Subject:
|
Re: USB Host and Slave
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 12 Jan 2006 11:34:41 GMT
|
Original-From:
|
PeterBalch <PeterBalch@compuserve.&ihatespam&com>
|
Viewed:
|
1509 times
|
| |
|
|
> > There's been a lot of discussion of how hard it is to make a USB host
> > compared to a Slave.
> > What if it were a "special purpose" host.
>
> In lugnet.robotics, dan miller <danbmil99@yahoo.com> wrote:
> look up "USB on the go", it's basically what you describe -- a minimal • host
> for peripheral-to-periperal connections.
I don't think that "USB on the go" would do the job. That's for connecting
two slaves and it's very limiting as to what the two slaves can do to each
other - it looks to me as though only bulk transfer is available.
There are lots of very cheap USB peripherals available now. Let's say I
wanted to use a USB optical mouse for odometry, I don't see how USB OTG
would help.
If the main processor on my robot was a PIC or if I wanted to use a PIC to
make an interface between a mouse and an RCX or NXT then I could probably
use a very cut-down "special purpose" host that only knows how to talk to a
mouse. Maybe.
Similarly, could I build a version of the CMU CAM that uses a USB webcam?
Webcams are very cheap and not all vision applications require huge amounts
of processing: line following, "find the orange ball", barcode reader, etc.
If I want to use a cheap processor to control an RS232 device, it takes me
about 10 minutes to write the program and design the hardware. I2C takes an
afternoon to write from scratch. UDP took me about a week to get working on
a PIC. TPC/IP would take twice that. But if I want to control a USB device
then will it take me 10 days or 10 weeks? Is this progress?
Peter
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
