Fuel Injection system

Download the RANGE ROVER 14cux manual here.

Note this is far from complete but based on my experiences that you wont find elsewhere- if you think something is wrong or want to add more information on the 14CUX, please drop me an email. Mark@blitzracing.co.uk

Note these notes are based on the 1992 ECU. Further design changes did take place after this. 

Overview

14CUX was the latest in a long line of fuel injection systems made by Lucas. Lucas licensed the designs from Bosch, the inventor of the system. Earlier iterations of this system were called 4CU, 13CU, and 13CUX.   14CUX is a fuel injection system only, meaning that it only controls the fuel injection on the engine - newer engine management systems do much more. These systems are pretty primitive in terms of diagnostics but they have the benefit of being serviced without special tools. They don’t have many sensors, and those sensors can be monitored with fairly common test gear and ss Lucas electrical systems go, these 14CUX systems are actually quite reliable. Sensor wise, these are air flow, fuel temperature, engine temperature, and throttle positions,  road speed and on catalytic cars a lambda input. The "hot wire" air sensor gives an electrical signal depended on the amount of cooling air that passes over a preheated wire, that's then compared with a reference hot wire that has no cooling air flow. The various sensor inputs are then fed into the ECU,  and converted into a digital signal that address's a pre determined fuel map held in a 27128 or 27256 type Eprom to give an injector pulse signal appropriate for the fuel requirements of the engine at that point. The ECU control's the fuel allowed to each cylinder by changing the time the injectors are open for (batch firing them) , combined with a variable fuel pressure regulator on the end of the fuel rail. This regulator alters the fuel pressure dependent on what the inlet vacuum is at any given moment and  will vary from 24- 36 psi .  This simply maintains the fuel pressure at at a set value above the pressure in the plenum chamber, so exact fuel metering can be maintained. To maintain a steady tick over, there is a stepper motor that controls an extra air path past the throttle body.  The Fuel regulator is set to 2.5 bar or 37.5 psi if the inlet vacuum pipe is open to the atmosphere.

Initial start up sequence.

  The sequence to restart the engine actually starts as you turn off the ignition switch. As the ignition voltage is removed from the ECU, the unit sends out a signal to the stepper motor attached to the plenum chamber to wind it fully backwards and allow maximum air into the plenum chamber. This can be heard as a buzz from the stepper motor as the engine dies. As the ignition voltage has now gone, the engine simply stops with the stepper motor in maximum air position.

 On turning the ignition back on, a short pulse  (about 1- 3 seconds) is sent to the fuel pump to pressurise the fuel rail. Once the starter motor starts to turn the engine, a 12v pulse is fed back to the ECU from the negative side of the coil as it the ignition amplifier switches. The ECU then turns on the fuel pump and energises the fuel relay that provides a fixed 12 volt supply to all the injectors. The transistors in the ECU  starts to ground the  injectors with a longer pulse than the normal idle pulse for about 3 seconds. This provides enough fuel to start the engine, combined with the stepper motor still being in its wide open position. Once the engine has fired, the air flow meter then takes over feeding the air flow volume back signal to the ECU and the injector pulse width is reduced to match the fuelling requirements for the engine at tick over. The stepper motor is also wound in to stabilise the idle at around 800 RPM. This system accounts for the short burst of higher RPM at tick over as the stepper goes from wide open to part closed during the start process.

 One side effect of this system is if air leak develops any where in the inlet system, the engine will start and run for 3-4 seconds and then die. The initial over rich mixture will allow the engine to run, but once the air flow volume comes into play, (Now reduced because of the air leak) the injector pulse width is reduced to the point where the is insufficient fuel to keep the engine running, so it dies. The whole pipe work and breather system around the plenum chamber is pretty finely balanced and can be easily go out of tolerance should an air leak develop.

Fuel control system.

Once the engine is running the ECU then starts to take the various sensor inputs into account.  These are:

Water and fuel temperature from 2 thermistors. 

A variable voltage for the overall mass of air entering the system from the mass airflow sensor.

The engine RPM from a wire leading from the coil trigger.

The throttle position from a potentiometer mounted on the throttle butterfly.

Road speed from a "chopper disc" transducer that's inline with the throttle cable.

