[Gmecm] Ignition off fan

[Robin Handley]

Been bashing my head against the $58 Fan code...

The 749 circuit diagram clearly shows just one Fan line, which turns on the
fan at full speed (i.e. not via an in-line resistor) when the ignition is
on. If the Fan line is active and the ignition is off, then the Fan is
driven at low speed, via a resistor.

According to my understanding of the fan code, and observation via simulated
CTS inputs, while bit 4 of FMDBYTE1 (V5 cooling fan request) is OFF (which
is how it stays in my installation), only the Hot Fan ON/OFF thresholds are
used to activate and deactivate the Fan line.

I see no distinction between the 1st time hot flag and the Fan ON flag.

If the V5 cooling Fan request is ON, then the (non-Hot) Fan ON/OFF
thresholds can be used to activate and deactivate the Fan line. The V5
cooling Fan request never gets turned ON in my installation and I don't know
what could turn it ON.

I haven't fully got my head around the Fan code because there are some
status bytes that come into play, mostly when the V5 cooling Fan request is
ON, which I don't know the meanings of ($0029, $002C, $0032, $0033). Also
$0102 and $0178 are used - which I think are Fan timers.

I can only assume that my Fan flicking ON and OFF a few times, as (real -
i.e. not simulated) CTS approaches the Hot Fan ON threshold is due to
sufficient noise to defeat the ON/OFF hysteresis, and that the 1Hz ALDL log
isn't showing all of the dips. It should take only one value below Hot Fan
OFF to turn the fan off. I will see if I can temporarily improve the CTS
ground to see if this goes away...

Robin

----- Original Message -----
From: "Scott Peitzsch" <jlg-sep at comcast.net>
To: <gmecm at diy-efi.org>
Sent: 28 June 2006 22:17
Subject: Re: [Gmecm] Ignition off fan


> > So, since my fan goes straight off when I turn the ignition off, would I
> > be
> > right in deducing that the $58 code goes through all the motions of
> > controlling 2 fans, but the code for the 2nd (high) one achieves nothing
> > on
> > the 749 h/w, but _does_ work on my 727 h/w? :-)
>
> That is how I understand it.  You also need to be sure that the high side
> power source for the fan relay is on battery power, not ignition switched
> power in order for it to be able to stay active after ignition-off.
>
> > If this is right, then it's probably worth me trying to find the #1 fan
> > pin,
> > to get the ignition off running. Shame there seems to be no cross
> > reference
> > for this! I guess I just check the quad driver pins...
>
> This is the cross reference I use:
>
> http://www.cruzers.com/~ludis/triplugecm.html
>
> You can also see the '727 pinouts for fan control here:
>
> ftp://ftp.diy-efi.org/pub/gmecm/manuals/1227730/90-7727V8TPI-6.jpg
>
> > BTW: I'm also observing my fan flicking on then straight off a few times
> > as
> > the engine temperature rises to the high fan ON temperature that I've
set.
> > The CTS is a little noisy (mostly +/- 1 C, with occasional dips of 6C
!);
> > but, looking at the plots of CTS vs. fan ON, there isn't a 100%
> > correlation
> > between the CTS spikes and the fan state transition. I'm still trying to
> > understand the code (not helped by not knowing what half of the zero
page
> > locations, particularly statuses, mean), but I've been wondering whether
> > the
> > $58's high fan control logic isn't as sophisticated as that for the
> > control
> > of the low fan (i.e. high fan has no delayed start up - which might
reduce
> > the likelihood of the 'flicking'). If the $58 only ran on h/w that
didn't
> > use the high fan logic, then GM may never have noticed this behaviour,
or
> > didn't care(?)
> >
> > Having said this, I'd expect the natural hysteresis of the ON/OFF
> > thresholds
> > to avoid this kind of 'flicking' problem...
> >
> > I also plan to look at my CTS wiring. I 'cheated' by wiring the GND side
> > to
> > chassis. This massively added to the CTS noise when I had had bad engine
> > grounding (now fixed), but still may be adding some. I'd be interested
to
> > see others's CTS plots to compare noise.
>
> This is a very bad practice.  Sensor grounds should always observe proper
> "star" or "single point" grounding technique.  The correct sensor ground
pin
> to
> use on a '727 is C10, as shown here:
>
> ftp://ftp.diy-efi.org/pub/gmecm/manuals/1227730/90-7727V8TPI-2.jpg
>
> -Scott
>
> _______________________________________________
> Gmecm mailing list
> Gmecm at diy-efi.org
> http://lists.diy-efi.org/mailman/listinfo/gmecm
>