Post production changes (February+March 2020)
After noticing that the clock signal is a factor 10 too high the decision was made based on spice simulations and oscilloscope measurements to add a resistor between the two LVDS clock lines on the Kapton Flex connector with size of 20 Ohm.
These changes were performed in March after initial tests were performed in February. As advertised the resistor dampens the clock from 700 mV mean peak to peak down to 120 mV mean peak to peak (measurement below is measuring the spikes that is why it is higher). This was done for 2 of the 3 boards.
Post production changes (May 2019)
- The cassette boards provide 1.8V for DVDD, we want them to provide 2.0V to change this one resistors needs to be exchanged R2 with a 154k Ohm resistor. A 160 kOhm resistor results in 2.03V which is also fine
- The humidity and temperature sensors are not currently equipped with the necessary resistors to run correclty, we need to add R7 and R8 with 4.7kOhm resistors (as written in the schematics, simply not equipped). One board was already changed
- During the test beam we noticed that we cannot run two KPiX on the same sensor in self triggering. They trigger each other repeatedly, the idea is to move/remove a resistor to ensure it more properly reflect the system with the old DAQ to see if this problem persists.
- Two of the boards were modified using silver epoxy to produce a ground connection between the board and the cassette.
Parts and Pieces:
New design from SLAC (Up to date as of 07.02.2020)
The newest version can always be found on the SLAC confluence ( https://confluence.slac.stanford.edu/display/AIRTRACK/KPIX+PCB+Designs )
Updated on 5 October 2018:
Because of the stiffener on the Kapton Flex, all boards had some minor changes, some capacitors were exchanged with smaller version to ensure every cable can be plugged in.
Master Cassette board schematics (final version):
- Master board schematics v.14: desy_tracker_cassette_master_c00_new.pdf
- Master board schematics: desy_tracker_cassette_master_c00.pdf
- Slave board schematics: desy_tracker_cassette_slave_c00.pdf
Note from Ben on :
- Each cassette sends its KPIX_DVDD (1.8V) and local DVDD (2.5V) back to the DAQ board so that that the DAQ board can isolate the digital signaling onto those voltages/grounds.
- Cassettes now regulate a separate AVDD for each KPIX.
- - Rather than 6 a separate LT1763 regulators, I now use 3 LT3029 dual-output regulators. This saves a lot of space.
Reply from Ben on
- The extra 0.1uF caps on the LT3029 probably aren’t necessary. We used them before on the LT1763 because it can help with noise and we had space to add them. I think the noise performance will be ok without them.
- The 0.1uF cap on BIAS_P/BIAS_N is part of an RC filter. It is the same filter type as the 3-level RC filter on the cassette. You might think of this as a fourth leg of that same filter.
- How cassette ground tie to AGND?
- Earth Ground is not shown on any of the board schematics. We will attach a wire from some Earth Ground reference point to a solder pad that ties to AGND on the cassette.
- Each cassette has 2 ground planes – AGND and DGND. The AGND symbol is a triangle, the DGND symbol is a triangle of 3 horizonal lines.
- AGND and DGND need to be tied somewhere, and the plan is still to tie them at the end of the flex cable only. We have 0Ohm resistor options to tie them in all sorts of other places though.
OUT OF DATE: Technical drawings showing connector positions and sizes needed to fit into the cassette (mechanical limitations) of the cassette boards with the connectors are below:
Need connector look:
Below are the drawings from SLAC about the cassette board. Due to size limitations they have to change the used linear regulators to smaller ones. Therefore the drawings may change.
Latest schematics from SLAC, update on Jan 29, 2018