Simulation waveforms of strip signals show a wide variation in pulse shape dependent on event position with a unit cell. The goal is to find a signal processing technique that can be used to determine which strip the event occurred nearest to.
The strip signal peak-to-peak amplitude is not proportional to X. In fact it tends to increase as the event position moves away from the center until it reaches the edge of the cell. The shape changes drastically near the corner of the cell presumably due to the strip collecting charge instead of a pixel.
These images show relative amplitude of the strip signal vs. X,Y position at a few different Z depths. The lower left corner of each image is the center of the pixel. The white space in the corner represents positions where there is no pixel signal and the strip collects charge.
Using peak-to-peak gives a more even response for events near the anode. This image shows the difference of the falling edge (peak-to-final).
An approximation to peak-to-peak can be generated by differentiating and inverting. The weakest signal is for Z=0. Something around 100ns time constant should give the best signal to noise for the slow pulse seen at Z=0.
These images show the relative response using the peak from a shaper with a CR differentiator and RC integrator.
The faster time constant looks slightly better but it would also pass more noise and require more power.