[time-nuts] NTGS50AA, better than Thunderbolt?

Mark Sims holrum at hotmail.com
Mon May 21 02:41:52 UTC 2012


Much of the improvement seems to be in the better oscillator temperature compensation...  no telling how the phase noise is,  but the Nortel osc has a VERY flat aging curve and no detectable DAC changes with temp.   I have several tbolts and they all have a very temperature dependent DAC curve.  The Tbolt DAC curve seems to follow the temp curve very closely.  It does seem to be tracking the osc and not the temp sensor...  I tried freezing the tbolt osc can with cold spray and the DAC showed changes the temp sensor did not.  It is obviously compensating for temp changes in the oscillator.

The Tbolt and the Nortel look like they use the same DS1620 temp sensor,  but the Nortel firmware does not show the temperature sensor glitch.  Once the oscillator settled in,  I was seeing rms errors in the OSC plot in the sub 10 ppt range and the PPS plot in the 2 nsec range.   Even a temperature stabilized tbolt was several times this.

I have not run it a lot in holdover mode (except when I had the antenna disconnected so that I could do some Resolution SMT work).  I was not recording data in that mode.  The Nortel unit seems to be designed with more fault tolerance and redundancy then the Tbolt.  For instance,  it has multiple copies of the firmware and can select a good copy.  You can also hot-upgrade the firmware.   I have not recorded a lot of data from the Nortel since I am still playing with updating the program and am interrupting the runs quite often.
----------------------
Do you have data/plots on this and your previous NTGS50AA? Several posts imply it is "better than a Thunderbolt". If you have some phase noise plots or locked phase/frequency difference or holdover raw data that you could send I'd appreciate it. 		 	   		  


More information about the time-nuts mailing list