h a l f b a k e r yThe word "How?" springs to mind at this point.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
As you electronic geeks know, the USB interface is not really a bus. A bus allows a nearly infinate number of devices to be connected one the SAME data lines.
USB requires expensive hubs to connect more than 2-5 devices.
I propose an interface where all devices are daisy-chained along the same
cable. Nearly every device will have at least one extra port on it to allow other devices to link off of it. (Something we were promised with USB)
In addition, add a 12volt line as well. That would make things easier for some devices like hard drives.
Of course, linking many devices on the same bus will quickly draw too much power, so power extender hubs will be needed.
Unlike USB hubs, these merely use an AC adapter to power the devices and contain no expensive circuitry, and would be very cheap to manufacture. (or build yourself)
Externally powered devices like printers will have a power extender built in.
[link]
|
| |
While USB does require the use of Hubs, which some may find annoying, they're not particularly expensive and the hub-based design offers many advantages compared with a "true bus" architecture--especially when dealing with mixed-speed devices. |
|
| |
In the absense of switching devices (hubs), the speed of a bus will generally be limitted to the speed of the slowest device on it. If your mouse had to share the same bus segment as a USB hard drive, either the mouse cable would have to be of high enough quality to handle a 480mpbs signal without radiating excessive noise or capacitively loading it, or else the hard drive would be limitted to communication at 1.5mbps. Neither scenario seems all that great to me. |
|
| |
Further, from my understanding, USB2.0 hubs are able to combne multiple low-bandwidth streams into a high-bandwidth stream, so it would be possible to have multiple 12mpbs devices plugged into a single USB 2.0 port and have a combined bandwidth that significantly exceeds 12mpbs. Even if a 'true bus' architecture could deal transparently with multiple speed devices, a 12mpbs device which used 50% of the available bandwidth would deprive a 480mbps device on the bus of 50% of its bandwidth. |
|
| |
USB works. While certain aspects are annoying, I consider the point-to-point nature of the architecture to be a plus rather than a minus. |
|
| |
The Commodore64 used to have a daisy-chaining peripheral linking system, using a 16-or maybe it was an 8 DIN wire, with each peripheral taking some responsibility over data transmission etc. I never tried linking more than 3 things together, and everything still had a separate power supply, but the system appeared to be limitless (though probably limited to 8 or 255 depending on the allocated peripheral index space) |
|
| |
Oh, I was hoping it would actually be a public transport vehicle. |
|
| |
//The Commodore64 used to have a daisy-chaining peripheral linking system, using a 16-or maybe it was an 8 DIN wire, with each peripheral taking some responsibility over data transmission etc.// |
|
| |
The protocol was loosely based on IEE-488 (which was a good protocol) but was unfortunately dog slow. It is maddening how much better the protocol could have been designed with no extra hardware, and how much better yet it could have been with a very small amount of extra hardware (even by 1980's standards). |
|
| |
There is a data protocol used in the entertainment industry called DMX512 that supports daisy-chaining up to 32 devices, and infinite cable-splitting (within practical limits of voltage). |
|
| |
It's also totally unsuited for anything PC related, having very low actual data throughput, and being totally unused anywhere else, but it's an example of a similar concept. |
|
| |