Fuel mixture from a lambda probe in each exhaust.

Although its of mid 1980 electronics and programming, its still quite smart with the Catalyst fueling maps. On  starting the engine the lambda probe output is measured at tick over, and if it averages rich or lean the ECU alters a "base "setting in the ECU that either enrichens or leans off the entire fuel map, unit the idle mixture is correct. This takes around 15 seconds to achieve, and could roughly be described as a "leaning process". Once this base level is fixed, the normal closed loop part of the lambda control circuits then takes care  of the "instantanous" mixture control throughout the rest of the fueling map above tick over. The ECU has the ability to alter the overall fueling over a wide range with this system, but if you start to deviate to far from its original "fixed" fuel map, the car can become hard to drive as the ECU constantly makes large corrections. When full power is needed at over about 3/4 throttle or  about 3400 rpm,  (both settings are programmable) the ECU goes open loop, and simply supplies a richer mixture, dependent on the fuel map values and fuel pressure.

There are 3 parts to lambda controlled  fueling.

1)    Tickover- Adjust the overall fuelling- long term fuel trim value
2)    Mid range - Closed loop control- small and rapid corrections- short term trim value
3)    Full Power - Open loop, no lambda control, both probes should show .7 volts or more and remain there to show mixture is around 12:1 for maximum power and engine safety.

The other variable the ECU can "learn" is the voltages generated by the throttle position potentiometer over its maximum and minimum range.

Idle control.

This is a fairly crude affair, utilising a air bypass screw (called the base idle setting) and a simple stepper motor controlled air valve to keep the idle steady as the engine loads vary. The stepper motor has 180 different step positions, and each time the ignition is turned off, the stepper motor pulls the air valve wide open, by being pulsed 200 times. As only 180 steps are available, it will always reach a "home position", and from this point the ECU keeps track of its position by counting pulses from the home position. One weakness of this system, is if the stepper motor sticks, the ECU looses its correct position, as there is no feedback to say where it is which leads to an unstable idle. Another weakness is how it crudely controls the tick over. If the engine is running above the required RPM at tick over, a burst of pulses is sent to the stepper motor to reduce the air supply. The ECU then waits a few seconds for the mechanics of the engine to respond. This time "constant" depends on the engine dropping its RPM in a fairly controlled manner. Further tweeks then take place if the RPM is still outside tolerance. A problem occurs if the engine RPM drops faster than predicted due to fueling errors, or ignition problems, so the engine RPM drops too far. After the wait time the ECU detects the RPM is now too low, and winds the stepper motor back again and waits again, at which point the RPM goes too high. The process then repeats itself, so the idle remains very unstable.

Road Speed sensor.

This has the function of turning off idle control when the car is moving, so to maintain engine braking when slowing down, operating the fuel cut off on the over run, and in the case of of the 4x4 map, a top speed limit.

ECU  Fuel maps. 

The fuel map is held in a 27C128 or 27C256 type Eprom, dependent on the revision of ECU. Although the 27256 device has twice the storage capacity of the 27128 type device, it appears that you can still put the smaller device in place of the larger one as the highest address line is not used. The device itself contains multiple maps, such as catalyst, non catalyst, limp home, and some country specific ones for particular fuels. These are selectable with a plug in programming resistor found under the passenger dash board.  The maps are referred to by the colour of covering plastic on the resistor wires,  red, yellow, green, white and blue, and no resistor- (this defaults to the catalyst map.) Changing these values alters the base address used to find the relevant fuel map in the Eprom.    On the early 14CUX units the Eprom device is soldered in, and does not have the classic quartz window in the chip. Later units have a plug socket, and plastic cover over the Eprom. To reprogram the early units the old device has to be cut out and the chip legs carefully removed one at a time. Its almost impossible to get the chip out in once piece, so once you have started there is no going back ! Its recommended that a “turned pin ” type chip socket  is soldered into the board as these have better electrical connections in the harsh environment of a cars foot well where the ECU lives. Unfortunately due to the very restricted data available on this ECU, many people have simply given up trying to tune it, to switch to an after market fuel injections control unit like Emerald or Megasquirt, where programming data is freely available.

ECU trim values.

