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

[John Parsons <parsons AT nevis.columbia.edu>]

Hi Gustaaf, all,

        I give a few comments below about LAr below.

                John

- Sec. 2.7.3 - I suggest to add the question about whether they used the 
full production quantity for the FEB2 PCB fab and assy quotes.  Even 
that would not explain their high numbers, but it would go in the right 
direction at least.
  (I note that using the ACTUAL LTDB costs paid by BNL for Ph. 1, which 
has a much smaller quantity than FEB2, would still give variances of 
220% and 50% for fab and assy respectively, compared to the ICE post-rec 
numbers. So it is still a mystery why their quotes are so much higher 
than ours.)

- For the rad-tol customs parts, they accepted our CERN pricing for 
lpGBT and VTRx+ (Tab. 2-8), which is good.  However, for some reason 
they did not do the same for the ADC (Tab 2-6). Instead, they state "the 
cost reduction for the production ADCs is changed from 95% to 97.8%. 
These cost reductions are applied to account for sole source pricing of 
these devices."; this makes no sense to me and seems completely 
arbitrary.  It is true the CERN Frame Contract with TSMC is sole source, 
so if they want to assume a higher risk that is fine. However, as for 
the lpGBT, all that really makes sense is to assume the CERN pricing as 
the central value.  You really can't expect producing your own custom 
chips to cost the same as buying commercial chips.  We should ask why 
they do not accept the CERN quote based on the TSMC Frame Contract.

- as Mike pointed out already, they made an error in Tabs. 2-5 and 2-6, 
where the ATLAS costs should of course be the same for pre-/post-rec. 
Looking at the latest BOE (ie. for FDR), our M&S costs for ADC 
production are $1070k, which somehow sit between the 2 values used by 
ICE, so I am not sure exactly what they are using (I couldn't quickly 
find again the table of costs that was prepared for ICE reconciliation, 
which perhaps was still using an older version of the BOE?). So, the 
post variance is likely even higher than the 27% they are quoting, and 
is more likely ~50%, making even more important the question above about 
why they don't accept the TSMC Frame Contract quote.


On 1/11/20 6:21 PM, Gustaaf Brooijmans wrote:
> Hello all,
>
> Here's a first draft set of comments and questions for the ICE team.  We 
> need to send this on Monday.
>
> Best,
>
> Gustaaf
>
> _______________________________________________
> Usatlas-hllhc-l2deputymgmt-nsf-l mailing list
> Usatlas-hllhc-l2deputymgmt-nsf-l AT lists.bnl.gov
> https://lists.bnl.gov/mailman/listinfo/usatlas-hllhc-l2deputymgmt-nsf-l

--
______________________________________________________________________

John Parsons
Nevis Labs,             Email: parsons AT nevis.columbia.edu
 Columbia University    Phone: (914) 591-2820
P.O. Box 137            Fax: (914) 591-8120
Irvington, NY 10533     WWW: 
https://urldefense.com/v3/__http://www.nevis.columbia.edu/*parsons/__;fg!!P4SdNyxKAPE!Utp-tbIwBUyw7qKUtn017T6FwbYsXvlbttnSJJ9291lnLgNhA-zvE7Zn5SQ85zsbtrCj3VVXYOrP_rcsy6XLMUEpClzM$
        

______________________________________________________________________