Subject:
|
Re: TLG & MIT Media Labs spawn another product: PicoCricket
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 31 Aug 2006 15:42:52 GMT
|
Reply-To:
|
danny@orionrobots.co+NoMoreSpam+.uk
|
Viewed:
|
3098 times
|
| |
|
|
On 31/08/06, Dick Swan <dickswan@sbcglobal.net> wrote:
> On Thursday, August 31, 2006 7:03 AM danny staple wrote:
> "Question - does anyone know how many output or input modules could be
> attached, and if there are any plans for seperate motor control
> boards? I strongly agree with Brian that having the motor controller
> separate from the mainboard could be quite handy."
>
> The PICO cricket had four ports. The ports are interchangeable -- you
> can connect either inputs or outputs to the port. And any device can be
> plugged into any port and the PICO cricket will find the device!
So bar power requirements, what is the addressing range of that? It
would be cool to be able to daisy chain elements, and jsut add
additional power where needed.
> Each peripheral device is intelligent and equipped with its own small
> microprocessor. There's a standard bus between the PICO cricket and the
> devices. This common bus is used for both sensors and motors. There's a
> good description of the Cricket bus somewhere on the MIT web site and
> presumably this is what PICO has used.
Each device having its own micro explains the expense, I take it that
only the main one can be reprogrammed and the otherones are merely
"slaves" with fixed code.
> From memory, I recall the bus as proving power, ground and a half-duplex
> serial link. The packet protocol on the bus included an address field,
> so it should be possible to daisy chain multiple items on the bus
> although the PICO cricket does not have the mechanical capabilities to
> do this. The bus was fast -- 100s of Kbits per second. THE MIT paper on
> the protocol was particularly "proud" of the design innovation allowing
> a bit-banged implementation on small micros that did not have hardware
> UART -- I recall they stretched the 'start' bit.
Oh - sorry asked that just up. I would still like to know how large
the address space is. Of course that is academic if it couldnt be
programmed. Do you think the default programming system could be
overridden with a fancier system, like NQC or similar. If it is based
on a standard micro, maybe there is a C compiler out there somewhere
for it. There is something about this being as modular as it is that
screams hackability to me.
> The motors/sensors have self-identification capabilities. For example,
> if you program a motor, it works whenever the motor is plugged into any
> of the four ports. The self-identification is the "type of device" and
> not an address.
Nice. But that would mean you may not be able to have two or more
motors connected as they cannot be uniquely identified. Unless this is
cleverer even than that, and each motor has a unique ID which remains
the same when their port is changed. Something not unlike udev and MAC
addresses on NIC's.
> For example, the base kit comes with two light outputs. If you do
> nothing when you send a light control command then it is sent out over
> all four ports. The default action is that a command applies to all
> ports. There is a feature that allows you to specify that an
> input/output command should use the single port specified.
Okay. Again, answering my above question.
> The ports use a funny little connector. It looks like the same physical
> shape as the tiny USB connector used on my digital camera. Logically and
> electrically it is not US.[B]
I would not expect so, but that tiny port may in itself define the
line power abilities, as I cant imagine such small connectors can deal
with much of a current.
> PICO cricket devices have features you can't find on the NXT. It has a
> really neat user controlled light output; it looks like a tri-color LED
> with variable intensity for each color. It has better sound playback --
> you can select from 24 different instruments: piano, trumpet, bells,
> vibe, rooster, horse , etc). Rooster/horse? -- remember this is a kid's
> toy! You can also control the tempo (playback speed. 100 steps) and
> volume (100 steps). You can also define your own tune/melody for
> playback.
I picked that up. I thought the NXT sound playback could be primed
with a sampled sound from a PC? Does it share any file
formats/protocols with the NXT?
> Two PICO crickets can message to each other by "beaming" infrared
> packets between them. This is not the 'crippled' low speed IR found in
> the RCX. I think it is much higher bandwidth.
Which aslo implies that it may not be IR compatible with it, unless
the IR can be clocked down to deal with it. I take it we are talking
about the clock frequency, and not the carrier, or is the carrier
different too?
> Yes. I too think it would be neat if the PICO cricket devices could be
> connected to the NXT. There seems to be two approaches:
>
> 1. Develop a new NXT "sensor" to beam IR packets to a PICO cricket.
>
> 2. Develop a new NXT "sensor" that contains one (or more) of the PICO
> cricket connectors. There's probably a powering issue -- especially
> with motors.
Also, couldn't yet another method be created using one of the NXT
motor ports, since there is already two way communication there due to
the motor feedback. Could motor port lines on the NXT be bit banged,
and maybe a direct interface to the cricket bus be made?
> All power to ports is supplied from the main PICO cricket. It takes
> three batteries (AA I think, but maybe AAA -- I can't remember). So I
> think the motors are maybe a little underpowered for use in a Robot. It
> only supports legacy LEGO motors.
Thats not such a bad thing, I have plenty of those laying around.
Being underpowered could be awkward, and it would be nice if the motor
boards could be given some separate power.
Danny
--
Danny Staple MBCS
OrionRobots
http://orionrobots.co.uk/blogs/dannystaple
(Full contact details available through website)
|
|
Message is in Reply To:
9 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
