how fast is the C-Bus ?

Discussion in 'General Discussion' started by 2SC, Jul 9, 2009.

  1. 2SC

    2SC

    Joined:
    Oct 10, 2006
    Messages:
    345
    Likes Received:
    0
    Location:
    Athens, Greece
    Being a System Integrator of C-Bus the last 5 years we have commissioned more than 60 c-bus installations some of them with more than 80 - 100 units.
    I know that the C-bus is really fast (I mean the rate of transmitting signal) in reverse of EIB that can transmit only 7 commands per second.

    Is there available that kind of technical data about C-Bus because a E/M design office asks about these data...
     
    2SC, Jul 9, 2009
    #1
  2. 2SC

    ashleigh Moderator

    Joined:
    Aug 4, 2004
    Messages:
    2,397
    Likes Received:
    26
    Location:
    Adelaide, South Australia
    The cbus transmission speed depends on the SIZE of the commands.

    First - we will assume lighting commands, which are small.

    There are 2 cases to consider.

    Case 1. How many SINGLE lighting commands can you send?

    Answer: A lighting command is either 2 or 3 bytes long, depending on the command. The message header, routing, checksum, and so on is BIGGER than the command. But in this case, sending about 20-30 commands / second is possible. It takes a lot of very very very careful software coding to achieve this. Using a PCI as the interface, you will most likely in normal cases achieve 10-15 commands/sec by trying very hard and being very careful. It is possible to do more, using a PCI, but the techniques to do this are unusual and require a lot of effort. Clipsal don't do that extra effort.

    Case 2. What about scenes?

    In the case of scenes, several lighting commands can be packed into a single message. For example, a single message might be able to control between 4 and 10 lighting groups at a time. The number depends on the network routing, and the commands. Using 6 at a time as an average is a reasonable guess.

    In this case, the overhead in sending the message is the same as the case of the single control, as above. But you can control 6 times as much per message.

    Because the messages are a bit bigger, they take longer. Typically these big messages, in a scene, might take twice as long to send. So by being really careful, you can send about 8 to 10 BIG messages per second. But each of those BIG messages controls 6 groups on average. This means you can switch or ramp approximately 60 groups / second.

    Doing this is difficult, and typically you won't get this performance. But being able to switch or control about 15 to 25 groups / second using big messages (as used in scenes) is reasonable by writing code to drive a PCI, and doing it very carefully. Being very very very careful and using some obscure techniques you can increase this to perhaps 20-30 groups / second, or maybe even a bit more. However, this is theoretical and clipsal dont do this.

    SUMMARY:

    It depends on what you want to do. If you cover raw messages / second, then it depends on the message and its size. 10-15 messages / second is achievable with a lot of care, provided those messages are short.

    If you look at what the messages do... that is HOW MUCH CAN I CONTROL, then the messages get bigger but each message does more. 15-30 control actions / second is achievable.

    These numbers are all theoretical, and to achieve them in practice needs careful writing of software.
     
    ashleigh, Jul 9, 2009
    #2
  3. 2SC

    Don

    Joined:
    Aug 4, 2004
    Messages:
    429
    Likes Received:
    0
    Location:
    Townsville, Australia
    .. more info on speed

    Everything Ashleigh says is correct, but it may be of interest also that messages generated by simple units themselves ( Neo, Saturn, Multisensor, bus coupler, etc) are processed at the highest possible speed.

    If a key is used to initiate a scene of 40 groups ( the maximum that can be put into a single scene in a current production wired C-Bus unit), that scene will change the state of all the groups in well under 1 second, even on a moderately busy network. This is achieved because these units assemble long messages to make the best use of the bus bandwidth (limit is around 100 lighting commands per second this way).

    If an E/M design office is concerned about traffic originating from simple units, this might be useful.
     
    Don, Jul 10, 2009
    #3
Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.