Fullscreen HTML5Med Res (??MB)HTML4

Electronic Engine Management Pt.5: Oxygen Sensors – by Julian Edgar A major incentive for adopting engine management systems was to meet the strict exhaust gas emissions legislation enacted in several geographical areas – notably the huge Californian market. To meet these strict emissions levels, manufacturers had to start tuning their cars to meet these criteria, rather than optimising for power and economy. The initial response by manufacturers to Australian legislation was often half-hearted, with Australian Design Rule (ADR) 27A back in the mid-1970s giving us cars which drank fuel with a voracious thirst, overheated and stalled in traffic. This reflected poor design adaptation rather than any intrinsic prob­lems with the new regulations. Unleaded petrol The coming of unleaded petrol (ULP) in 1986 meant that engines had to be redesigned to run on lower octane fuel which lacked lead. For some local makers, their old engines simply couldn’t be updated and so new engines were introduced. Holden replaced its venerable red/blue/black 202 (3.3 litre) engine with the Nissan 3.0 litre straight six, for example, before switching to an American-designed 3.8-litre V6. As well using the new fuel, the car manufacturers also had to use a catalytic converter. A catalytic converter changes the “colour” of several of the more noxious pollutants to “green”, thereby benefiting the environment. However, leaded fuel will poison a catalytic converter and so must not be used. (Inciden­tally, ULP will always give a black tailpipe – irrespective of mixture strength). Air-fuel ratio This engine uses a single-wire (unheated) oxygen sensor. It is shown bolted through the top of exhaust manifold, just to the right of the turbocharger assembly. 42  Silicon Chip Also required for efficient catalytic converter operation is an air-fuel ratio that’s very close to stoichiometric (14:1). This means that, for the catalytic converter to work best, 14kg of air (or 10,000 litres) must be mixed with every litre of petrol. Incidentally, the stoichiometric ratio – where theoreti­ cally best combustion occurs – varies from 14:1 to 14.7:1, ac­cording to the reference used! The authoritative Bosch Automotive Handbook lists it as 14:1. Fig.1 shows the relationship between varying air/fuel ratios around stoichiometric and the production of the pollutants carbon monoxide, hydrocarbons and oxides of nitrogen. An example of a heated oxygen sensor from a Nissan engine. Note that there are three leads running back to the plug connector. The stoichiometric ratio isn’t, however, the best for either maximum power or economy, with the mixture needing to be richer or leaner respectively to achieve this. Mixture feedback loop Car manufacturers were therefore faced with a dilemma – did they design for power, economy or emissions? They solved this by using a feedback loop which allowed them to have their cake and eat it too. At constant throttle settings (that is, cruise), the exhaust gas is monitored for mixture strength and information from the sensor fed back to the ECM which in turn controls injec­tor pulse width openings to give a stoichiometric mixture. Fig.2 shows the structure of the feedback loop. At full throttle (sensed by the throttle position switch), the system goes open loop, with the exhaust gas oxygen (EGO) sensor ignored and the mixture suitably enriched for power. Conversely, lean mix­tures are used during a trailing throttle. The EGO sensor keeps track of all Fig.1: the relationship between air/fuel ratio & the production of various pollutants. February 1994  43 CONTROL UNIT FEEDBACK SIGNAL INJECTION PULSE of that point. Fig.5 shows the voltage response of a typical EGO sensor. Note that its output voltage does not directly follow oxygen concentration, especially for lean mixtures. OXYGEN SENSOR Mixtures revealed OXYGEN SENSOR INJECTOR FUEL INJECTION COMBUSTION ENGINE Fig.2: the EGO sensor feedback loop. At full throttle, the system goes open loop & the EGO sensor is ignored. of these mixture varia­tions. It can be one of two types – titanium or zirconia oxide. The zirconia type is more frequently used and generates a voltage output. A cross-sectional view of a typical zirconia EGO probe is shown in Fig.3. Its operating temperature is from 300°C upwards and it is sometimes electrically heated to bring it up to this temperature. Its performance in unheated mode is usually satis­factory, though, and so some manufacturers run it like this. The other type of EGO sensor – the titanium probe – must always be electrically heated. Instead of generating its own voltage output, the titanium probe changes its resistance in response to different oxygen levels in the exhaust. It is mounted