Adhoc BI-TB for LIU BWS power electronics

Name: Adhoc BI-TB for LIU BWS power electronics
Start: 2020-04-15T11:00:00+02:00
End: 2020-04-15T12:00:00+02:00
Location: CERN

Wednesday 15 Apr 2020, 11:00 → 12:00 Europe/Zurich

CERN

- 11:00 → 11:30
  Power electronics for LIU-BWS 30m
  
  Speakers: Federico Roncarolo (CERN), Jonathan Emery (CERN)
  
  20200415_BWS_LIU_electronics_Adhoc BI-TB.pdf
  https://indico.cern.ch/event/907313/
  
  Sum up from Jonathan’s presentation
  
  Each LIU BWS control drive (IDC=Intelligent Drive Control unit) hosts 4 main boards:
  
  1-VFC
  
  2-Mezzanine BWSAIF (mounted on the VFC)
  
  3-Motor inverter BWSMIB
  
  4-Charger BWSCPC
  
  VFCs are of course part of the BI production and stock
  
  Components for ALL other cards are procured already
  
  Original plan (in line with pre-Covid-19 injectors master plan):
  
  Produce PCBs, assemble boards, test individually and then as whole unit for 10 Pre-series units (10=8 for PSB + 2 for FW and SW development)
  
  After pre-series validation: produce PCBs and mount series (40 units for PS, SPS, ESS + spares)
  
  Table summarising costs
  
  excluding mezzanine: PCB production and assembly ~= 31k (money lost if need to redesign not working)
  
  decision on Mezzanine BWSAIF go ahead ~independent from other two cards, but: latest version not fully tested, some risk remains)
  
  Work advancement ok until Feb
  
  individual boards tests ~completed
  
  at the moment of testing full IDC: repeated (6 times) problem with BWSCPC when controlled (as by design) by inverter BWSMIB
  
  Only when controlled by BWSMIB (i.e. charger proved several time to work when controlled by external circuit, BWSMIB always there to invert)
  
  Happens after few seconds of operation
  
  burning of Flyback shunt, MOSFET and driver
  
  after extensive investigations and tests: problem traced back to a sum of design flaws, control link between B WSCPC and BWSMIB not functional.
  
  put in place TEST SETUP in which control link between two boards is cut, external (timed) relay used to enable charger, fast discharge resistor to simulate discharging during a scan
  
  last two days at CERN (Miguel): 150 charge-discharge cycles without problems (without motor connected)
  
  Long term option: Redesign of BWCCPC based on alternative topology (mostly Ervin so far)
  
  Discussion on technical part
  
  Enrico suggests to investigate AC/DC converter on the market before embarking into new design
  
  look at existing commercial products
  
  look at companies available to make custom made solutions even for a small series
  
  look at CERN TE-EPC designs
  
  Ervin’s work already going into direction of first two points, 3rd point to be addressed, TE-EPC to be consulted ASAP
  
  Manfred argues on conceptual design, starting from 3-phase conversion instead of using transformer.
  
  Enrico’s comment in favour of compact design can work: we have high peak powers but for short times.
  
  General agreement on that ‘conventional' (non-switching) charging via transformer would work, but one of the key points of the project is to have a compact system.
  
  David argues on that in TEST SETUP the relay acting on the fast discharging resistor may suffer of high peak power flow (I think whit was the point…)
  
  Answer: will not be there when motor connected. Only here to simulate load
  
  General agreement (?):
  
  Continue in parallel with re-design studies
  
  Extensive tests: long cables, heating, go back to full IDC test with all different control settings (speeds, rep. rate etc…)
  
  Minimal test: repeat test with real motor as load
  
  ‘TEST SETUP' solution (cut inverter-charger CTRL link, charger enabled by external relay or enabled all the time is plugged) could work. But we need to complete validation.
  
