Peter presented his studies about the alignment of the chips by using straight tracks. There is some strange pattern in the residuals, when you plot them with respect the the z-position of the hits. We could not find an explanation... He thinks that spoils the sagitta resolution, which should be a factor of 2.2 better than what he gets from data.

Amir shortly explained what he is trying to do. He wants to reduce the number of hits for track finding (and olny for finding), by building "superpixels". Depending on the condition, such a superpixel can have a variable size. So e.g. 50x50 real pixel. From all hits in this area, a single hit is calculated which the center of gravitiy.
He hopes that one then can use track finders like they are already used for the pad based readout, as those are in general better developed then the ones for pixel.

Michael just showed one slide, in which he displayed the number of dead chips together with the activation of the magnetic field (the shaded area in the plot is when B was on). It looks like from the end of the first runs with magnetic field, chip dying starts. At the test beam, there was a fast ramp down of the magnet at that time due to a interlock break. The inspecion of the dead chips, however, did not show anything suspicious, also not at a brief look at the wire bonds, which could possibly have moved. But Michael thinks he needs to have a far more detailed inspection to find some hits. 
In one of the last meetings, he was asked, what could have caused the sudden start of chips dying and Jan proposed, that the magnitic field could be a reason.
So he did this plot, and somehow, there is a relation. He does not know if the Bfield directly is the reason for chips dying.
So he proposed to put so chips in a magnet, while they are operation and then cycle the field. Jan said that at NIKHEF, they have a small 1T magnet, where this could be done.
But we did not talk about who exactly should do this.

Next meeting will be on the January 28.