Starting Problems (electric and rather interesting)

[rdh@sli.com (Robert D. Houk)]

Some further thoughts (staring at the ceiling, trying to get to sleep
last nigth) on no voltage to the fuel pump.

Another possibility is that the computer is pissed off and refusing to
switch the fuel pump on. Again, this discussion/hypothesis is based
on my '83 UrQ, and may or may not have anything to do with any other
species of the genus Audi.

There are three separate sensors the computer uses to determine the
primary ignition timing (which is purely electronic, no "mechanicals"
are involved, unlike the fuel system which is basically mechanical
injection with computer-controlled fine-tuning...). There is a sensor
which counts the teeth on the ring gear (flywheel), which is used as
the determinate of timing. There is a second sensor which detects the
TDC mark (pin) on the flywheel, which in turn sychronizes the primary
sensor (counting the ring gear teeth). There is a third sensor in the
distributor which is used to determine which TDC impulse is 0-degrees
vs 360-degrees (of the 720-degree four-stroke cycle), and is thus the
master synchronizing pulse. If any one of these three sensors fail (or
the cable harness/connectors/etc. disconnect), then the computer can
not determine engine timing, and will shut down the fuel pump. I *be
lieve* that the fuel pump should always cycle on for a second or two
when you first switch from OFF to ON/START (and thus is another "self-
test" you can perform) before the computer decides to shut it off a-
gain, but can't tell you why I think this is so (and even if true for
my car, might not be for yours). If any sensor fails, voila! cranking
to heart's content but no fuel pump (and no ignition).

In other words, if anything anywhere fails, everything everywhere is
doomed!

Now, imagine a room full of clever German engineers, poring over the
latest Bosch catalog, oohing and aahing about the latest sensor arrays
and challenging each other to see who can incorporate the most new
sensors in the cleverest ways into the latest engine management sys-
tem . . . then take a look at your engine compartment, and the wiring
harness, and all those dangling connectors, and use your imagination!

					-RDH

P.S.	'nuther random thought -- on my car, there is a two-conductor
	"test socket" (rectangular, female connectors) dangling near
	the coil. This is in parallel with the Freq Valve signal, and
	should yield a duty cycle of 50%(*) when working properly. When
	starting, it can go as high as 95%. So if you see any signal
	present there, you know the engine computer is "working", more
	or less properly. If you see no signal there, either the com-
	puter has no voltage, or has "shut down" to due faulty sensors,
	excessive cosmic rays, etc. This is an easy-to-access test,
	as opposed to pulling the computer to get at the master wiring
	harness to start checking voltages/resistances/etc. presented
	to the computer connector.

P.P.S.	Check the simple and obvious first. For all their inability
	to use wire and connectors and relays and fuses in a safe and
	sane manner, I have never heard of an engine computer failing
	(I'm sure they do, but with nowhere near the frequency of just
	about everything else in the car), so you can probably assume
	the basic computer is behaving the way it is supposed to be-
	have ("is not broken"), and thus the problem's a "simple" one
	of bad wire/connector/sensor/etc. You just hafta find the sim-
	ple and obvious problem . . .

(*)	OK, before I get flamed -- 50% is "ideal" reading; actual will
	vary depending on what the computer thinks about the exhaust
	composition, in static conditions it will typically vary in
	a 5% window; full-throttle should "lock" at 70-80% range. On
	my car, at idle, if you actuate the WOT switch the duty cycle
	should lock at 50% solid (this is a quickie self-test of the
	idle and WOT switches (and computer) working properly).