 There are many values that can be altered  such as the upper rev limit, the amount of cold start fuel supplied, the point at which the system goes open loop, plus the overall fueling levels of the map. It is virtually impossible to study the binary data in the EPROM, and tweak it manually. If you are that way inclined the fuel maps can be studied by extracting the binary data with an EPROM programmer, and using a simple hex data editor to study the data. A suitable hex editor can be found here:  

http://www.chmaas.handshake.de/delphi/freeware/xvi32/xvi32.htm

Memory map is 64K bytes big (16 address bits), and the program/calibration chip sits at address range C000-FFFF. Tune data is at the beginning of the chip, from chip address 0000-07FF. Program (but not entry point) officially starts after at 800h, although there are several 'hidden' calibrations (like overriding rev. limit) just after 07FF. Actual program entry point can be found from the reset vectors at the top end of the chip with reference to the 6803 processor data sheet.

Fuel maps are 2-axis maps 16 by 8 sites, speed versus load with 16 speed sites and 8 load sites. Load is calculated from the AFM reading by a formula which includes a scaling factor from the engine speed. This means that (maybe) 30 litres per second could give max load (site no.7) at 1000RPM, whereas it could take (maybe) 150 litres per second at 5000RPM to give the same load site. (Those figures are illustrative guesses!). Speed axis is a simple set of breakpoints with RPM. The breakpoints I believe have always been the same as follows:

If you can adjust whatever you use to view the chip data to display 16 columns, the first map should be glaringly obvious. Since the data groups are not on 16-byte boundaries the others may be less so however, but the maps are almost identical so should be spottable. Directly after each map is the scaling factor applied to it, a 16-bit value usually of the order of (hex) 6590 or maybe 61AB. Fuel maps usually start with several instances of the same value, often (hex) 14 or 21. Each fuel map is at the start of a section of tune data applicable to that tune (as determined by the tune resistor) which extends down to the start of the next fuel map. Many things like acceleration enrichment, order of accessing analog inputs, cold start maps and other stuff are included on a per-tune basis, but some stuff found in the first (tune 0) area applies to all. There are six areas including tune zero. All are exactly the same size except zero.  Individual map  locations appear to be:

0x0000-0x007F
0x0267-0x2E6
0x0379-0x03F9
0x048B-0x050B
0x059D-0x061D
0x06AF-0x072F

To re program the 14cux, Lucas sold what was described as a "development box" that was basically a programmable ECU with a laptop interface and software. This unit replaces the standard ECU, and by using a rolling road and lambda probe outputs, the fuel map can be tweaked until the best conditions are reached. This modified map can been be programmed into a blank Eprom that then fits into a standard ECU. It is not possible to communicate with the stock ECU in anyway to alter the fuel mapping. One reason aftermarket chips are so expensive is the development units where very expensive, and now pretty rare, limiting the number of places that can genuinely change the fuel maps. As the fuel maps are on standard Eproms, they are pretty easy to copy given an Eprom copier although some aftermarket Eproms are sold with a scramble chip, that needs the correct address sequence to mimic the initial start up of the ECU CPU before the fuel map is unscrambled. A standard Eprom programmer will not follow this sequence so the data is scrambled. 

  After market  performance chips.

It a light car like the G33 where the engine can rev easily, much more fuel is required through the higher rev  and load range (4500 rpm plus) to prevent damage. The original leaning off of the mixture was a "feature" of the Range Rover to try and save a little fuel at higher revs and keep the emissions low. An evaluation by the magazine Car and car conversions found one G33 to be down to around 140 BHP (at the rear wheels), and by simply increasing the fuel pressure an extra 20bhp was gained. As already stated there is also a speed limiter plus a rev’ limiter also built into the software, so there is good reason to upgrade the chip. At its most basic,  simple modifications are to increase the injector pulse widths above 4500 rpm to provide more fuel, and remove the rev’ and speed limiters. Beyond that there is also scope for improving the starting, mid range running and mid range fuel economy.  Getting maximum power normally involves making the engine run at around 12:1 air /fuel ratio, but at this point both fuel economy drops, and emissions go up.  So manufactures chose a compromise point of around 14.7:1, (shown as closed loop mode) when the engine will fill the emission requirements, and  fuel economy even though power will drop a little. So the first step for more BHP is to re-chip to richen the mixture. A richer mixture also makes the engine run cooler, and its possible to advance the timing a a couple of degrees to make most use of the extra fuel with lower combustion temperatures. 

