Automagical Holdover Time

Automagical Holdover Time

Posted on 19/12/2017 by Quentin Griffiths
What do you do when your clock loses sync?  Do you want it to continue operating for a period of time before reporting out of sync?  If so how long?


What is the drift rate of my internal oscillator?  When will it be to inaccurate?  How do I calculate all of this?


ARGH!  Wouldn’t it be great if you could just tell a clock how accurate the output needs to be, then for it to squelch the outputs once this accuracy limit has been reached?


We liked the idea so much that we added this to our latest firmware version for the NTS 03-G+ and the TCG 02-G range of GNSS clocks.


Here is an overview as to the new features we have added to our clocks;


1)     Maximum Output Inaccuracy Check


The first change you will see is that under the Outputs tab of the Tekron configuration software, there is a new setting called “Enable maximum inaccuracy check” with a drop-down box for the different options, as shown below.  This setting allows you to define an accuracy limit for the clock to work within and transition to an out of sync state, should this boundary be exceeded.





So how does this all work and does it replace the holdover timer?  Below is a short overview of the different operational modes of the clock, where you can use either the holdover timer, or the inaccuracy checks alone, or a combination of both to meet your timing goals.


If "Enable holdover timeout" is selected and "Enable maximum inaccuracy check" is not selected


This is closest to the traditional behaviour of our clocks.  When no sync sources are available a timer is started that, when expired, will cause the clock to transition to the "out of sync" state and squelch the outputs.  


If "Enable holdover timeout" is not selected and "Enable maximum inaccuracy check" is selected


At all times, including times when syncing to a source, the calculated output inaccuracy of the clock is compared against the configured maximum inaccuracy threshold within the software e.g. 10 µs.  If the threshold is exceeded the clock will transition to the "out of sync" state and squelch the outputs.   


If both "Enable holdover timeout" and "Enable maximum inaccuracy check" are selected


Then both the holdover timer must have expired, and the inaccuracy threshold exceeded before the clock will enter the "out of sync" state.   The holdover timeout then serves as the minimum time that the clock will remain in holdover.   This mode excludes the use of the inaccuracy threshold as a noisy source filter as out of sync will not be reached unless the holdover timer has started (and expired) which can only happen if no source is available.


If neither "Enable holdover timeout" or "Enable maximum inaccuracy check" are selected.


In this case the clock will (almost) never enter the "out of sync" state even if no sources are available, instead remaining in the holdover state until a source becomes available again.  The inaccuracy estimate will continue to be made and outgoing signals that can carry this information will report the current accuracy level of the clock that will degrade over time.   


2)     Individual Output Suppression


In the past, when a clock was to go into an out of sync state, users were given the option to squelch all of the outputs on the clock, or to leave them operating, with the sync state and perceived accuracy reported in each of the time codes. 


This all or nothing approach was fine for most substation applications, where a protection network only required a timing accuracy of 1 ms and was not critical to the plants operation. 


In modern applications, where a clock may be connected to a telecom’s network (high accuracy), as well as a substation protection network (sequence of events - lower accuracy) there is a need for some outputs on the clock to be supressed when the holdover timer expires or the inaccuracy check is exceeded, whilst others can continue to operate.


To address this need, our latest firmware release allows you to define how the output ports react, when the clock loses sync.


The available options for when to suppress the output are;


Holdover timeout expires
The output will be suppressed if the holdover timeout is enabled and has expired.   If the holdover timeout is not enabled then the output will never be suppressed.
Inaccuracy threshold exceeded
The output will be suppressed if the inaccuracy threshold is enabled and has been exceeded.  If the inaccuracy threshold is not enabled then the output will never be suppressed.
The output will never be suppressed


This option is easily configured under the I/O tab, by selecting the output, then using the “Suppress Output When” drop down box, as displayed below.






Please note that not all outputs support this option, so please refer to the firmware release notes for further details.


The release notes for the firmware changes, the configuration manuals and more can all be downloaded here


If you would like to learn more about our new features, want help upgrading, or would like a demo on the new features please contact us!


Automagical Holdover Time