If all power and communication lines are shut off to a particular drop of a DeviceNet network and then re-applied after a certain time period, why would I see different time periods in which the communication is re-established with the nodes on that drop?
Answered by Ray Romito, DeviceNet Trainer, e-mail: raromito@ra.rockwell.com .
A180) There are several factors that affect this: Reconnecting with a slave requires a very specific sequence of messages. The scanner must first attempt to converse with the node using the UCMM method (which is optional). If that does not successfully complete within one second, it tries again. If the second attempt also fails then the scanner reverts to an alternate method which will probably work on the first try. Thus, if the drop has a mixture of UCMM and non-UCMM nodes, the non-UCMM nodes will take at least 2 seconds longer than the UCMM capable nodes. This process is defined in Vol 1, chapter 7-9.1 if you have the Specifications and would like to know more. Another factor is how often the scanner attempts to reconnect to a missing node. This will vary from one scanner maker to another, so it^s hard to characterize it across the board. Some scanners may attempt to reconnect sequentially by node number once every so many seconds while another scanner may do it in a different order, more quickly. A node may have just missed an attempt, while the next just makes it. The difference between reconnecting in this case is how long the scanner takes to retry. There is no specification as to when communications should be re-opened after a time out, or how often it should be retried. The last items is how long it takes the nodes to power up, do their power on self-test and get themselves ready to accept messages from the wire. If all the devices on this drop are the same, then this is obviously less of a factor than if they were all different types of devices, potentially from different manufacturers. There could be quite a bit of variation.