[Gmecm] Clearing fault codes

[Robin Handley]

Following the Turbo P4 doc, I've had another go with the Mode 4 command;
setting the discrete output enable control and state words to $00 and the
mode control enable word to $40, but the MALFs didn't go. There's something
odd about Mode 4 with the Sunbird, though, as when I tried using it to turn
the fan on, I succeeded but a very odd message came back (2nd byte = $52 -
i.e. less than the normal minimum of $55!) followed by ~ 20 data bytes. This
unlocks my test program. Anybody any idea why this 'length' byte is so odd?

I was also being really dopey with the manual approach to MALF clearing. I
(thought I) was disconnecting the main ECM power and wondering why the MALFs
weren't clearing (and wondering if my EPROM changes were wrong), and I
wasn't! Now I've rediscovered how I wired the ECM in to the car (which
provides a really easy method of disconnecting the constant +12v), I'm
successfully clearing the MALFs, so the s/w method isn't so important. The
power disconnect cleared my MALF 33 and it hasn't come back, so I think
raising the value of KKPM33 to 100kPa was the right thing to do - since I'm
running a 1BAR MAP sensor with the code in the 1BAR mode.

I also discovered another interesting 'feature' at the w/e. With ANHT/$8D
s/w (in my 1227727 h/w) my fan was driven by the fan #1 (low fan) code and
thresholds. With Sunbird/$58 s/w, my fan is apparently driven by the high
fan code and thresholds! This is the first apparent wiring difference that
I've encountered. TPS, MAP, HEI, ALDL, and CTS are all ok. All that leaves
for my (later) injection installation is MAT and the injector group fire
wires, I think. I'm hoping these will be as for the 1227727 pin-out as well.
:-)

Thanks,

Robin

----- Original Message -----
From: <davesnothereman at netscape.net>
To: <gmecm at diy-efi.org>
Sent: 10 June 2006 13:56
Subject: Re: [Gmecm] Clearing fault codes


> I haven't had any luck using aftermarket laptop software to clear 7749
> MALF's.  I can't remember if the dealer scantool was effective for
> this.  I generally pull the plug if I need to clear the codes right now.
>
> There is code in the $58 which causes the ecm to automatically clear
> MALF's from memory after a specific number of key cycles.  I set that
> to 5, figuring that a current code will continue to be current, and a
> passing code due to my error with an eprom or running too much boost
> shouldn't hang around at all.
>
>
> Zaphod
>
> -----Original Message-----
> From: Robin Handley <Robin at FuryWorld.fsnet.co.uk>
> To: gmecm at diy-efi.org
> Sent: Tue, 6 Jun 2006 09:47:50 +0100
> Subject: [Gmecm] Clearing fault codes
>
>    I've been trying to get rid of some spurious fault codes and I've
> realised
> that I'm not exactly sure how they latch.
>
> If the fault (e.g. 33 - MAP too HIGH) goes away, after it has been
> present,
> should the flag automatically clear? Or is a total ECM reset required?
> I've
> had no luck with using ALDL commands to clear malfs. Is it required to
> remove all power from the ECM? If so, for how long?
>
> Robin