Catalyst versus non catalyst fuel maps.

The 14CUX has all the hardware to use Lambda probes to keep the fueling spot on to best suit the catalysts, and keep the emissions low. Any G33 produced pre August 1992 however was not fitted with the lambda probes or catalysts as it was not a legal requirement, and not running a catalyst can have a slight performance advantages, by not restricting the exhaust gas flow.  Having said that the catalyst map goes into open loop mode at anything more than 3/4 throttle or around 3000 rpm, so the fixed map section can be tweaked to run a bit richer in open loop as there is no lambda control.  Its is not true to assume that running the ECU in fixed map will provide any more power than a correctly modified catalyst map. Things have moved on since the nineties when the G33 was produced, and TVR have done considerable work with the 3.9 engine fitted to the Chimaera with catalysts and have produced a very good fuel map, but it does require the installation of Lambda probes and wiring. Its worth noting that due to there being at least 5 maps in the Eprom, only one map may be modified. Don't assume that a TVR map switched to non cat will provide any performance advantage, as it may well be just a standard Range Rover map.

The catalyst fuel map has been rolling road tested (rear wheel figures) with the non cat G33 (read Range Rover)  and TVR map and the results are as follows:

Max Power
Standard 179bhp
Modified 193bhp

Max Torque
Standard 198 ft-lbs
Modified 207 ft-lbs

Rev Limiter
Standard 5200 rpm
Modified 5680 rpm

The torque curves for both ECU's are virtually flat with max torque flat between
 3200 and 3800 for the standard ECU and 2400 & 4000rpm for the modified map

 Car ran to 84mph in 3rd with the modified map  and 80mph with the standard.

Why bother switching to a catalyst maps over your fixed map ?

The standard fixed map needs its idle mixture set up accurately on the air flow meter adjustment to get the car to idle well and not cause problems with an unstable tick over. The normal procedure is to set the CO trim value to 1 -1.5 volts on the air flow meter for the non catalyst cars. Unfortunately any changes in the fuel pressure, fuel map, or valve timing (such as cam change) can throw these settings out, and the fixed system may not be able to cope, affecting the cars low speed drivability. By changing to the catalyst map these variances can be automatically compensated for by the ECU without a full and expensive remap. On the idle carbon monoxide settings I have tested from 0 - 3.5 volts on the CO trim value, and the system can cope within this range, without dropping out of "lambda range". From this I believe the CO trim value is ignored on the catalyst setting.  On the fixed map, it really needs to be setup with a CO meter, as its quite possible to foul the plugs if this setting is out.   Running the catalyst map is good for fuel economy, emissions and low speed drivability. Above about 3400 rpm, maximum power can still be maintained as long as an upgraded ECU chip is used such as the TVR Chimaera one already mentioned.  The important thing here is you have nothing to loose as its the physical catalyst that reduces the power due to the restriction in gas flow, not the fueling map.  The only down side is it can introduce a slight snatch in the power delivery at low speed, when the airflows are very low. This does not show on the standard Range Rover cam, but on cams with longer durations, the airflow is lower "off cam". This can cause the mixture to run lean, combined with a system the ECU uses to keep the catalyst clean. This involves constantly "sweeping" the mixture rich and lean around Lambda 1 and during the lean sweep, it can run so lean, the mixture does not ignite reliably so the car snatches. On the plus side, fuel economy can be much better.

Lambda probes

These are fairly conventional units generating a voltage dependent on the amount of remaining oxygen in the exhaust after the burn has taken place. The voltage should be within the range of .3-.7 volts , with .3 being lean and .7 being rich.  The only unique thing about them is the M12 thread, that allows the price to be kept up, as most generic units will not fit this thread. They also have a heater coil fitted that is powered from the fuel pump relay, so give a rapid warm up time. Once the  ECU is powered on, will not modify the fueling until it receives its first changing signal from the probe, but from this point it expects the signal and if the signal fails or goes out of range it will generate fault codes showing fuel supply/ injector problems  or lambda faults. If you source probes in the UK, they are VERY expensive at between £60 – 100 each. All is not lost however, as they can be sourced at around £14 each from the USA from a company called Global automotive: 

http://stores.ebay.com/GLOBAL-AUTOMOTIVE 

Search for Land Rover Oxygen sensor, and you will find them listed for the 3.9 engine.  

