Timing
Bunches are separated by 11ms. Normal operation will have 12Hz of spills but for special dedicated g-2 running we should support 24Hz.
- 2 tracker stations
- 1000 straws (channels) per station
- 10 hits/fill/tube (*not* uniform!)
- 8 bytes per hit
Assuming we read out in the 11ms between bunches, the data rate for a 16-channel TDC would be 8*16*10 = 1280 bytes / 11ms = 115kB/s. This is average. If we want i.e. a factor of 5 safety this would be 580kB/sec or about 5Mb/sec. The largest detector has 4 TDCs per side, so if all shared a readout link this would be 20Mb/sec.
A total of 18 such links would be needed per tracker station, or 36 total. The total estimated bandwidth for one tracker station would be 580k * 58 = 33.6MB/sec. With a 5X safety factor this becomes 168MB/sec. Half of this is 84MB/sec for each side, which plausibly fits in 1GbE. So, at a minimum we would have one 1GbE cable per side per station (4 total).
It is probably more conservative to double the number of GbE links. A module (VME?) could be designed with 6-8 links and one GbE output.
Each station would have about 9 straw detector assemblies. One preliminary design allocates the straws as follows:
| 'Station | Straws | ASDQs | TDCs |
| 1 | 32 | 4 | 2 |
| 2 | 64 | 8 | 4 |
| 3 | 96 | 12 | 6 |
| 4 | 96 | 12 | 6 |
| 5-9 | 128 | 16 | 8 |
| Total | 928 | 116 | 58 |
