industrialNETworXnetx

houzenan

houzenan

| 25.11.2009 | 16:56 | 4 replies

About the healing time of redundant ring topology

according to the SerCos III standard, the healing time of redundant ring topology is 25us.

Does comx conform to this?

Andreas Jacob

Andreas Jacob

Hilscher Gesellschaft fuer Systemautomation mbH

| 25.11.2009 | 17:06

Hi,

please have a look into the SERCOS API manual. This information should be there.

houzenan

houzenan

| 25.11.2009 | 17:33

:( I have hunted for this in sercos iii API master&slave document twice.
:oops:

Johnny

Johnny

| 26.11.2009 | 09:57

Hello houzenan,

The answer is a little complicated:

1.)
Hilscher SERCOSIII Master and Slave supports the redundancy. Redundancy means that you have an "double ring". You can remove ONE of the cables at any time without disturbing the communication (maybe one bus cycle is broken, but not more).

2.)
If a cable is removed, both affected ethernet ports detects a Link Down. If link down is detected too late, the frames would be sent over a dead port, so it must be detected quickly. The detection time depends on the used Ethernet PHYs. The PHYs used in the netX (netX is used in all Hilscher SERCOSIII products) have a very small detection time (< 25 us). I guess you meant the link down detection time.

3.)
Double ring recovery takes longer as 25 us in any case. How does it works:
The user plugs the missing cable again. Ethernet Link up is detected (takes up to 2 seconds), master detects the changed topology and request from the affected slaves to change the telegram flow (ring healing). Slaves shall do this immediately.
Double ring recovery is not time critical:
User gets a new ethernet cable from the stock (5 minutes)
Master heals the double ring (3 seconds)

Johnny

houzenan

houzenan

| 26.11.2009 | 16:35

:lol: yes! I mean "link down detection time", not "healing time". I've made mistake.

how professional you are !

Login