# Picosecond range White Rabbit Digital Time Scale

- Institution: CERN
- Speaker: Sougandh Kannoth Mavila
- Year: 2026
- Language: English
- Source: [CERN](https://videos.cern.ch/record/3025806)
- Licence: [CERN](https://copyright.web.cern)
- Who can watch: Teens profiles and older (sources give no ratings; placed by Invisible Network from the source's own words)

We have developed an upgrade to White Rabbit (WR) that solves ambiguities in networks were the active Grand Master (GM) is backed up by several 'passive' GMs, each synchronized to their own reference clock. In such networks, each clock defines a different reference time (often with relative offsets of nanoseconds to microseconds). These offsets will cause severe time jumps when failing over from the active GM to a back-up GM, and raise the more fundamental issue of how to define a single network time scale from the ensemble of connected reference clocks. Using modified WR software and off-the-shelf OHWR hardware (WRSv3), we demonstrate the generation of a virtual network time with ~100 ps residual time errors from a four-clock ensemble and stability that is superior to that of the individual clocks. Using redundant paths, clocks, and time-scale algorithms, we essentially remove all single points of failure from the WR network, while adding a software-defined, virtual time scale that can be realized in digital (PTP) as well as physical (10 MHz, 1 PPS) form in any node of the WR network.

00:00:00 Slide 1

00:00:27 Slide 2

00:01:12 Slide 3

00:02:52 Slide 4

00:03:29 Slide 5

00:03:51 Slide 6

00:04:06 Slide 7

00:06:04 Slide 8

00:06:17 Slide 9

00:06:32 Slide 10

00:07:28 Slide 11

00:08:06 Slide 12

00:08:39 Slide 13

00:09:04 Slide 14

00:10:41 Slide 15

00:11:14 Slide 16

00:11:59 Slide 17

00:12:14 Slide 18

00:12:44 Slide 19

00:13:06 Slide 20

00:13:31 Slide 21

00:15:54 Slide 22

## Lectures

1. [Picosecond range White Rabbit Digital Time Scale](https://invisiblenetwork.ai/learn/cern/3025806/70f444f94c48), 21 min (CERN's player)