
Wed Aug 19 15:56:19 UTC 2009


Optimum performance from a GPSDO requires two things:  feed it with quality inputs and keep disturbances to a minimum.

Use a very high quality antenna.  My testing shows that the higher the quality the antenna,  the better the performance.  A geodetic quality antenna is best,  a survey grade antenna is good,  an automotive type patch antenna without a ground plane is horrible.

Know where the heck you are down to a gnats ass and get that location into the unit with full accuracy.   Errors in the saved antenna position from its true position have a definite effect on the quality of the output.  Ideally have the antenna location surveyed in WGS84 coordinates.  

If an antenna survey is not possible (it will cost you quite a bit of coin) use a good self-survey to determine the antenna location.  Standard receiver self surveys just average fixes for an hour or so...  not good.  Any self-survey should take place over 24 hour time intervals.  This will help mitigate the effects of satellite and atmospheric cycles.

Get that accurately known position into the receiver with full accuracy.  The Trimble Thunderbolt allows one to manually enter a known location,  but it only uses single precision (32 bit) floating point numbers.  This can introduce several feet of error.  The next version of Lady Heather should have some code to do a precision self-survey and save the position to good accuracy.

Mount the antenna where it is far from anything.  Move to the middle of a lava field.  Things that are higher than the antenna can reflect distorting signals into the receiver (multipath).  Trees are bad,  bad,  bad.   Become one with the chainsaw.  Buildings are bad, bad, bad.  Become one with the wrecking ball.  

Make sure you properly measure and specify your antenna cable delay.  Don't forget the delay in the output cable.  Use identical cables from the same production lot.  Make sure the output signals are driving identical loads. 

Use identical setups.  Use the identical antennas.  Use identical receivers with the same revision hardware/firmware.  Make sure your Thunderbolts have the "good" temperature sensor that produces high resolution temperature reading.

Keep the unit at a constant temperature and will isolated from the environment.  Active temperature control to say 0.001 degrees would be nice.  Otherwise isolate it from the environment.  Air conditioning and heating cycles are bad.  Sunshine on its shoulders does not make it happy.

Find a corrugated cardboard box around three inches on a side bigger than the unit.  Mount the unit in the box on a piece of foam, etc so that it is centered in the volume of the box.  Cover the box with aluminium foil.  This keeps my unit's temperature stable to around 0.06 C despite 1C A/C cycles.   Be careful not to insulate the unit so well that it cannot dissipate it's own internal heat and it cooks itself in its own juices.

Feed your unit with good,  clean,  stable power.  Use a high quality power supply,  drive the supply with clean, stable input voltage.  Isolate the supply from the environment.  Using the standard TAPR Thunderbolt supply,  about 1/3 of the temperature related effects seem to be coming from the power supply.  

Use caution when thermally isolating the power supply.  Supplies are known to go up in smoke and flames.  Don't keep it in a flammable box, etc.   If you are paranoid,  don't isolate the receiver in a flammable box.  If you are really paranoid,  tin foil in the baseball cap helps.





_________________________________________________________________
Get back to school stuff for them and cashback for you.
http://www.bing.com/cashback?form=MSHYCB&publ=WLHMTAG&crea=TEXT_MSHYCB_BackToSchool_Cashback_BTSCashback_1x1


More information about the time-nuts mailing list