RE: A86: TI-Network
[Prev][Next][Index][Thread]
RE: A86: TI-Network
In that case, no offense, but I'll continue with my own specification for the network, as I'd rather see a freeware release. I figure that since these are just calculators (powerful ones, but calculators nonetheless), no one should be kept from taking advantage of these designs.
I will elaborate on the general protocol and other aspects right now, however.
Most work will be performed off-board, from one of the Atmel AVR microcontrollers. The system will handle up to 255 (1 reserved for "send to all") devices connected, from calculators to expanders and possibly other components. most components will require a special "locked burst transfer" mode, in which a device controls the entire bus for direct communication, unless compliant drivers are written to access these devices. For example, an expander being run as a standard device will be located on the bus, and the communications will be effectively locked for other calculators until the expander<->calc transfers are complete.
Error handling will be provided via a 4-bit checksum (sum of bytes, separated into "nibbles", readded to crate the 4-bit value), and each packet will consist of the following portions:
[startpacket#][destination address][packet x of y][number of data bytes][data(up to 4 bytes)][checksum][source address][endpacket#]
startpacket# is a predefined value that defines the beginning of a packet, address is the assigned # of the destination device, and I think most of the others are self explanatory.
The packets are transmitted cyclically, with "good" packets being removed from subsequent cycles. The retransmit packet is sent as follows
[startreTx#][source address][# of bad packets][bad packet #'s][checksum][destination address][endreTx#]
Where source and destination are the same for each packet type
Excess data is stripped off, and the final, cleaned result is transmitted to the calc via the TI protocol (standard TI checksum algorithm, etc.), which is ultimately processed by an assembly driver to either store the data in mem and edit the VAT appropriately, or use it as a returned value in a program.
The first demonstration of this technology will be exposed in a chat program, and later, to transmit settings across the whole network for demonstration purposes in class. For example, the teacher can set up a graph on one calc, which will subsequently be sent to every student, demonstrating the exact data sets needed. This would be very useful in an AP statistics course, which for the most part becomes a calculator demonstration class.
Christopher Kalos
raptorone@stuytech.com
Executive Director/Administrator
Virtual Technologies Developer's Group
----------
From: Grant Stockly[SMTP:gussie@alaska.net]
Sent: Thursday, December 04, 1997 7:06 PM
To: assembly-86@lists.ticalc.org
Subject: RE: A86: TI-Network
It is being worked on by Grant Stockly(me) and Travis Brady. In order to
keep the project confidential, devolpers won't be able to program for it
untill we have addiqute patents and copyrights. It is 10Base-T and kinda
works.
If anyone has a www site on the internal workings of an ethernet system I
would appreataite getting the info!
Thanks!
Grant Stockly, ApleCyber
Travis Brady, BradySoft
- Warning
- Could not process part with given Content-Type:
application/ms-tnef
Follow-Ups: