Art Interface Device logo   Project description  Development   Mailing List  Contact


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: AID: Bug in AID Firmware

From: Mel Wilson   mwilsonthe-wire.com
Date: Wed, 17 Nov 2004 18:37:42 -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

In article <20041025040843.GB3459@zooid.org>,
Stefan Wiechula <stfn_aidp@zooid.org> wrote:
>On Sat, Oct 23, 2004 at 01:49:46PM -0400, mike2004 wrote:
>> Some observations that cannot be confirmed due to the lack of a
>> programmer are:
>>
>> 1) a newly programmed PIC seems to function correctly and only fails
>> after use. In fact it seems to fail only after use of the AD command.
>>
>> 2) a board I am using only exhibits the problem when I use AD commands
>> and does not get stuck in this state. On power up it behaves normally.
>
>Have you tried using code compiled with aid_lib.c revision 1.6?  Last we
>talked about this someone (you?) noticed that the ADC module was left
>enabled after a conversion and might be sinking or sourcing a lot of
>power.  The "fix" in revision 1.6 is line 728, here:
>
>http://cvs.sourceforge.net/viewcvs.py/artinterface/Core/Code-Firmware/AID_LIB.C?annotate=1.6

   As of Nov.14th, Ian, Jeff, Mike and I have concluded:

   In aid_lib.c version 1.1 there was code that set the
ADCON1 bits PCFG2:0 to binary 010 .. this decoded meant that
all the PORTA pins (that is, ADC0 to 4) were set for analog
input, but all the PORTE pins (that is, ADC5, 6 and 7) were
set for digital input.  That meant that an AID mainboard
running v1.1 of aid_lib.c with analog voltages connected to
ADC5, 6 or 7 could run the PORTE input gates in a halfway
state where both the high and low transistors of the gate
were conducting, and *that* meant the possibility of a short
between Vss and Vdd through the PORTE input gates, with the
chance of excessive current drawn and damage to the gates in
the microprocessor chip.

   The problem was eliminated in version 1.2 of aid_lib.c,
and won't have appeared since.

        Mel.

:      messages saved at http://www.interaccess.org/aid/list
:      unsubscribe/help requests to mailto:Majordomo@interaccess.org