  Discussion on planning, schedule, budget
  
  Main question:
  
  Reminder: money possibly wasted if production finished =31k (table above) when considering inverter and charger
  
  Can’t answer today, but meeting was very useful, we need to re-iterate possibly with in hand PSB - PS tentative planning and BI-PM plan for lab tests
  
  do we go ahead with series production (=PCB fabrication & mounting and then reception/tests) before fully testing ’test setup’ solution above?
  
  Fact 1:
  
            Jonathan’s delays estimate of series production from go-ahead day:
  
         Design Office
  
             1) + 3 weeks min. for standard PCB production
  
             2) + 1 month min. assembly + transport abroad!
  
         Reception by BE-BI-PM
  
             3) + 40 days work BWSAIF tests
  
             4) + 40 days work BWSMIB & BWSCPC tests
  
             5) + 60 days for assembly of complete units (2 days per unit)
  
             6) + 40 days of complete IDC crate under heavy operation
  
  is hardly compatible with PS schedule as known today
  
  Fact 2: In case of re-design: it is very likely that many of the already purchased components will not be needed
  
  Fact 3: CERN with higher priority w.r.t. ESS
  
  Next actions:
  
  Consult TE-EPC on charger advise and/or existing designs (Jonathan, Miguel)
  
  LS2 committee on Friday: see if any news on schedule (Ray)
  
  Investigate option of installing less then 8 systems in the PSB in favour of at least 1H+1V in the PS (Federico, Jonathan, Ray, Ana)
  
  Define steps and milestones to declare ’TEST SETUP’ viable at least for PSB (pre-series) (Jonathan, Miguel, Ervin)
  
  …
  
  Heating
  
  Different settings (scan speed, rep rate)
  
  Charging / discharging cycles with motor connected
  
  All this high priority for BI-PM: study if and how we can restart as soon CERN is at least partially re-open. (Federico)
  
  ….
  
  No FSUs (?)
  
  Respect instructions by health authorities (e.g. max number of people, distances)
  
  Probable constraints:
  
  Investigate with design office (Jonathan via William V ?)
  
  ~Real cost of speeding up production (e.g. going to local - more expensive - companies)
  
  If production delays affected by present worldwide situation
  
  Continue with charger re-design studies (Ervin, Jonathan, Miguel)
  
  This includes looking for commercial solutions or companies available for custom made solutions
  
  Among already purchased components purchased for inverter/chargers : are there class of components likely re-usable for future projects? (PM engineers)
  
  https://indico.cern.ch/event/907313/
  
  Sum up from Jonathan’s presentation
  
  Each LIU BWS control drive (IDC=Intelligent Drive Control unit) hosts 4 main boards:
  
  1-VFC
  
  2-Mezzanine BWSAIF (mounted on the VFC)
  
  3-Motor inverter BWSMIB
  
  4-Charger BWSCPC
  
  VFCs are of course part of the BI production and stock
  
  Components for ALL other cards are procured already
  
  Original plan (in line with pre-Covid-19 injectors master plan):
  
  Produce PCBs, assemble boards, test individually and then as whole unit for 10 Pre-series units (10=8 for PSB + 2 for FW and SW development)
  
  After pre-series validation: produce PCBs and mount series (40 units for PS, SPS, ESS + spares)
  
  Table summarising costs
  
  excluding mezzanine: PCB production and assembly ~= 31k (money lost if need to redesign not working)
  
  decision on Mezzanine BWSAIF go ahead ~independent from other two cards, but: latest version not fully tested, some risk remains)
  
  Work advancement ok until Feb
  
  individual boards tests ~completed
  
  at the moment of testing full IDC: repeated (6 times) problem with BWSCPC when controlled (as by design) by inverter BWSMIB
  
  Only when controlled by BWSMIB (i.e. charger proved several time to work when controlled by external circuit, BWSMIB always there to invert)
  
  Happens after few seconds of operation
  
  burning of Flyback shunt, MOSFET and driver
  
  after extensive investigations and tests: problem traced back to a sum of design flaws, control link between B WSCPC and BWSMIB not functional.
  