To fit probes to the down tubes, you need to remove the pipe that connects the exhaust manifold to the first exhaust box. A hole big enough to take the probe tip needs to be drilled on the side or upper section of the pipe, as near the manifold as practical. Make sure you have sufficient clearance for the probe and wiring once its reassembled.  Don’t install the probes on the lower parts of the pipes where condensation can run down into the probe sensor. You now need to weld a M12 fine nut (or half of one) to the down tube to hold the sensor.

 Wiring the probes

 The connections are already installed in the engine bay for the probes, and can be located jammed down amongst the wiring loom that runs along the engine bulkhead wit a pin circular connector for the left and right hand probe. As the probes come without the relevant connectors, you need to cut the connectors off and fit something suitable like bullet or spade connectors. I also left 3 test connectors free for subsequent testing. The Lambda control wire is screened, as the signal values are very low, and external electrical noise will interfere with the signal. If you need to extend the cables, its recommended that you use screened audio cable for this one signal. 

Wire Colours 

Switching the ECU tune resistor.

You could simply buy another plug in resistor pack, or just remove the original 470 ohm resistor replace it with a 3900 ohm resistor (half or quarter watt). The leads are long enough to use a terminal block to join the connections if you don't have access to a soldering iron. Alternatively a better bet to make a switch able option is to insert a simple on off switch in series with the tune resistor, and then solder a 3.3k ohm resistor across it. Switching the switch "on" selects the non cat' map, whilst switching it "off" puts the resistors in series, switching to the cat' map.

ECU modifications

As already stated, the ECU will need the standard Range Rover fuel map chip removing, and an upgraded TVR chip fitting. There is little point in switching to the cat map without this.

Testing.

You need preferably an analogue test meter, with a suitable range to measure up to 1 volt or over. With the engine running for a few minutes, the meter needs to be connected across the white and black connectors on the probe outputs. This should show a fluctuating voltage of between . 3 and .7 volts as the engine idles. If its not fluctuating, and sticks near 0 or 1 volts there's is a problem and the ECU is not compensating for the mixture variances,  some sort of fueling problem or possibly a failed lambda sensor.

Further fueling adjustments.

Firstly and most important. Dont try to make any fueling adjustments without monitoring both Lambda probe outputs. It is impossible to guess what the fueling is doing without this, and could prove fatal to your engine if it runs lean under load. 

The lambda probes have quite a narrow range of operation of around 15:1 air fuel to around 12:1 air fuel, but as long as the origional map is not too far out the readings will be good enough. In normal operation the voltages range around .3 - .7 volts, but even on a fully functioning system, you may see glitches as high as 1.2 volts. Due to the rapid response times, an LED bar graph display is the best item to make a cockpit gauge that can be viewed as the car is driven. Although you can buy such gauges "off the shelf" for around £25 (per probe), these will display only one side (or one probe) at a time. I found a low cost alternative (around £15)  was to buy a stereo audio VU meter kit so both banks could be seen side by side in real time. The particular kit I purchased did require some tweeking of the input resistors to get it to give a full scale reading of 1.2 volts. I also managed to locate some old Lambda plugs and sockets that allowed the unit to plug in without splicing into the loom. It also conveniently provides the 12v supply needed to drive the gauges from the lambda heater supply rail. 

Viewing the normal fueling

This can give some interesting in site into how this ECU works.  Firstly run the engine up until its warm, the disconnect the ECU for a few seconds to remove any fueling preset values. Reconnect the ECU, and start the engine and observe the results. The Lambda probes should already be hot, so the rising voltages should show rapidly BUT, with no fueling offset yet applied it may read very rich or lean. The ECU will then adjust the "base" mixture over a period of about 15 seconds, so the gauge should then start to fluctuate rapidly as a correct fueling point is reached. This short period can give you some idea of how much correction to the basic fuel map is applied by Lambda feedback, which is useful if you are adjusting the fuel pressure to minimise the amount of base correction applied. 

