Re: Waveform Record Device Support

[Burkhard Kolb <b.kolb@gsi.de>]

Am 08.06.2010 um 11:00 schrieb <graham.cox@stfc.ac.uk> <graham.cox@stfc.ac.uk>:

> Thanks for the replies.
> 
> Now knowing what the philosophy of the implementation of the waveform
> record is I have the data successfully transferred into EPICS.  Reading
> the NORD field to get the actual data size in my clients is not a
> problem, however I have followed Juraj's suggestion and separated my
> individual channels of data using subArray records.
> 
> To expand, RE Emmanuels request -
> 
> The BPM system is for EMMA, a non-scaling FFAG accelerator.  Each
> injection of beam into the accelerator could potentially be for a
> different specified number of turns around the ring before extraction at
> a particular energy.  Also the ADC clock for the data acquisition is
> derived from the beam directly off the vertical BPM pickup.  So,
> potentially the beam could be outside of the range of detection for the
> pickup, but still within the acceptance of the lattice.  In this case
> the ADC stops triggering and the acquisition run for that BPM is
> complete.  So each BPM could receive data anywhere in the range from 0
> to 1000 turns depending on the number of turns EMMA is setup for.  In an
> ideal world, all BPMs would see all turns, but this number would still
> be variable.  The BPM system is a turn-by-turn in house design
> specifically for EMMA.
> 
> As a follow up question, what would be the recommended way of performing
> calculations on array data within the EPICS database?  Do I have to
> resort to subRoutine records?  In my individual channels of data
> (subArray records) I have multiple quartets of data, one for for each
> turn captured, I need to perform calculations on each of these quartets
> to determine beam position for each turn.
> 

For the calculation you can use the aSub (formerly genSub) record.
Burkhard

> Thanks,
> 
> Graham
> 
> -----Original Message-----
> From: emmanuel_mayssat@lynceantech.com
> [mailto:emmanuel_mayssat@lynceantech.com] 
> Sent: 07 June 2010 22:39
> To: Cox, Graham (STFC,DL,EID)
> Cc: tech-talk@aps.anl.gov
> Subject: Re: Waveform Record Device Support
> 
> On 10:06 Mon 07 Jun     , graham.cox@stfc.ac.uk wrote:
>> I am integrating a turn-by-turn BPM system into EPICS and would like 
>> to store the raw BPM pickup data in a waveform record.  Due to the 
>> nature of the system, each run can result in a different number of 
>> samples for each BPM pickup, so the array size is unknown in advance 
>> and could potentially be different every run.
> 
> Given my limited experience with particle accelerators, I am just
> wondering what kind of system are you working with?
> Storage ring, linac, ... pulsed, not pulsed?
> If the waveform length is changing, it can only be because data
> acquisition is asynchronous.
> Are you sampling CW signals? I can't think of a system that would work
> with those requirements.
> Could you expend?
> 
>> Is it possible to resize the number of waveform record elements in my 
>> device support, or do I have to size the waveform record in advance 
>> for the maximum number of samples that could be expected.  Most of the
> 
>> examples of waveform device support I see read the nelm field and 
>> return this fixed amount of data via the records bptr.
> 
> Another solutions may be to use FIFO waveforms, aka compress records.
> The compress record is part of epics-base.
> In this case, the number of samples per channel acquired can be
> different, but the length of the FIFO stack can be the same.
> 
> Good luck,
> 
> --
> Emmanuel
> -- 
> Scanned by iCritical.
> 

--
Dr. Burkhard Kolb                              
Pleikartsförster Str. 38
69124 Heidelberg
Germany
E-Mail: B.Kolb@gmx.de 
Tel: +49 6221 718640 / Fax: +49 6221 718641