[Supras] Whipping noise, throttle control, stock fuel piping

[berniek at technicaldevelop.com]

Jim and group:


    In summary, I'm just beginning to see a very small amount of fuel 
mixture leaning at 18 PSI boost.  It could be the boost extrapolation 
mapping of the stock ECU, but as you pass 350-375 flywheel HP (just a 
guess at this point) with low compression ratio, fuel requirements start 
to climb to about 3/4 GPM or slightly more if you run really rich. 


    First, I know there is gross distrust of devices such as the Dawes 
LED A/F indicator which reads the last rich side millivolt levels from 
the stock O2 sensor.  It took me a fair number of hours to get to trust 
it as an approximate indication of rich end A/F, by comparing its LED 
presentation with reading of an in-car portable Sun thermal conductivity 
exhaust gas analyzer.  The Sun has two thermistor based thermal 
conductivity bridges, and positive displacement pumps based upon 
loudspeaker like driven pistons with spring loaded intake and discharge 
balls as valves for sampling.  Between the two bridges in the pumped 
exhaust gas path is an activated charcoal filter to remove HC, but leave 
CO untouched (it requires reactivation every so often).  So by doing 
some simple addition and subtraction, it provides analog meter readouts 
of HC and CO.  The amount of CO produced is proportional to richness 
beyond stoichiometric mixture and is a good indicator of A/F ratio 
richer than stoichiometric since he engine does not use exhaust air 
injection.  It is a rarely used instrument about 20 years old, with 
analog circuitry based upon simple LM324 operational amplifiers.


    I also have a single bridge Marquette analyzer in addition to yet 
another Sun analyzer which is also single bridge.  I used to tune my GTO 
and Firebird with the earlier of these instruments.  Although they were 
(are, in the case of the Firebird with a Chevy BB in it) normally 
aspirated, maximum power occurs at about 12.5:1 A/F ratio with 
considerably greater than factory initial spark advance, reduced to 
about 38 degrees at high engine speed. 


    Further, the oxidation potential principle behind operation of a 
stock zirconium sensor is well known and predictable, albeit very flat 
at mixtures which deviate from stoichiometric.  When I first obtained 
the Dawes indicator, I bench tested it with a precision voltage source 
and found it to be exactly as advertised, with the rich, slightly rich, 
and optimum mixture LEDs separated in operation by almost exactly 40 
millivolts, corresponding to the rich end oxidation potential voltage, 
near one volt, of a zirconium O2 sensor.  The richest LED operates over 
the range of richer than 9.5:1, the next at 9.5-11:1, and the next at 
11:1 to 12.5.1.  There is a fourth LED which is a lean warning, 
indicating mixtures of 12.5:1 to about 13.5:1.  The warning LED covers a 
range of about 40 millivolts corresponding to the knee of the rich end 
curve characteristic of the zirconium sensor. 


    Until installing the upgraded IC and hardpiping, I ran 13-15 PSI 
boost with 550cc injectors, 57 trim CT26, 3" exhaust. Lex, Walbro, and 
the usual supporting modifications.  Afterwards I increased boost to 
16-18 PSI (at the 3000 pipe), which is the maximum I'm going to go with 
stock pistons.  I noticed that in transient fashion, fuel mixture was 
entering the 11:1-12.5:1 range, whereas previously that never happened 
at 35-40F ambient temperature before.  It was always richer.  Yes, the 
stated range is best for power, but I'd rather run richer for 
protection, at least until I get water injection installed.  Later 
yesterday I hauled the Marquette analyzer into the car, stuck the probe 
up one of the HKS exhaust outlets and drove for about 20 minutes doing 
some WOT runs on a favorite back road.  Sure enough, the A/F indication  
when pinching off the sampling hose to hold the reading was close to 
11.5:1 to 12:1, again enough proof for me about validity of exploiting 
the rich end 120 millivolts of the stock sensor.  Perhaps after the next 
state inspection I'll install a wideband sensor just upstream of the 
(now) single cat, as it won't last at the temperatures present at the 
elbow (the inspection police inspect the underside of the car with 
convex mirrors).  Having spent considerable time with exhaust analyzers 
in the car, however, I do no believe I will gain much which I do not 
already know. 


    In summary, I do not believe the ECU to level off in boost 
extrapolation levels, as it must track for safety up to fuel cut with 
the stock AFM (never hit fuel cut with the Lex).  The remaining and 
plausible explanation is that I'm near the capability of the 1/4" fuel 
piping, which at this point seems likely.  Moreover, under 18 PSI boost 
conditions, fuel pressure needs to be about 63 PSI (45 PSI static).  So 
available delta P for the fuel feed line is reduced as well.  Curves for 
the  Walbro GSS342 indicate that it should be a reasonable performer at 
that pressure.  Nonetheless, I now am beginning to understand why Arlene 
Lanmann has dual Walbros and bigger fuel line with a larger turbo. 


Bernie


   

Jim Jobe wrote:
> Ok I'll bite...
>
>   
>>     One more item of possible interest:  The stock fuel piping just 
>> barely starts to become inadequate with a genuine Walbro pump at about 
>> 18 PSI boost with 550cc injectors, a 57 trim CT26, and 45 PSI static 
>> fuel pressure.  Can describe how I know if anyone is interested.
>>
>>
>>     Thanks for the ideas.
>>
>>
>>     Bernie
>>
>>     
>
>
>