4.1. Distributed Parameter Database
Each controller in a Servo Network is part of a distributed parameter database that is invisibly managed by the master controller. GSB and GIO boards are not controllers, so they do not contain a parameter database.
Slave controllers keep a local copy of their parameter database, but the values are automatically written and read to/from the slave by the master controller and a complete set of the configuration data is kept in .pac files on the master controller’s flash disk. Any .pac files on the slave controller’s flash disk are ignored. When the master reads or writes the .pac files from its flash disk, or a user displays or modifies configuration data via the web pages on the master controller, the master controller dynamically exchanges data with the appropriate slave controllers in a transparent manner to update them. Therefore, slave controllers must be present in the network in order for their data to be displayed or modified.
If any slave controllers are inactive or missing from the network, the .pac file data entries corresponding to those controllers are displayed as 0 or null strings (""). It is important to not save configuration data to flash memory unless all slave controllers are active.
4.2. Datalogger Performance for Serial Bus Boards
Since the datalogger streams data on the serial bus from servo boards, the amount of data that can be logged from a serial bus node depends on the number of nodes present due to the overhead of continuous communications between the master and slave boards. The table below shows the measured and estimated number of bytes that can be simultaneously datalogged. The number of data bytes needed for a given database parameter depends on its data type.
Table Of Data Bytes That Can Be Datalogged From A Single Node Given The Trajectory Rate, # Of Serial Slave Nodes, And Sampling Interval
| Trajectory Period | # of Serial Slaves | Datalogger Sampling Interval | |||
|---|---|---|---|---|---|
| 125 usec | 250 usec | 500 usec | 1 msec | ||
|
1 msec |
1 |
none |
8 |
16 |
32 |
|
2 msec |
1 |
16 |
40 |
80 |
160 |
|
|
2 |
8 |
24 |
68 |
|
|
|
3 |
none |
12 |
32 |
|
|
|
4 |
none |
none |
4 |
20 |
|
4 msec |
1 |
20 |
60 |
120 |
240 |
|
|
2 |
20 |
56 |
|
|
|
|
3 |
16 |
40 |
|
|
|
|
4 |
12 |
32 |
|
|
The table above assumes logging from only one GSB node at a time. If you log from more than one node, there is additional overhead that will reduce the number of data bytes that can be logged. For example, if the table shows you can datalog 40 bytes of data from a single node, you cannot simultaneously log 20 bytes of data from one node and 20 bytes from a 2nd node. If you attempt to log more data than is permitted, GPL issues the error: "Too many remote data items" (-1558).