Signal passing through Network Bridge

Hoping someone can shed some light on a strange little situation I've seen. I have two networks connected together through a network bridge. The connect application is checked on both sides of the bridge so commands may be sent between the two networks. The local network has a PC interface (CNI actually) on it which is connected to a Crestron controller. The Crestron controller can see commands being sent from the remote network ok (I've added the extra byte to allow for the bridge routing) and everything seems to work well.
I've also noticed that when a bridge passes a command, the indicators flash twice. Likewise, the PC interface indicator will flash ( I have it in smart mode) to say it has passed the information on too
This works ok until.....

If either side of the bridge loses comms (ie you briefly unplug the bridge) and then re-establish comms (plug it back in), the first signal sent thereafter, does not seem to be passed by the bridge. The indicators only flash once and the PC interface doesn't signal at all!!! It seems to kick back in after that ok.

Is this correct operation or a bug?

 

This is in my completely UNeducated opinion, but I'm guessing that what you are describing is normal operation.

I cant give you a definative answer, but coming from an IT background, it sounds entirely feasible for it to be normal operation.

It could be something as simple as the first command "initiating" the connection..

Hope one of the guys on here can give you a definative answer, sorry if I havent been any help to you, lol

 

It's hard to say

The bridge is constructed with two isolated and identical circuits, which communicate via an optical link. Messages are passed between the two halves according to a defined protocol.

If the power is removed from one of the circuits, the whole bridge will simply stop working. If one side is still operational, there is a good chance that communication will be attempted, and retries will be initiated, etc. When the other side is powered up again, it will initialize all buffers to be empty, and will expect a valid message form the other side. I suspect what is happening here is that brute-force re-connection may result in some transient signals in the optical links, which might be interpreted as negative acknowledgements, or invalid and partial messages. These invalid segments will just sit there until a message with a valid delimiter (end-of-message marker) is seen. The garbage plus legitimate message is probably being treated as an invalid message.

Not sure here, but it sounds pretty normal under the circumstances.

Don

 

Thanks Don

I have tried it quite a few times with the same result each time. The simple setup only had a key input unit on the remote network so no traffic was attemting to go through the bridge whilst it was 'offline'. It was only offline for a second. I simply reconnected it, pressed the key input unit (I tested this by pressing the button immediately and also waited up to 30 seconds with the same result) with the first on (or off) command not being passed by the bridge.

This might be important on a particular project where C-Bus power dropouts on a remote network are common.

Cheers


Chris

 

does it behave the same way when the entire "remote" network is powered down, then up again? Disconnection and re-connection of the bridge may not result in the same behaviour because C-Bus power supplies have a limited slew rate on power-up.

 

Hmmm Good question Don. I'll need to check that one out. I'll set it up again tonight and post back tomorrow