Once out on the road in light to mid throttle, the lambda probes should hold the fueling close to the optimum point, and the gauge with constantly fluctuate. Once you go above around 3000 rpm or 3/4 throttle, Lambda control is released and the ECU goes open loop. At this point the gauge should rise to .7 volts or more and stay there. This is most important as any reduction from this value shows there is a problem with the basic fuel map, fuel pressure or fuel pump delivery. A lean mixture is very bad news as it causes the combustion temperatures to rise, and can lead to "micro" welds between the piston rings and bores, that can make a liner lift a fraction, which can destroy an engine block. YOU HAVE BEEN WARNED.

Adjusting the fuelling. 

The first and foremost thing to remember is Lambda feedback is king, and the ECU will always try and keep the mixture correct whatever you do to the various inputs. In an ideal world, the unaltered part of the fuel map would be so near perfect, that very little correction has to take place, but this is rarely the case. The name of the game here is to alter a sensor input to try and minimise the amount  of correction taking place. 

1)      Fuel pressure.

 By increasing (or decreasing) this by a couple of PSI will alter the fueling through the engines closed loop range. The ECU’s idle adjustment system will pick up this change in mixture, and alter the base setting to bring it back to lambda control range. If you go to far either way you will exceed the ECU’s compensation range, and it will go into a fault status. The cars will also become hard to drive before this point is reached, as the ecu battles to keep control. Once the ECU goes open loop near full throttle, the lambda control is ignored, so increasing the fuel pressure slightly can be used to provide more fuel under full load. The normal fuel pressure is 37 psi, and no more than 40-41 psi is recommended, as an optimum point for the standard fuel injectors.

Setting or measuring the fuel pressure has to be done WITHOUT the vacuumed pipe connected to the plenum, as the plenum pressure will alter the reading. Simply disconnect the vacuumed pipe, let the car  tick over and take the reading. 36-37psi  is normal, and a couple of PSI ether way will make a big difference to the open loop fueling.

 2)    "Altering"  engine temperature

 Crude to say the least. The first problem is the resistance / temperature curve of the water temperature sensor is not linear, so by fitting extra fixed  resistance distorts this curve. An additional 200 ohm resistor will make the engine temperature drop by an apparent (and approximate) 20 degrees at 80’, or only  about 10' degrees at 60’. The effect of  this is to fool the ECU into thinking more fuel is needed as the engine is colder than it is, but much the same as changing the fuel pressure, but  lambda control will have the same effect, pulling the mixture back to 14.5: until it goes open loop. Personally I prefer to alter the fuel pressure, as the results are more linear. 

 Air flow meter overview (Electrically speaking)

This is frequently seen as restrictive for best performance on the Rover V8, due to the size of the restrictive area that forces air into the hot wire chamber, an the turbulence generated by it. This may be true on the 4.5 - 5 litre engines, but on a mildly tuned 3.9 its not a major problem, up to around 280 bhp. It also has the great advantage of measuring the exact and true air flow used by the engine, and can compensate for engine modifications (within limits), unlike most ECU's that would require a remap. It has only one adjustment, a hex head bolt that sets a DC output voltage from the sensor that sets the idle carbon monoxide level (or mixture setting). Without lambda feed back this setting will significantly alter the mixture to the point of the engine running very lean or the other way, carbon fouling the plugs and it can cause low speed running problems if its too far out. With the catalyst fuel map, this setting has no effect as the ECU will trim the mixture to keep it within the correct range.  The setting of the DC voltage has no effect on the overall voltages produced by the air flow meter output for any given air flow. 

There are various generations of airflow meter, but in real terms very little difference and pretty much inter changeable. The Hitachi / Lucas 3AM and 5AM are plug compatible, even though the number of connector pins is different. The extra pin is simply not connected.

Testing the A.F.M .

With thanks to Mark Adams of TornadoSystems.com 

Note that you can drive without an airflow meter in case of emergency (i.e. airflow meter disconnected), because the system will drop into a default (limp-home) mode based on throttle opening.

