TCIM Changes in August 2001

1) This change was suggested by Pat McGaughey. The 74act827 which drives the 10 MHz clock signal on the VME backplane is not spec'ed for the amount of capacitance it sees when trying to drive 20 DCIM boards. Therefore, we "doubled up" two input and outputs from the chip on the TCIM (sorry I do not know its name on the schematic right now) to improve the drive capability.
2) This change was also suggested by Pat McGaughey. This was not a change to the TCIM, but a change in the termination of the clock lines on the VME backplane. The changes attempted to pull the clock line to ~ 2.5 Volts. Some connectors, which fit on the backplane, were made which did something like this: 

   view at back of crate

  +5                        +5
  |                          |
  R                          R
  |                          |
  -------clock-line-----------
  |                          |
  R                          R
  |                          |
  Ground                     Ground

  where R = 1000 ohms. Impedance = R/4 = 250 ohms.

Actually, we did not use exactly 1000 ohms, but I forgot what we did use. We changed the diagram a little because the power supply to the crate is at 6.5 V, not 5.0 V and we wanted the clock line at 2.5 V.
3) This change commected previously independent grounds for the Glink part of the circuit board and the rest of the board. The orignal clock comes in in the Glink part of the board. The signal then goes to the rest of the board which distributes it to the MCMs (from the front panel) and the DCIMs (through the VME backplane). The signal went from one part of the board to the other, but their grounds are indenpendent on the board itself. There seems to be a substantial (0.5 to 1.0 V) 40 MHz oscillition between these two grounds. This is/was surely part of the problem. There is a little confusion about the magnitude of the oscillation because the wire I attached to the scope probes to reach beteen the two grounds picked up this clock also. We connected the two grounds together with a wire on the board. Probably this connection could be made in a better place.

John Sullivan
last update: 15-Jan-2002