Hi everyone,
After lurking on this newsgroup for a few months, I've found a topic
which I can contribute to from my own experience. As an over eager
EE/CS type, I've often fiddled with some of the wonderful sensors
available on modern engines. As far as Exhaust Gas Oxygen sensors
go, I've installed them on two of my cars and many a customers, and
use them for tuning, both on the road and on chassis dynos. Most
(but not all) oxygen sensors give a similar 0-1 volt output, switching
abrubtly around stoiciometric (14.7:1). As reported by others, they
operate correctly at high temperatures (>300 deg C). There are a few
wiring configurations that I've encountered:
- single wire - signal wire only
- double wire - signal and ground
- triple wire - signal, two for heating the element, no ground
- four wire - signal, two for heating and a ground wire
The output from the signal wire is the same for each, once the sensors
are hot enough. The ground connection is critical and I suggest
connecting a ground wire to a washer at the sensor to the body of the
car. My 74 Daimler runs on leaded fuel and I expect the sensor to die
shortly (it has worked reliably for about a month with no problems).
I've heard some people advertise lead tolerant sensors but haven't had
the chance to test one. One option (here in Australia) is to switch to
high octane unleaded fuel while testing and then switch back after
removing the sensor. The sensor takes around thirty seconds to reach
operating temperature without any electrical preheating. Modern
applications require closed loop (lambda) control to come into effect
before this period, and hence the heated type. Wideband sensors are
available on some very high tech J******e cars but are very expensive.
They also require temperature compensation which turns a $60 oxy gauge
into $2500 - $15000 precise air fuel ratio meter.
Oxygen sensors have this step switching characteristic for a number of
reasons. Firstly, three way cats require precise air fuel ratios,
normally at stoiciometric, for efficient operation. The only info the
EFI system requires is at this switching point and a step characteristic
is very easy to process electronically. Secondly, the output from the
sensor varies greatly over temperature in such a way that the whole
output characteristic moves up or down by a few 100 mVs. The step point
remains the same, and as long as the switching threshold is always
between the full lean and full rich values, the post warmup temperature
will have little effect. Modern cars generally run in closed loop
mode while they are driven in a manner reflected by the EPA drive cycle
tests. In Australia, this doesn't include highway cruising or wide
open throttle conditions for which EFI computers can swith to open loop
mode and run lean or rich respectively.
Making a LED display is trivial for anyone who has put circuits together.
Bar graph driver chips are available and in conjunction with 10 or so
multicoloured LEDs and some protection and filtering circuitry can be
made into a very visible oxy gauge. Any volt meter would work, but the
signal does tend to flicker around the switching point. I do make and sell
such gauges locally for around A$60 but will be happy to give any hints
to those wanting to make their own. The bargraph displays and the plastic
casing are the most difficult parts to source.
Tuning a car using an oxy gauge requires a little bit of experience that
could be summed up in a paragraph. After connecting the sensor and
ensuring that the ground connection is very good, start the car and let it
warm up. Driving the car heats the exhaust many times faster than idle
warming. I set up the gauge to light green LEDs when the AFR is much
richer than 14.7:1 and red when much leaner. Note that the span of the
sensor output is around 14.2:1 - 15.0:1 (my guess). This means that rich
on the gauge is probably NO WHERE NEAR RICH ENOUGH for wide open throttle
especially for old, racing or turbo engines. I use it as an indication
for tuning while cruising and for diagnosing various faults such as
over lean or over rich conditions. (Over rich conditions tend to "wash"
the sensor with unburnt fuel giving either a full LEAN indication or a
flickering display which, from experience, differs from normal running.)
Idle is generally full rich for an old engine design. High load
conditions are also tuned to be relatively rich for performance and
engine longevity. I tend to lean the fuelling out for cruising at or
below highway (100 kmh) speeds when the engine is lightly loaded for
efficiency. An EFI injector pulse width duty cycle display on my EFI
hand controller allows me to find the optimal operating point for maximum
efficiency for constant road speed. An on-line programmable EFI and
ignition system makes this task easy (but is not neccessarily legal).
I have no idea how carbies are tuned or whether it's possible to adjust
the fuelling for different operating conditions. IMHO carbies and
distributor are the first thing to go if an engine builder has
performance, reliability and efficiency in mind - but then, I may be
biased as I design and build programmable EFI/ignition systems for a
hobby.
If anyone is interested in a pair of 2 inch SU carbs and inlet manifold
to suit a Jag XK engine please let me know. As far as their condition
goes - they worked well last time they were bolted to the Daimler and
were once very shiny.
- Robert
'74 Daimler Soverign
'55 Austin Champ 4WD (+ one for spares)
'76 + '77 Toyota SillyCar
PS. Exhaust Gas Oxygen is commonly reffered to as EGO. My initial gauge
had the label "EGO METER" and started many a conversation as it went up
and down depending on how fast I wanted to drive.
Robert Dingli dingli@mullian.ee.mu.oz.au
Melbourne University
Australia
|