Most airflow meter faults will cause the engine to run excessively rich. However if the airflow meter remains connected whilst defective then the vehicle will probably not run. In most cases the output from a defective airflow meter will be in the range 2.0-2.5 Volts, which is a viable value. This represents a moderate load and will cause heavy over-fuelling without setting a fault code.

Testing is performed in the following manner. Peel back the rubber boot on the airflow meter connector and leave it plugged in to the airflow meter. Set up the digital multimeter to read voltage. Insert the negative probe into the Red/Black wire (sensor ground), and the positive into the Blue/Green wire (Airflow signal).

Turn on the ignition, but do not start the engine. The meter should immediately indicate a reading of approximately 0.3-0.34 Volts. Most defective airflow meters will overshoot to 0.8 Volts or higher, and take at least 2 seconds to come down to the correct voltage.

Now start the engine, and the reading should rise to 1.6 Volts (3.5 Litre engine) to 1.75 Volts (5.0 Litre engine).

The next test is full load, and as with the fuel pressure test it will require use of a rolling road or a steep hill in the same manner. Under full load the voltage should rise to 4.45 Volts (3.5 Litre engine) to 4.95 Volts (5.0 Litre engine).

On this injection system, the idle CO mixture adjuster is provided on the airflow meter. It is located in a boss on the top of the airflow meter, pointing towards the engine. Leaving the multimeter negative probe in the Red/Black wire, move the positive probe to the Blue/Red wire.

Now turn on the ignition but do not start the engine. Observe the voltage. The normal adjustment range is between 0.0 and 3.6 Volts, with the higher Voltages producing higher idle CO values. There are approximately 20 turns of the adjuster screw to cover the entire range.

Annoyingly, the adjustment may be clockwise or anticlockwise to increase the value, and this varies from meter to meter! For this reason it is always preferable to have the multimeter connected in this manner when adjusting idle CO, so that you see can something is actually happening.

Typical Voltages that would be found at this point are between 0.9 to 1.4 Volts for non-catalyst cars. This Voltage is always factory pre-set to 1.8 Volts for catalyst vehicles. A value near to 3.5 Volts will generally produce an idle CO value of 9-10%. These Voltages may be used as safe initial values particularly if no CO measuring equipment is available.

Idle stabilising system and its problems.

 The stepper motor is always the first culprit for unstable tick over, but due to a very crude "pulse and wait" system used by the ECU to stablise the tick over, other factors like wrong CO settings, air leaks and wrong timing can cause the the engine revs to rise and fall as if the stepper motor was sticking. If you have cleaned the stepper motor shaft (as below) then look else where before replacing the stepper motor

Servicing the stepper motor.

I have now perfected a way of getting them to bits to clean, without damage:

1) Get the car warm 

2) Remove the stepper from the plenum and reconnect it.  Block the hole where the stepper was with some sort of bung or strong tape  to prevent air ingress.

3) Hold the stepper in your hand, so you can catch the central cone and spring when it comes out.

4) Get someone to start the car. Without the stepper in place it will rev at around 2k- 2.5k (This is noisey!!), but the ECU will try to slow it down by powering the stepper motor and pushing cone outwards. It will do this in a series of pulses, a few seconds apart, until the cone and spring drop out. (assuming they are not completely stuck). It may take a bit of manual help to get the cone to fully release. Now stop the engine and let it cool.

 6) Clean all the muck of the shaft, and lightly lubricate the screw thread inside the motor. 

7) To reassemble, wind the shaft back into the motor so far, but align the slot in the shaft with the plastic keyway as best you can by rotating the shaft in its screw thread 

8)With the engine cool, then reconnect the stepper motor. Power cycle the ignition on and off, (don't start) and the ECU will try to pull the shaft back in each time the ignition is turned OFF. This is the slightly tricky bit, as if the keyway and shaft are not quite aligned it wont pull home, so you may need a few attempts. 

Other factors in idle control.

