5/13/04 LG

 

 

Some Constraints on the Flex Printed Circuit Cable for the MVD and Current Status of the Design.

 

 

Mechanical Constraints

 

The MVD detector is designed to fit into a very constrained space between the (new) beam pipe and the existing SVT detector in the radial direction and is constrained by the forward TPC in the beam direction. The mechanical access is quite limited and the space constraints are severe. This is illustrated well in the proposal.

Due to these space constraints, we are unable to locate printed circuit boards of any significant size (much larger than the cable itself) at the point where the ladder ends. The signals must be carried approximately 0.6 m from the end of the ladders to reach a point where we can make transition boards, perhaps LVDS to optical fiber link as well as locating a control/clock/housekeeping board for each ladder. The exact size and requirements for such a board are still in design. Based on this, the cable that we will be building will be approximately 1 m in length.

In addition, we need to keep the interaction length of the cable to a minimum. A study of the interaction lengths of the various detector components including cable and carrier designs is available here http://www.lbnl.leog.org/pixel_rad_length.pdf . We have recently found a vendor who will give us aluminum on kapton flex pc with vias. This is the structure of our intended cable design as shown in figure 4 of the above writeup. The calculated radiation length for this cable with the materials shown is 0.10%.

 

 

Cable Design

 

September Test

For the September test, we intend to produce one complete carrier based on the use of MIMOSA5 detectors and, some discrete electronics bonded to a carrier/cable assembly that will be a reasonable approximation of the final cable. It is intended that we gain the necessary experience in detector handling, aluminum conductor flex-pc, assembly techniques, etc. to ensure that we can make a final detector.

To this end we will produce a test ladder that consists of a carrier large enough to hold the discrete components that we need and a cable to carry the amplified differential analog signals to the altro based digitizer chips. A schematic mechanical design is shown below.

 

 

A simplified electronic schematic for this September test is shown below.

The schematic for each MIMOSA5/buffer/diff driver is shown below here http://www.lbnl.leog.org/SCHEMATIC1 _ PAGE1.pdf

 

A list of the signals/power carried on the September test cable is shown here

http://www.lbnl.leog.org/signal_list.pdf

As one can see from the list, we expect to run a separate analog and digital power traces to each MIMOSA detector to allow remote malfunction remedies at the detector level.

The parameters for the test cable are as follows.

 

trace width

0.005” - 0.008”

trace space

0.005” - 0.008”

cable length

~1 meter

Aluminum thickness

0.0007”

Kapton thickness

0.001”

Layers

4

 

The present status of implementation is that the cable is in layout. The carriers are in fabrication.

A schedule document is posted here

http://www.lbnl.leog.org/pixel_ladder_3.pdf

 

Final Cable Design

The final cable design is expected to include Robin’s readout chip. The mechanical constraints described above still apply and the cable length and construction techniques will remain the same. We will, however, need fewer traces to read out the data. A preliminary list of the signals and power can be found here.

http://www.lbnl.leog.org/signal_list_final_cable.pdf

In the case that we would need to drive the analog signals down the cable, a list similar to the September test cable can be expected. In this case, the actual shielding and impedance of the cable is much more of an issue.

 

Impedance

We are using standard Texas Instruments THS4501 differential drivers and are designing the cable to have a pair impedance of approximately 100 ohms. This is achievable using conventional construction techniques and materials. Some impedance calculators can be found here;

for striplines:

http://www.emclab.umr.edu/pcbtlc/

 

for differential pairs:

http://www.ideaconsulting.com/dstrip.htm

 

We appear to be operating at the edge of common construction techniques and abilities to reach the 100 ohm pair impedance that we believe that we can achieve. In order to significantly increase the pair impedance we would need to move to very small trace widths/spacing, large cable thicknesses (giving much larger radiation lengths) and exotic materials such as PFTE based prepregs and boards. Construction of cables of this type with aluminum conductors is a significant engineering challenge.