Date: Thu, 20 Dec 2001 12:38:37 EST
Reply-To: FrankGRUN@AOL.COM
Sender: Vanagon Mailing List <vanagon@gerry.vanagon.com>
From: Frank Grunthaner <FrankGRUN@AOL.COM>
Subject: Re: Boston Bobs day at the dyno
Content-Type: text/plain; charset="US-ASCII"
Doctor Tim, nice to see you back.
Many of the sentiments expressed by DrT are close to my own. Clearly the
depth of engineering behind the stock setups is unappreciated by many on the
list. However, incremental improvements can be made at some pain by the
dedicated amateur, as one moves away from the factory compromises and on to a
more specific goal (better performance in a particular rpm and load range,
better fuel economy at cruise, etc). With fuel injection engines, the
traditional enhancements done by hot rodders over the years (all designed to
improve the efficiency of the hot air pump we use to propel these vehicles)
are harder to properly exploit because of sensors and system limitations.
I applaud Boston Bob for taking the real step of verification of performance.
There is no substitute for measurement and the accelerometers genetically
implanted in the human buttocks are notoriously nonlinear. The performance
improvements often touted on this list remain colorful cracker barrel yarns
for the unwary until backed by data. The key to Tectonics Tuning has been the
use of a crude engine dyno to follow the effectiveness of engineering
modifications. The larger tuning houses generally have their own chassis dyno
for product development.
All this aforementioned notwithstanding, there are some lower cost approaches
to engine development and tuning. The least expensive is the g-tech
accelerometer. While not up to hard power numbers it is an excellent device
to compare modifications before going to the dyno. As a relative measure of
performance improvement, you only need to know mass and have access to a flat
stretch of highway. I use these numbers for most of my initial tuning
efforts. Recently though, I have built the Road Dyno kit, which measures the
instantaneous velocity of the crank as a function of time. Using a magnetic
pickup, knowledge of mass and road speed, and an inexpensive laptop, one can
measure the complete torque curve from 1200 to redline with a precision of a
few percent. The included software will also permit correction for
atmospheric density, temperature and relative humidity. Most recently, my son
and I have been tuning his supercharged VR6 ('97 GTi). It is perversely
amusing to see the true impact of improvements touted by many tuning groups.
Sometimes, on a good day the improvement can be kept to no power loss at all!
Watching my self-confident Caltech engineering major (with a 4.3 on 4.0
average) son weep after eight hours of bolting and rebolting has a certain
perverse pleasure component.
Boston Robert, I have encountered the
lean-at-full-throttle-after-increasing-head-throughput phenomena before. Of
course, not on the WB engine. My hat is off to a stronger masochist than I!
Anyway, I have encountered several different conditions. The simplest: the
full throttle switch not triggered, so the fuel injection system doesn't go
open loop. Second case: Air flow meter maxed out, no measured change in the
last 20% of the flow range. Now the fuel management computer is just running
on its internal map and putting out fuel as function of rpm only. Solution is
higher pressure. Larger injectors are not needed for 20% higher airflow. Your
observed lack of change with higher rail pressure suggests dirty injectors.
Check the spray pattern.
FWIW, I have just had the fuel injectors that I'm using for my turbo Audi 3A
project flowed, cleaned and calibrated. Cost was $85 here in the LA area.
There was a 35% variation in flow rate before cleaning, now they are matched
within 2% up to 85 psi. Patterns now identical.
Third, exhaust air leak near the O2 sensor. Higher flow can draw more
atmospheric oxygen to lean exhaust composition.
Some suggestions: O2 sensor based on heated regular Bosch unit is a guide
only. Really only accurate in a narrow range above and below stoichiometric.
Limited confidence in measured lean and rich AFR. Much preferred is the use
of a EGT. This can be placed right at the #1 exhaust port and can easily be
monitored for true lean conditions. I routinely monitor the O2 sensor voltage
cycling, but only to watch the control loop pattern. If it changed, out comes
the oscilloscope! I also monitor the EGT. With the Turbo 3A I will also
monitor the fuel pressure in the fuel rail, the injector duty cycle and the
signal from the knock sensor.
Don't give up,
Frank Grunthaner
|