![]() |
|
AID: RE: peripherical card
From: cis (by way of Michelle Kasprzak) cis
artsens.org
Date: Tue, 12 Nov 2002 06:25:41 -0800
: Sent via the Art Interface Device mailing list: aid@interaccess.org : Use your "Reply All" to reply to the list, "Reply" for private response > There should be a note somewhere, though, that one should probably NOT > place the address on the address bus AND assert /R in the same instruction. > This would be possible, since they share the same port, but could result in > spurious reads from peripheral boards. Yes, it must be sequential, and you must keep in mind that in some PIC (because of the internal technic of bcf, and bsf instructions and pipelines specifications) it is not recommended to access 2 times consecutively to the same external port it is better to put a nop (or any instruction) between puting adress and read if you code in C in differents fonctions for place adress and put read if they are non inline fonctions, the call to the function and return from function will do the wait for you > This works as a hold, to ensure that the data on the bus is that which > was in the 373 when the /R signal was asserted (or, about 31ns thereafter). > I would leave the creative use of LE up to individual designers, myself. > I can see applications where you'd want the latches transparent during the > read or write, and others where you might not want that. if you prefer, but if you want another data, you just need to read again by puting read up and read down, and you will have the next most recent data I don t thing you want so big data traffic that you need to not loose time on puting up/down a pin moreover, this could be a synchronisation between card but I have always the same question : how the main card knows the sort of card it is reading and it must do with the data cis@artsens.org : messages saved at http://www.interaccess.org/aid/list : unsubscribe/help requests to mailto:Majordomo@interaccess.org