• The position accuracy of the D-substrates with respect to their nominal position should be 25mm along the input bonding pad row. The precision orthogonal to the pad row should be better than 100mm.
• Reference marks on the M-substrate (thick film) have a precision of about 25mm. The M-substrate has reference marks on the position of all DS corners. Along its edge the M-substrate has reference slots (laser-drilled)
• The DS must be removable from the M-substrate (in case of failures) with no damage to the M-substrate.

Gerrit proposed to see if the DS positioning along with the bonding can be done elsewhere. Bernie will contact relevant companies.

An in-house assembly would require a "pick and drop" assembly setup where the DS are taken from a slot by vacuum jig on a positioner , the MS is positioned below the arm and the DS placed on the MS. Bernie proposed to align the DS with respect to piece of silicon which has the bonding pad rows on and use them as reference

The main problems seem to be (a) glue spreading over the IO lines and (b) how to remove the DS if it doesn’t work

1. Bernie proposed to use adhesive film, which has no glue spread, but doesn’t allow any removal
2. Heinz proposed to stick the DS with heat conductive paste to the MS and fix the position with 3 small drops of epoxy glue. By heating up the glue drops this may allow to remove the DS. The heat conductive paste might spread out under the DS

Both methods need to be tested.