close to the engine in the exhaust manifold to ensure that it is quickly heated to operating temperature – see Fig.4. Both probe types are calibrated so that their output chang­es rapidly around the stoichiometric point and is symmetrical in response to either side BUSHING (ELECTRODE) The most interesting aspect of EGO sensors is that it is easy to access their output and then see for yourself the mixture variations that occur as the car is driven. It’s a bit like gaining sight after being blind – suddenly you can see the cold-start and full throttle enrichment cycle working, the overrun injector cutoff, the time when the computer is in closed loop mode, and when the computer goes open-loop. And in a car running modified EFI – whether by chip rewrit­ing or cruder means – it can be clearly seen where rich or lean points occur in real driving conditions. Obtaining a readout from the common zirconia EGO probe is easy, because the commonly-available LM­ 3914 LED display driver IC seems almost custom designed for the purpose. By following the attached circuit, a 10-LED display mixture meter can be easily and cheaply constructed – see Fig.6. The voltage output from the EGO sensor is usually between 0-1V, with the sensor in most cars giving 0.5V at the stoichiometric point. The IC uses an internal reference of 1.25V and this is easily reduced to 1.0V by a trimpot (VR1). Inside the LM3914 is a series of op amp comparators and these each compare the signal voltage from the EGO with a divided reference signal. Each op amp in turn drives an LED (LEDs 1-10) and this produces a moving TERMINAL SUPPORT (LEAD WIRE INSULATION) LEAD WIRE ATMOSPHERE SPRING EXHAUST MANIFOLD Fig.3 cross-sectional view of a typical EGO sensor. 44  Silicon Chip Fig.4: the EGO sensor is bolted into the exhaust manifold, close to the engine. O2 SENSOR VOLTAGE Obtaining a readout EXHAUST GAS ZIRCONIA PIPE EXHAUST MANIFOLD CO CONCENTRATION O2 CONCENTRATION RICH THEORETICAL AIR/FUEL RATIO LEAN Fig.5: the output from a typical EGO sensor in response to O2 levels. display as the input voltage rises or falls. Pin 9 controls the display mode. Leaving pin 9 open cir­cuit produces a dot display, while tying pin 9 to pin 3 produces a bargraph display. The 680Ω resistor sets the display brightness. The components can be bought individually and mounted on a board, or the Jaycar Car Battery Monitor kit (which uses the same IC and comes with 10 square LEDs) can be modified to work in the 0-1V range. The circuit shown is about the simplest possible. Variations include using diodes to limit voltage spikes and slowing the response time of the meter by using capacitors to filter the input signal. Connecting the meter to the sensor is straightfor­ ward – just connect it in parallel with the ECM. If the EGO sensor is a 3-wire type, then use the workshop manual (or a high input-impedance multimeter) to sort out which is the sensor output wire. If the EGO sensor is a variable- LED1-10      10 11 12 12V  13   14 15 16   17 18 1 3 INPUT FROM OXYGEN SENSOR VR1 5k 5 IC1 LM3914 6 7 2 4 8 680  Fig.6: the readout for the oxygen sensor is based on IC1, an LM3914 dot/ bar display driver IC. It functions as a simple LED voltmeter. resistance type (rare), then obviously the LED meter will be inappropriate in this form. Finally, connect 12V and earth and the meter should come alive when the sensor is up to temperature The meter’s output display will depend on the type of ECM your car uses. In closed-loop mode (with the EGO sensor having an input into injector pulse width decisions), the mixture will cycle rich-lean-rich-lean, either at a few Hertz or almost in­stantly back and forth. Alternatively, some cars will cycle for a few seconds and then settle at the “correct” mixture, holding it at the point until a throttle change. Others will require perhaps 60 seconds of constant-speed cruising before holding the mixture steady on the display. However, flooring the right foot will instantly give a rich readout, as the ECM software commands for maximum power are invoked. If your car runs plain ol’ carbies, you can still use an oxygen sensor. It will help if you use ULP in your car, as other­wise the sensor will be prone to lead fouling. Oxygen sensors are quite expensive when new but a car wrecker importing engines directly from Japan will have used sensors available. I bought two sensors in this way for $15 for the pair. Depending on the design of the sensor, either a nut or mounting plate will need to be welded to the exhaust to allow it to be fitted. Place the sensor as close to the engine as possible, making sure that it will get the gas flow from SC all cylinders. The completed mixture display meter. It connects directly to the EGO sensor. February 1994  45