PS BGI - meeting with OP about timing requirements
ccc
CERN
From: Denis Gerard Cotte <Denis.Cotte@cern.ch> , Sent: Friday, January 17, 2020 17:39
To: Steen Jensen <steen.jensen@cern.ch> , Cc: Alexander Huschauer <alexander.huschauer@cern.ch>; James Storey <james.storey@cern.ch>
Subject: RE: bgi cps - timing requirements
Hi Steen, Concerning the number :
T revolution @ 1.4Gev = 2.29us , T revolution @ 2Gev = 2.21us
PIX.WRF occurs -2 ms before injection. It count PIX.TRF train :
@ 1.4Gev = 286ns , @ 2Gev =276ns
Cheers, Denis
From: Steen Jensen <steen.jensen@cern.ch>
Sent: 17 January 2020 15:33
To: Denis Gerard Cotte <Denis.Cotte@cern.ch>; Alexander Huschauer <alexander.huschauer@cern.ch>; James Storey <james.storey@cern.ch>
Subject: bgi cps - timing requirements
Dear all,
Here are my notes from our meeting about CPS BGI timing needs this morning:
- PIX.WRF
Needed to get a jitter-free, beam-synchronous start signal. Precedes beam arrival by 2ms - RF clock
If this signal cannot be cabled in time, a fall-back solution is using the internal 40MHz clock of the CTR which will introduce some amount of jitter.
Needed for bunch-by-bunch measurements
Denis explained this is always the H8 clock.
The harmonic can change during the cycle and hence we need to know when this happens.
One option is to read the “harmonic-function” from a certain GFAS.
However, Denis preferred to be able to define this information manually via the FESA class - Revolution clock (~280ns@1.4GeV, ~300ns@2GeV)
Needed for multi-revolution measurements
If phase adjustments become necessary, one option is to use a Fine-Delay module.
However, James proposed that Hampus can implement similar functionality in the FPGA to be used from the FESA class
Action points:
- James/Steen: Request from Emmanuel Said (CO) the cabling for the above signals to the BGI crates in building 353
- Hampus (in absentia): Implement a fine-delay in FPGA along with software interface for calling it
- Steen: Define and create the required LTIM instances
Please let me know if you have comments.
Cheers, Steen