U.S. ATLAS HL-LHC Upgrade Level 2 and Deputies-NSF only Management Mailing List

[Gustaaf Brooijmans <gusbroo AT nevis.columbia.edu>]

Hi John,

> 1. the ATLAS SVR for 6.4.3 BE is scheduled for Nov. 2020
>      - I will argue that the BE is based very closely
> on Phase 1, so the "v1 preproto design" which will be
> complete before Apr. 2020 will allow us to meet the
> "construction ready" mark.  However, given that the
> HL-LHC Specs for the BE will not even be reviewed
> until Nov. 2020 leaves this argument quite open to
> challenge
>      (I guess there are also some other subsystems
> with some SVR dates after start of MREFC?)

Yes, but there is no technological risk here, eg the 25 Gbps is now 
proven, and the first LASP board will be made at the end of this year. 
I don't see a real problem here, at least not more than for any of the 
other ATCA boards.  LAr is in fact ahead of HTT and L0MDT on the ATCA 
front.  Another way to look at this is that the comparison standard is 
FEB2, not ADC.

> 2. a low probability but high impact risk for the ADC
> in 6.4.1 is that it fails somehow and we have to fall
> back to a COTS device (with much higher cost, ...)
>      - I will argue the 2 rounds of preproto so far
> have reduced this risk, and the v3 preproto that will
> be submitted and tested still as part of preMREFC
> will get us to where we can retire (or be forced to
> unfortunately realize) this risk.  However, what if we
> are not 100% confident to make this decision before Apr. 1,
> particularly given the NSF restriction on spending
> contingency unless you have an associated risk?  ie. if
> we retire it, and then find a problem a month later
> that anyway forces the COTS backup solution, where do
> we get the ~$2M needed??  To guarantee we don't fall
> off this cliff, I would of course like to keep the
> risk active, albeit with a Very Low probability.  But
> how do we say this does not contradict the clause to
> "eliminate all technological risks" ?

Also here I don't see the problem.  Of course you can still have this 
risk.  The point is that you have a fall-back solution if your preferred 
option fails.  So you know you will be able to build FEB2.  Only if you 
did not have any fallback would you be in trouble.

Best,

Gustaaf