Speaker
Description
To meet the High-Luminosity LHC demand for enhanced phase stability and replace the outdated RF and timing distribution system (called TTC backbone), an upgrade to a White Rabbit-based system is planned. A compliance test showed a phase variation of less than 30 ps in a proof-of-concept system. In order to assess the quality of the new system in real conditions a test campaign was started in several points of LHC monitoring its phase stability with respect to the RF-master frequency and comparing it to the TTC backbone currently in operation. The test campaign will be described and results presented.
Summary (500 words)
The timing distribution of the LHC is based on obsolete components and the
High-Luminosity (HL) upgrade of the LHC requires a higher phase stability,
of the order of 10s of picoseconds. Therefore it will be replaced by a White
Rabbit (WR) based network, similar to the one employed at SPS. This has
the advantage of being scalable due to its Ethernet topology and providing
high phase stability even under temperature fluctuations and different operating
modes.
A preliminary compliance test was done on a proof-of-concept system re-
alized in a lab by a cascade of switches and White-Rabbit to RF (WR2RF)
boards developed for the SPS RF frequency of 200 MHz. It revealed a peak-
peak phase stability of less than 28 ps under normal operation and with network
disturbances , suitable for the timing requirements of the HL-LHC experiments.
In SR4 where the LHC RF control system is located, the RF frequency
is encoded in binary format and sent to the master switch of the WR timing
network in the form of a Frequency Tuning Word (FTW). This master switch
is located 7km away in CCR, close to the CERN Control Center. It distributes
the FTW through the WR network to the experiments, where a WR2RF board
recovers the bunch clock (one tenth of the RF frequency) and orbit from the
FTW with very low jitter (< 3ps rms) and high phase stability (< 30ps).
The planned tests involve acquiring phase measurements at two locations.
One location is SR4, where it is possible to measure the phase variation be-
tween the original master RF frequency and the bunch clock recovered from the
FTW of two different WR networks: a reference network, fully located in SR4,
consisting of one WR switch (WRS) and a WR2RF board which regenerates
the bunch clock from the FTW; and a realistic WR timing network in the form
of a second WR2RF board (located in SR4) connected to the master switch of
the WR timing network (in CCR) through a cascade of 5 WRS, to mimic the
network connection to the experiments.
The second part of the test takes place at the CMS experiment. There a
third WR2RF board also connected to the WR timing network through five
layers of switches retrieves the bunch clock and orbit from the FTW send by
the frequency master in SR4. The phase between these clocks and their counter parts from the original TTC backbone is measured.
This new WR-based architecture is analyzed, characterized, and compared
with the legacy architecture, aiming to assess its quality in terms of phase
stability for the HL-LHC requirements.
