Post by thread:(Fwd) Re: 12C508 too fast partial solution?

Again, this message was inadvertently sent only to me; I'm forwarding it to the PICLIST. -Andy ------- Forwarded Message Follows ------- Date: Fri, 31 Jan 1997 22:39:45 -0600 (CST) To: spam_OUTfastfwdTakeThisOuTix.netcom.com From: Todd Peterson <.....elabKILLspam@spam@netins.net> Subject: Re: 12C508 too fast partial solution? At 01:01 PM 1/31/97 -0800, Andrew Warren wrote: >Todd: > >Yes, there's someone (I forget his name now) from CCS on the list. > >That is, he USED to be here... I haven't seen any messages from him >since someone posted a long list of CCS-compiler bugs to the list >three weeks ago. I have found two distinct problems here; the first was a CCS compiler error; however they did graciously fix the compiler bug problem immediatly. The second is that, no matter what value is written to OSCCAL (the upper 4 bits are the only ones that matter), I cannot change the speed of the clock. Yes, my scope is perfectly calibrated; it has a 1KHz calibration plug that shows exactly calibrated; and besides I should see a change anyway when moving the PIC's calibration constant from 10 to 70. My question is: do others have this working? If you think you do, you might not; the device does power up to a default of 70 and you might not be really calibrating the RC value from that. I do read the hex file back from the programed chip every time to verify that the last location (the cal. constant) was programmed correctly. I get a .8 mS high time from the output pin regardless of the cal. constant loaded into OSCCAL. I hate to mess with the delay loops I know work, 'cause unsolved problems like this invariably come back to bite you. I would actually be happy if I could verify that someone IS actualy tuning their R/C properly, and not just using the powerup value of 70. Be careful, the last location of memory reading back as your constant is not proof it is working; you must actually watch signals on I/O pins. -Todd

Hi Andy, &c. > Again, this message was inadvertently sent only to me; I'm forwarding > it to the PICLIST. Your mailer sends out a reply-to header and some list servers leave it there if it exists and otherwise put in the list address as reply-to. If you send mail from only one account then you could remove your reply-to header if this symptom starts to get to you. {Quote hidden}> The second is that, no matter what value is written to OSCCAL (the > upper 4 bits are the only ones that matter), I cannot change the speed > of the clock. Yes, my scope is perfectly calibrated; it has a 1KHz > calibration plug that shows exactly calibrated; and besides I should > see a change anyway when moving the PIC's calibration constant from 10 > to 70. My question is: do others have this working? If you think you > do, you might not; the device does power up to a default of 70 and you > might not be really calibrating the RC value from that. I do read the > hex file back from the programed chip every time to verify that the > last location (the cal. constant) was programmed correctly. I get a > .8 mS high time from the output pin regardless of the cal. constant > loaded into OSCCAL. > > I hate to mess with the delay loops I know work, 'cause unsolved > problems like this invariably come back to bite you. I would actually > be happy if I could verify that someone IS actualy tuning their R/C > properly, and not just using the powerup value of 70. Be careful, the > last location of memory reading back as your constant is not proof it > is working; you must actually watch signals on I/O pins. I have written a short program to do what you need. Works like a bomb and as it only uses 9 bytes you can use it in a OTP and then just program the bytes to NOPS after you are done if you must use it later. I realise when I was testing that the B on the lid of one of my devices was infact a 3 and this could explain some of my own timing problems I had while testing code. The graph illustrates the way the frequency changes. It shows percent from the ideal values. Check out the following URL : http://www.ip.co.za/people/kalle/pic/default.htm#caltest Cheers -- Kalle Pihlajasaari kalleKILLspamip.co.za http://www.ip.co.za/ip Interface Products P O Box 15775, DOORNFONTEIN, 2028, South Africa + 27 (11) 402-7750 Fax: 402-7751 http://www.ip.co.za/people/kalle DonTronics, Silicon Studio and Wirz Electronics uP Product Dealer