  put in place TEST SETUP in which control link between two boards is cut, external (timed) relay used to enable charger, fast discharge resistor to simulate discharging during a scan
  
  last two days at CERN (Miguel): 150 charge-discharge cycles without problems (without motor connected)
  
  Long term option: Redesign of BWCCPC based on alternative topology (mostly Ervin so far)
  
  Discussion on technical part
  
  Enrico suggests to investigate AC/DC converter on the market before embarking into new design
  
  look at existing commercial products
  
  look at companies available to make custom made solutions even for a small series
  
  look at CERN TE-EPC designs
  
  Ervin’s work already going into direction of first two points, 3rd point to be addressed, TE-EPC to be consulted ASAP
  
  Manfred argues on conceptual design, starting from 3-phase conversion instead of using transformer.
  
  Enrico’s comment in favour of compact design can work: we have high peak powers but for short times.
  
  General agreement on that ‘conventional' (non-switching) charging via transformer would work, but one of the key points of the project is to have a compact system.
  
  David argues on that in TEST SETUP the relay acting on the fast discharging resistor may suffer of high peak power flow (I think whit was the point…)
  
  Answer: will not be there when motor connected. Only here to simulate load
  
  General agreement (?):
  
  Continue in parallel with re-design studies
  
  Extensive tests: long cables, heating, go back to full IDC test with all different control settings (speeds, rep. rate etc…)
  
  Minimal test: repeat test with real motor as load
  
  ‘TEST SETUP' solution (cut inverter-charger CTRL link, charger enabled by external relay or enabled all the time is plugged) could work. But we need to complete validation.
  
  Discussion on planning, schedule, budget
  
  Main question:
  
  Reminder: money possibly wasted if production finished =31k (table above) when considering inverter and charger
  
  Can’t answer today, but meeting was very useful, we need to re-iterate possibly with in hand PSB - PS tentative planning and BI-PM plan for lab tests
  
  do we go ahead with series production (=PCB fabrication & mounting and then reception/tests) before fully testing ’test setup’ solution above?
  
  Fact 1:
  
            Jonathan’s delays estimate of series production from go-ahead day:
  
         Design Office
  
             1) + 3 weeks min. for standard PCB production
  
             2) + 1 month min. assembly + transport abroad!
  
         Reception by BE-BI-PM
  
             3) + 40 days work BWSAIF tests
  
             4) + 40 days work BWSMIB & BWSCPC tests
  
             5) + 60 days for assembly of complete units (2 days per unit)
  
             6) + 40 days of complete IDC crate under heavy operation
  
  is hardly compatible with PS schedule as known today
  
  Fact 2: In case of re-design: it is very likely that many of the already purchased components will not be needed
  
  Fact 3: CERN with higher priority w.r.t. ESS
  
  Next actions:
  
  Consult TE-EPC on charger advise and/or existing designs (Jonathan, Miguel)
  
  LS2 committee on Friday: see if any news on schedule (Ray)
  
  Investigate option of installing less then 8 systems in the PSB in favour of at least 1H+1V in the PS (Federico, Jonathan, Ray, Ana)
  
  Define steps and milestones to declare ’TEST SETUP’ viable at least for PSB (pre-series) (Jonathan, Miguel, Ervin)
  
  …
  
  Heating
  
  Different settings (scan speed, rep rate)
  
  Charging / discharging cycles with motor connected
  
  All this high priority for BI-PM: study if and how we can restart as soon CERN is at least partially re-open. (Federico)
  
  ….
  
  No FSUs (?)
  
  Respect instructions by health authorities (e.g. max number of people, distances)
  
  Probable constraints:s
  
  Investigate with design office (Jonathan via William V ?)
  
  ~Real cost of speeding up production (e.g. going to local - more expensive - companies)
  
  If production delays affected by present worldwide situation
  
  Continue with charger re-design studies (Ervin, Jonathan, Miguel)
  
  This includes looking for commercial solutions or companies available for custom made solutions
  
  Among already purchased components purchased for inverter/chargers : are there class of components likely re-usable for future projects? (PM engineers)
- 11:30 → 11:40
  
  Next steps 10m