Private:Weekly Group 10/31/2007

= Minutes =

Matt's Update
(busy week, esp. studying for GRE Physics) otherwise:
 * looking for literature in Simulated Annealing (SA).
 * focus: understanding how to direct our search
 * general perspective on how real problems are approached/broken down with SA

Carl's Update
Order almost worked out. I am still looking for a company that carries less then 4800 pipette tips per case.
 * trying to deposit a small/consistent enough drop on tip of fiber with liquids of similar viscosity to that of optical epoxies using conventional applicators (e.g. toothpick)
 * as expected - small amount application possible, but no consistency.
 * looking into better applicators - types of pipettes.
 * &plusmn;0.04mL precision graduated pipettes found with &plusmn;0.01&mu;L precision graduated tips available. The pumps are added separately. Fine control (non-discretized) pumps found for these pipettes. From past experience, reasonably confident that it will have the fine control to load up the tip in a measured way. (Note that for a drop ~mm2 = &mu;L, so &plusmn;0.01&mu;L precision suffices)
 * status of gluing apparatus:
 * considerations of cutting and stage 1 (cutting and gluing scintillator to waveguide): 2mm channel has been widened by the machine shop to avoid scratching the fiber. The edges still need to be rounded. Note that with the current scheme of the curing joint suspended in a gap of the channel (no epoxy in channel itself) eliminates mold-release issues of this gluing stage. (Release from this channel was the driving motivation for solid PTFE block.)
 * considerations for stage 2 (gluing 5 fibers together): machine shop insists that the small wedge (1cm2 cross-section with sharp, thin tip) will be troublesome to machine out of solid PTFE but the current aluminum wedge can successfully be coated (after machining down the wedge appropriately). It is very easy to reapply PTFE tape on the edge of the thin slab (which applies pressure on fibers while curing) so this component is not an issue.
 * considerations for stage 3 (gluing rows in staggered formation with the above mentioned wedge taken out): the relevant area of the bed of the apparatus can have an inlay of PTFE.

Igor's Update

 * Some old reusable modules adjusted for previously discussed timing issues. All modules for which it was seen inconvenient, impractical or downright counterproductive to delay "Go" pulses on which data is latched (e.g. some registers) the latching now occurs on falling edge of clock. All this was tested with various shifts, delays etc. Functionality in dependencies not effected because all retrieval happens on rising edge.
 * Brendan's DAC controller was reviewed, including issues he had with timing (he had to insert an artificial delay in DAC Sync signal because the accompanying shifter was always late by one cycle). To ensure that this delay is reproducible, the parallel data (that is serially shifted out) is made to latch on the falling edge. The module was simulated and shown to behave as before with timing more assured, in Igor's opinion.
 * The DAC Programmer (which drives the DAC controller) is nearly complete and in the process of simulation and debugging. After this module, the Transmitter, Transceiver and Reset remain. The Transmitter should work analogously (and in reverse) of the Reader and Programmer and will not involve learning much of anything new. The Transceiver and Reset, on the other hand, represent the next stage of understanding of the Ethernet Controller. The former involves a new bus protocol and the latter needs to take into account all the reset needs of other chips and modules on a proper schedule and handle the addressing scheme we have devised.
 * Brendan's pages on VHDL work are slowly being augmented to include development in the last two months.