The season for politics is well and truly upon us, and I’m not referring to Nigel ferriage & the
entourage, I am however referring to the IAAF interview recently screened on TV, in a desperate
attempt to convince the general public that new diesel technology is clean! It’s a hell of a lot cleaner
than it was, however for all our sakes let’s clean up our act and treat the motoring public with the
respect and the integrity they deserve!
The combustion of diesel is so inefficient it now requires three post combustion processes to meet
current emission standards .the assumption is that these critical systems maintain total efficiency at
all times! It has been established that Nano soot particulates pass through the treatment process
and present a genuine health threat. A more illustrative TV programme appeared to illustrate that
real time emission analysis revealed that manufacturers emission claims bore no resemblance to
actual in service reality.
I feel like a turkey voting for Christmas, 80% of our repairs relate to failed diesel emission systems, if
the IAAF want to act like the king with the magic coat that’s fine, but nudity is always ugly in public.
Keeping the theme of emissions in mind and both feet firmly on the ground let’s take a look at how
we might interoperate various forms of data when diagnosing petrol emission problems.
The first problem is that emission control is adaptive, it’s supposed to be! This means that superficial
analysis may disguise a problem. The design intention is to adjust fuelling control within a near
perfect lambda 1 window. Treating residual emissions similarly to diesel! With three way catalytic
conversion and separate NOx reduction.
The clues are always there if you know where to look. Begin with serial data, no! Not dtc’s actual
data and specifically, adaptive correction. Fuel trim is calculated by feedback values from oxygen
sensors pre and post catalyst. I use the expression oxygen deliberately, as they have no possibility of
measuring ratios. They can be fooled by environmental problems, especially air leaks.
If you are to rely on lambda feedback, then focus on the post cat sensor as it does not lie. Trim
correction should not exceed +\- 2%. The maximum deviation is around 25% at which point a dtc
will be generated. It’s often a good idea after collating data to reset adaptive data and re assessing
the true state of exhaust gas emissions less the effects of correction.
There are many reasons for trim related deviation,
Induction system leaks
Hydraulic\ electronic injector faults
Fuel pressure, flow, cavitation faults
Sensor range errors
Wiring problems
Adaption errors
Software
The responsibility of the technician is to eliminate in a methodical progressive manner all the above
possibilities. The first steps should include a pressurised induction system smoke test, this can
further be applied to the exhaust system. In the absence of any system leaks check fuel trim
correction. The correct response of the pcm is opposite of the lambda signal output.
I am lean go rich, I am rich go lean. However when making assumptions about mixture ratios
remember the sensor only understands oxygen ratios. Next carefully evaluate sensor feedback
values against good known examples. Wiring and environment problems will effect sensor outputs.
Fuel pressure flow and cavitation must be checked using a suitable gauge.
Many years ago we developed a means of evaluating fuel trim in real time using an oscilloscope, this
offered a distinct advantage as it was instant real time response. I still remember a particular author
stating that it was not possible to measure an air mass meter in such a way! Form your own opinion
from the two following pico waveform samples. They are taken from a bmw with dual lambda
feedback, pre and post sensor evaluation and the air mass meter load input.
Image 1, short term trim lean deviation.
Red channel bank 1 sensor 1, green channel bank 2 sensor 1, black channel post cat sensor
Left to right , fast idle normal activity , note post cat sensor high voltage @ 0.8v = no residual
oxygen. Throttle off excessive low voltage = excess oxygen, correction taking place prior to normal
switching and wide open throttle response. Note a normal rich response from all 3 sensors.
Following a return to natural idle and excessive oxygen. Conclusion short trim, lean deviation. Air
mass meter output correct, full load simulation correct, conclusion not a problem with physical fuel
delivery.
Compare improved fast idle switching frequency, short term trim recovery now just over 5 seconds,
normal switching prior to wide open throttle, final natural idle normal recovery.
Just in case of remaining heretics, air mass meters have a unique profile, when compared with
lambda and throttle signal input request they form a unique and predictable response within 175
Ms.