[time-nuts] Z3801A / Z3812A leapsecond funkiness

[Mark Sims]

As discussed earlier there is a bug in the firmware of these devices where they report the wrong date of a leapsecond if it is announced more than three months in advance.  Mine were reporting leapseconds will occur on 30 Sep 2016.    

Well, the event passed and the devices reset their leapsecond pending flags as expected.  I was expecting them to see that the satellites were still broadcasting a leapsecond pending flag and expected them to recover and show the new/proper leapsecond date.  Well, they they did not automatically recover and report the correct leapsecond at the end of this December.   After a day (long enough to get a fresh almanac) they still had not recovered.  A soft reset command did not cause them to recover.  I had to cycle power to get them back to reporting a pending leapsecond.