MOST2 chip design meeting

Asia/Shanghai

1. Tianya reported some issues in the digital cell design.

     i. The delay of the read bus seems large, and the delay of the rising edge and falling is not balance (23ns and 15ns respectively). This is maybe due to the accumulation of the unique of the tpLH and tpHL of the buffers. (possible solution is to replace the buffer chain into inverter chain)

     ii.some issues with the address encoder. will discuss with Ping Yang about this.

  2. The design specification of the periphery was further discussed.

     i. The periphery only needs to prepare the triggered data in FIFO. the LVDS related design will be considered in the serializer and data interface block. It is better that the "data available" bit comes at first, and the possible parity bit comes at last. 

     ii. The readout scheme of the FIFO was discussed. Originally it was designed to have a "average speed data rate clock", but it can be "read the FIFO whenever it is not empty". Actually there will not be this average speed clock, but only trigger. 

     iii. the FIFO will only deal with 40MHz system clock, the 160Mbps clock will only provided to the serializer. serializer will use the "FIFO_empty" as the indicator to read the FIFO. It seems that 8b/10b coding is not the critical block in the current stage.

     iv. the size of the chip FIFO will be further calculated. Currently it is designed to be 512 events, and seems ok. The real size of the matrix is surely not possible to be 512*1024, and should be shrunk. However we expect that the periphery aim at the final functionality.

     v. In the first tapeout, analog+digital+periphery should be integrated. We will see if there is enough space to place 32 parallel output pads.

  3. Zhang Ying reported some issues in the analog design. The circuit timing performance is sensitive to the process, especially when we chose "standard Towerjazz"(used in MOST1), and "modified Towerjazz"(used in ATLAS upgrade). For CEPC, it is only possible to achieve the timing performance (and radiation performance ) by "modified TJ".

  4. Therefore, if, when, and how we can use the "modified TJ" process is very critical now. Zhijing will ask through the ATLAS tunnel.

  5. We agreed our meeting still fixed to 3:30pm Beijing time... 

There are minutes attached to this event. Show them.