88 5kq - Intermittant Hard Starting

[SJ]

>From: "Huw Powell" <audi at humanspeakers.com>
>Subject: Re: 88 5kq - Intermittant Hard Starting
>

> >>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.
**** What I meant was that if nothing is swinging(my case), you can watch
the O2 voltage while adjusting mixture until you see the O2 sensor start
swinging. Now you are in the ECU adjustment window, and can switch to the
DPR current for the final adjustment of swing around the center value of 0.0
ma.


> > 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.
***** RIGHT! The Probst book has a full page chart of DPR current vs A/F
mixture - but the chart is for KE. Five pages further, you find one little
paragraph telling you the KE3 range is -10 ma to +10.

> > **** 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.
**** But, I'm stubborn. Analog milliammeter needles dont go negative, so its
hard to figure out what the swing range is. I"ll see how much play I have
with the meter needle zero adjust.


> 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.
**** Actually, with some tweaking, I got the OXS to steady on 0.5 volts when
using propane.
The reading wont jump around if you are in open loop - OXS not connected to
ECU.
You adjust mixture till DVM jumps from a low to a high. Now you are in the
ball park. Note where the 3 mm T handle is pointing.
Back off the adjustment a wee bit till DVM jumps from High to Low. Note
where the T handle is pointing.
Adjust mixture so T handle splits the difference between the above two T
handle directions.

"Doing it right" means getting a CO meter. Cheap CO meters are known for
being inaccurate. And I dont want to go begging for a CO meter at the local
gas station.
Besides, I'm stubborn . . oh, I said that already.

Ok, I thought of another way of adjusting idle mixture. Using engine RPM as
an indicator to tell you when you are close to or at Stoich mixture.
Engine warmed  up.
1. Disconnect ISV
2. Attach hose pinch clamp to regulate bypass air for controlling idle RPM
3. Disconnect DPR
4. Have propane tank with hose on standby, propane hose leads to air intake
at front grill
5. Start engine and let it settle down
6. Adjust hose pinch clamp for idle 750 RPM
7. Start feeding in propane, while observing RPM
        - if  RPM increases, mixture is lean;  if RPM falls, mixture is rich
        - Adjust propane for max RPM
        - reduce RPM to 750 using pinch clamp on bypass hose
        - shut off propane
8. Adjust 3mm mixture adjustment to adjust RPM to max (750)
9. For final adjustment, use either a controlled vacuum leak and/or the
propane to establish if you are at peak RPM, then tweak the 3 mm Allen screw
to match.

I think I have an answer for a "tunable" controlled vacuum leak.
A half inch copper pipe is the right size to fit into the "gas vapor valve"
port on the boot.
Drill a line of fine holes lengthwise into the copper pipe
Slip a plastic 1/2" hose piece(plugged at one end) over the copper pipe.
Covering or uncovering the drilled holes with the vinyl hose regulates how
much air gets "leaked" into the intake.

I'll try out this experiment tomorrow.

SJ





-- 
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.407 / Virus Database: 268.12.9/457 - Release Date: 9/26/2006