Re: Ur-Quattro

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

   Help, I have an 83 Ur-Quattro with the IA computer mods. The car has a
   stumble at around 4500 rpms with the boost at 13 to 14 lbs. Now with the
   boost at 12, the car does not have the stumble unless the air conditioner is
   on (but its not as fast as I want). I want to crank the boost back up, but I
   need to take car of the stumble. 
   Its seems as though the computer is reading an over boost and is shutting
   down the fuel pump. I've checked all the sensors and replaced the air temp
   sensor which helped alittle bit. Also, when I disconnect the manifold vaccum
   line from the back of the computer the car runs great. 
   Could this be do to the modified exhaust system. I welcome any ideas.

Well, I can guarantee that you cannot confuse "stumble" with the computer
shutting down the fuel pump (which is akin to hitting a brick wall; the
engine is switched off, the drivetrain completely unwinds and winds up
in the other direction, driver, passengers, pets, other loose objects all
head for the windshield). It is not a genteel (not to mention gentle) pro-
cess at all! Especially when it happens after the suspension has taken its
set for a banzai sweeping onramp run . . .

If it hits at 4000 (aka 4500; those tachs are incredibly liberal in their
readings) rpm, that sounds like a bad intake air-temp sensor (bad idle
switch will also cause this).

If you can run 14psi at 3000 (or 3500), then 14 psi should work everywhere
you can get 14psi (i.e., over 2500 or so), and not trip the overboost fuel
pump shutdown.

If you can get the engine over 4500 (like 5-6000 rpm or so), then the com-
puter is not complaining about the air temp sensor [under boost conditions
only] -- can you hold 12psi to redline, for example? 14psi? 10psi?

A "stumble" sounds more like it's just not getting the gas that it wants
(or getting too much...). Air might be blowing out a weak injector seal (or
any of a hundred other seal points). Your fuel pump might be "weak" and 
"stumbling" at higher flow rates. You might even have a piece of "dirt"
blocking the fuel distributor orifices (slots) at just that point . . .

Does the stumble even out at higher rpms? Get worse? Power steadily
drop off?

Get a duty-cycle meter and see what the computer "thinks" the mixture is
like in those conditions. Ideally, will read 50% always (50% says the
mechanical injection system is spot on, needing no corrections from the
computer; lower duty cycle is leaning out the mixture, higher duty cycle
is enrichening the mixture; under full throttle, the computer kicks the
duty cycle up to provide a "nicer" (richer, cooler-running) mixture; on
mine, it kicks up to 78.9% (my meter's reading) for several seconds, then
drops to a steady 70% for as long as I have had road/courage to keep the
throttle wide open). Handy dandy self-test: if both idle and WOT switches
are closed, the computer locks the duty cycle at 50% (my meter: 49.9%),
so you can quickly and easily check switch operation -- at idle, just lift
the WOT switch lever and you should see 50%; open throttle a little, and
you should then see 70-80%; let go of WOT switch, you should see 40-60%
varying readings -- this will tell you if both switches open and close
and the computer sees them.

pYou can also try measuring the voltage from the O2 sensor to see what the
*real* mixture is. +0.5V is ideal, lower voltage is rich, higher voltage
is lean (I think; hafta check the books to be sure on that one); the 
transition is extremely steep, almost a step function, between lean and
rich, so what you should really see is a rapidly vacillating .2V to .8V
(or so) and back again signal as the computer in essence "dithers" the
mixture to maintain an approximately-correct mixture. (Several companies
make a "LED bar-graph meter" you can hook up to the O2 sensor to monitor
the mixture; I don't have any experience yet with these gizmos, but they
sure seem neat.)

					-RDH