Discussions

Baud Rates - Why don't we use 115200 all the time?

Posted by Nigel on May 22, 2011 at 5:07am

A general question about baud rates...

 

I have noticed that there doesn't seem to be any fixed standard for the 'best' baud rate for communicating between any particular device and the APM board.  Why don't we just stick to 115200?  What is the point in communicating more slowly?  For example, why has 57600 been selected as the speed for communicating with the XBee boards?  I see that the new ArduStation 'Reborn' project is using 9600...why use an even slower speed?  If we always use 115200 wouldn't that save a lot of confusion?

 

Don't get me wrong, there may well be very good reasons for the different baud rates...I'd just be interested to know what they are.

You need to be a member of diydrones to add comments!

Join diydrones

Email me when people reply –
Follow

Replies

  • Paul Mather May 26, 2011 at 8:17pm

    It is my personal opinion that the data being sent back and forth from the APM to the GCS is well suited for 38400. It's a good mix of 15-20% usage of the bandwidth and the reliability of the slower speed. 9600 is way too slow. 115200 is unnecessarily fast for a single UAV's purposes and I believe the 2.4GHz X-Bee's have a bug in them that will not allow them to work at 115200. The reason to select a slower speed is the slower the data is sent, the less likely the data is to be corrupted.

     

    According to RS232 specifications, the max length of a physical wire connecting two devices running 19200 is 50 ft. At 9600 this number jumps up to 500 ft. http://www.lammertbies.nl/comm/info/RS-232_specs.html The theory behind these distances were for noisy industrial environments, not over air. But the same applies. The faster you send the data, the more likely you are to have errors.

     

    The only real benefit I can see is the potential ability of having multiple UAV's all talking on the same X-Bee channel. In which case, you're going to need the faster bandwidth.... but again, for the single UAV, 38,400 is plenty fast and very reliable.

     

  • MikeInFrance May 23, 2011 at 8:21am
    History has it's part to play also:
    depending on the device you hook on serial, 9600 is about the entry point nowadays (think of serial GPS )
    When speed and datavolume are kept slow and small, all devices can communicate without having to use complex error correction routines.
  • 3D Robotics
    Chris Anderson May 22, 2011 at 7:30am
    It came about from a lot of testing. We got too many lost packets on the standard 900Mhz Xbees at 115200, and the telemetry team said they didn't need any faster speed than 57600, which is more reliable. We used to run them even slower at 38k, but after a lot of field trialing felt comfortable enough to raise that.
  • Developer
    Jani Hirvinen May 22, 2011 at 7:22am
    Jmd45, yes we know those but they are a bit heavyish.
  • Developer
    Jani Hirvinen May 22, 2011 at 6:46am
    Answer is simple 57600 used to work a lot better on earlier xbees. Also it really depeneds on what type of protocol you using to writr your telemetry things.

    More faster you go more you will get errors. If you dont have good error correction on protocol your gcs will work badly.

    Also distance is one issue that affects on speed. Sure we can go even 20 megabytes over 20 kilometers but then quatiom is wha tis the price for the modems. There are njce tdma modems than do life hd feed slong other thing and for long distances but problem is that they cost aboit 800usd each.

    9600 might be a bit too low for proper data feed. 57k then again lets us to put a lot of messages trough. 115k is good if we dont have interference.
This reply was deleted.

Activity