Requests for Quotation (RFQ) went out from LANL to three silicon vendors in February, 1995. The three vendors were: CSEM, Switzerland, Hamamatsu, Japan, and Micron Semiconductor, England. Prior to sending the RFQ, we met with each of these three vendors, and held detailed discussions relating to the specific design of the PHENIX MVD.
By mid-March, the vendors had submitted bids. Each bid includes engineering design time, mask production, prototyping, testing and production. The bid totals are:
| CSEM | $566,000 |
| Hamamatsu | $439,000 |
| Micron | $287,000 |
Based on the bids alone, Micron is the clear choice. However, not just cost, but deliverability must also be considered. This specific concern was raised by a member of the TAC in the review at BNL in March. Indeed, Micron has a very large commitment to D0; to supply them with 1200 silicon detectors over the course of the next year. In fact, their prototype and delivery schedule almost exactly overlaps our own. CDF expects to order about 750 silicon detectors in the next year or so. They have not yet made a vendor choice, but they anticipate that they will split their production among two vendors. In comparison, the MVD order is for 150 detectors. Although as part of their bid, Micron has agreed to a specific delivery schedule, clearly there is some risk against their ability to maintain that schedule, especially if the D0 production meets with some extraordinary problem. We have looked to our overall schedule, and found that we can accept a 6 month slip in silicon delivery under our current profile without affecting our completion date.
There are potentially some positive factors associated with the D0 silicon order. Micron has added an in-house ion-implantation machine to streamline their production capability in anticipation of large orders, such as this. Also, D0 has chosen a bias technology; polysilicon, and a detector feature; ac-coupling that is identical to those chosen for the MVD. The quality of our detectors will benefit from the processing development that has already gone, and will continue to go into the D0 detector production.
In conclusion, we have chosen Micron as the silicon vendor for the MVD. We acknowledge there is risk to our delivery schedule in choosing a vendor that has large outstanding commitments. We think that the 6 month cushion in our overall schedule allows us to accept this risk. The large differential in cost between Micron and the next lowest bid figures significantly in our evaluation of risk Vs $cost.