Menu
▾
▴
Re: [Gumstix-users] CANbus connectivity?
|
From: Sean M. <sm...@ap...> - 2006-02-06 17:14:38
|
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> <small><font face="Helvetica, Arial, sans-serif">Hello all,<br> <br> Thanks to everyone for your ideas, they basically confirm what I was thinking. Here are the options I see for CAN connectivity:<br> </font></small><br> <small><font face="Helvetica, Arial, sans-serif">Option 1:</font></small><small><font face="Helvetica, Arial, sans-serif"> Hang a full featured CAN controller off the 90-pin bus controller. I was thinking the Phillips SJA1000.</font></small><small><font face="Helvetica, Arial, sans-serif"><br> <br> The Good: W</font></small><small><font face="Helvetica, Arial, sans-serif">ell known part with lots of support and GPL drivers available.</font></small><br> <small><font face="Helvetica, Arial, sans-serif"><br> The Bad: </font></small><small><font face="Helvetica, Arial, sans-serif">You limit your other connectivity options. It would be *really nice* to have CAN and ethernet. (Why isn't the gumstix bus-header stackable?)<br> <br> <br> Option 2: Hang a micro with built in CAN on the NSSP (SPI) bus and implement a translator in software. There are many micros with built in CAN, pick your favorite vendor.<br> <br> The Good: Doesn't limit ethernet connectivity options.<br> <br> The Bad: Software translation, lower throughput, may not be fast enough for hard realtime applications.<br> <br> <br> NOTE: I am still new to the gumstix platform so forgive me if I made any bad assumptions above.<br> <br> <br> Sean<br> </font></small><br> </body> </html> |