Exact match. Not showing close matches.
'[72677.2623@CompuServe.COM: Re: Errors from MPALC]'
|At his request, I am forwarding a message from Don Lekei
<CompuServe.COM>. For the record, I completely agree with him, 72677.2623
and I use the TRIS instruction routinely. In my previous message which Don
quotes, I just presented all three ways of accessing the TRISA and TRISB
I should point out for the benefit of those using the 16C71 that the TRISA
register on the current parts doesn't actually have a bit for PA4. If you
write anything to PORTA with bit four low, it will drive PA4 low regardless of
what you have written to the TRISA register. The errata covers this, but in a
confusing manner. Mike Jedrezewski (sp?) of Microchip told me that there is a
new rev of the 16C71 silicon due out soon which is supposed to fix all known
errata, including the missing TRISA bit. I've also heard unofficially that
a 16C71A should be out in early '95, with improved performance (possibly
>20 MHz) and lower power consumption.
------------------------ Don Lekei's message follows ------------------------
>Date: 18 Jun 94 13:23:40 EDT
>From: Don Lekei <CompuServe.COM> 72677.2623
>To: "INTERNET:apache.telebit.com" < ericapache.Sunnyvale.Telebit.COM> eric
>Subject: Re: Errors from MPALC
In the previous message, Eric Smith refers to Microchip's documentation:
>Microchip is trying to discourage people from using the tris instructions,
>and claim that they may be removed from future parts.
This advice from Microchip is some of the highest ranked BAD ADVICE in the
history of PIC programming. It is right up there with starting your code at
location zero in any PIC (other than an '84)!
You should ALWAYS use the TRIS instruction, unless it is necessary to
modify the TRIS register (eg. bi-directional i/o or open-drain emulation).
1) It is better to maintain code compatability and portability to the
existing economical PIC parts (5x) rather than some mythical future part.
2) TRIS takes less code and fewer cycles than accessing registers in bank
3) MPSIM will not correctly track changes directly adjusting bank 2
4) If Microchip ever actually does remove the TRIS instruction on a future
part, then it is far more portable to define a macro to emulate the TRIS
instruction on that part, than to cut-off backward compatability with 16c5x
I just HAD to jump in here, it makes my skin crawl whenever I read that
part of the data sheet.
PS. Another handy way to start a lively thread is to mention using EQU to
define data addresses.
More... (looser matching)
- Last day of these posts
- In 1994
, 1995 only
- New search...