88 5kq - Intermittant Hard Starting

[Huw Powell]

>>To make this adjustment without a CO meter, you hook up a milliammeter
>>in series with the DPR and adjust the static fuel mixture at idle until
>>the DPR current swings back and forth (with OXS voltage) around 0 mA.
>>(it will go back and forth a few mA - you need a fine scale on your
>>meter).  It helps to make a harness from male to female for the DPR
>>plug, but you can get by without.
> 
> **** OK, this is the same as watching the O2 sensor voltage start to swing
> around 0.5 volts >  0.2v  <-->  0.9v

No, it isn't.  The DPR could be swinging around, say, -10 or +10 mA and 
the OXS would look the same.

> The DPR current wont swing if the O2 dont swing

That is true, yes.

> I think the DPR current is 10 ma at Stoich(14:7). Lean stop is at 0 ma, and
> rich stop is at 20 ma. In normal operation, the current should fluctuate
> around the center value of 10 ma. No?

No, that's CIS-E.  CIS-E3 fluctuates around 0 mA, which makes for a 
nicer "limp home" mode if the computer dies.

>>And as you say, it might need to be richer to enable the ECU to get it
>>to swing at all.

Right.  That is your first goal, and you could head for that by watching 
the OXS voltage, if you want.  Then fine tune with the DPR current.

> **** I was looking for some easy way to set the basic fuel mixture without
> having to use a CO meter, or a swinging milliampmeter.

Can't really be done.

> I was thinking of checking the O2 voltage on my 100Q with the O2 sensor
> disconnected. Then adjusting the idle mixture on the 5000q to get the same
> reading on the O2 sensor as the 100Q reading. But, an O2 sensor is not very
> accurate above and below its operating range. So, forget that idea.

Exactly.

> On the other hand, how about setting the basic(mechanical) idle mixture this
> way.
> 1. Disconnect O2 sensor from its harness to CPU
> 2. Apply 0.5 volts to the CPU, this will fool the cpu into setting DPR
> current at 10 ma
> 3. Monitor the disconnected O2 sensor with DVM
> 4. Adjust idle mixture until O2 sensor reads 0.5 volts(Stoich)
> Cool, eh?

Same problem as you just mentioned, at least unless you use a wideband 
OXS.  Which would be more trouble than a CO meter or mA meter...

> Oh wait, the CPU aint dumb . . if it dont see any O2 crossovers, it goes
> into open loop mode. Hmmmm, we could fool the CPU by using a sine or square
> wave generator . .. no . . no good, cause the DPR current is fluctuating and
> hard to read.
> Drats! Foiled again!!
> 
> Whats the difference between open loop mode and limp home mode? What is the
> DPR current for each mode?

Open loop is for two things basically - the warm up curves, where the 
OXS can't be relied upon and a richer mixture is needed anyway, and when 
the WOT switch is activated, for a richer mixture under hard acceleration.

Limp home is when the computer is busted and everything just runs at the 
static settings.  Can be a pain when the engine is cold, but runs OK 
when warm (due to the 0 mA thing).

> Open loop is when CPU does not use O2 sensor input, but uses the internal
> 0.5 volt reference and it uses the temp sensor to regulate DPR current for
> temperature only? With engine fully warmed up, the Open loop DPR current
> will be around 10 ma?
> 
> Limp home mode is when the CPU electrical system is dead? The DPR current is
> 0 ma? An unplugged DPR would simulate Limp home mode, right?

Yes, pretty much, see above.  Except substitute 0 mA for 10 mA.

> If Open loop mode works as described above, then I can use this mode of
> operation to set the idle mixture. Adjust the mixture until the DVM on
> disconnected O2 sensor reads 0.5 volts. Now, the Open loop 10 ma DPR current
> will coincide with a Stoich O2 reading of 0.5 volts. Will this work?

Why not just do it right?  You'll never be able to get an OXS to read 
right on 0.5 volts, it's going to jump up or down.

> I went shopping for a 3mm allen T handle adjusting tool. Nobody has one.

The one in the plastic holders places sell should be long enough. 
T-handle sets are sold fairly widely.  Or catch a "tool truck" and buy 
an expensive one.

> How long does the tool need to be in order to clear the top of the fuel
> distributor? Can a normal Allen wrench be modified to work?

Sure, it only needs a couple of inches to clear.  The longer t-handle 
version just makes it easier and prettier.

-- 
Huw Powell

http://www.humanspeakers.com/audi

http://www.humanthoughts.org/