With erratic idle, the stepper motor is normally the first suspect, as they get fouled up with carbon, and may need a clean. If this does not cure the fault the reasons can be many fold, but some basic trouble shooting steps can help isolate the fault. Firstly the reason the engine rpm cycles up and down is due to the crude system used to control the stepper motor. The ECU will try and hold the tick over at around 800-900 rpm when the car is stationary, and if the rpm is too high, if simply fires a pre programmed series of pulses at the stepper motor to reduce the airflow into the plenum chamber. The ECU then waits for several seconds for the engine to respond, and then applies a further burst of pulses should the RPM still be wrong. In an ideal world the drop in engine RPM in these waiting seconds should be uniform and controlled. So as an example the engine is running at 1200 rpm, and after the stepper pulses are applied, it should drop to say 900 rpm in say 3 seconds, and no further correction is required. This works well enough if everything is spot on with everything in the fuel injection system, air control, and ignition, but all goes very wrong if ANY parameter is wrong.

 The result of something being wrong is the RPM now drops much faster than the ECU program expects, so lets say the engine drops to near stall at 500 rpm. After the wait time the ECU rechecks the tick over and finds its too low, so winds the steeper motor back to let in more air to increase the RPM again. This may now go to high, so again the ECU tries to correct it back down and the cycle repeats itself.  Typically the engine can cycle between around 500 rpm up to 2000 rpm.

 To  get a perfect tick over all the following conditions have to be met:

 1)      Correct air fuel ratio.

 The ECU will get its total airflow reading from air through the AFM, and supply a relevant amount of fuel for this airflow. Problems occur if there is any air leaks anywhere in the plenum feed pipe, plenum chamber seal, stepper motor housing, stepper motor itself,  vacuum pipe to fuel regulator,  vacuum pipe to distributor, or a split diaphragm in the distributor advance mechanism. Everything is very sensitive, as the actual airflow through the AFM is very low at this point, and any air leaks will significantly reduce this reading so the ECU reduces the amount of fuel to the injectors. The engine then leans out and the RPM drop is rapid. This is  combined with a high vacuum in the plenum chamber as the throttle plate is shut, so any small leaks in this area have a much greater effect due to the greater pressure difference.  

To compound the issue, if the ECU is running the catalyst map and lambda feed back, it will then try to correct the lean mixture at tick over (although this takes over 15 seconds of running), but this correction is now distorted by the extra air getting in. basic fueling is now wrong, dependent on how bad and where the leak is.  

On the non catalyst fuel map, the CO setting on the side of the AFM has a huge effect on the tick over mixture. So if during normal running the mixture is near correct, once the engine drops to the tick over range the mixture can change significantly, so again the RPM can drop faster then expected. It can be  impossible to stablise an over rich mixture, as variable amounts of un burnt fuel remain in the inlet system that alters the burn properties , that will only clear once the throttle is opened, so the idle conditions constantly change.

 Another area of possible fueling issues can be the fuel pressure. The pressure regulator holds fuel pressure at around 37psi above the pressure in the inlet manifold. This vacuum level is fed to a control diaphragm in the pressure regulator  through a rubber pipe from the plenum chamber. If you measure the actual fuel line pressure at tick over, it may drop to say 20 psi when the throttle is closed, but will rise rapidly as you snap open the throttle. If anything in this system fails and the fuel pressure is not accurately controlled the mixture will alter from its correct values, changing the fueling.

2) Perfect ignition /burn.

Some what statement of the obvious, but any poor performance of any ignition component, ( That is HT leads, distributor cap, rotor arm, ignition amp coil, plugs,  etc)  could cause the engine to misfire at tick over, so the engine looses power, and the revs drop too rapidly. As the ECU applies more air and fuel, even if the spark is weak, the mixture may now burn fully, so the engines power suddenly picks up and the RPM shoots up too high. The ECU then overcompensates once more and the idle speed cycles.

 3)      Ignition timing.

Although a less likely cause if the engine is running OK under load, if the distributor advance curve is not functioning correctly, (distributor bob weights , springs or shaft wear), scattered timing at  tick over will cause the RPM to vary. As can be seen from the above, this is a complex control system, relying on part mechanical, and part ECU control, all working in close “harmony”, until something drops out of specification. 

 Practical steps.

 Firstly look for air leaks. Assuming none are found then try and get the engine to tick over without stepper motor control.

This is not a permanent fix, but if the engine ticks over nicely and does not stall then chances are the basic ignition system and mechanical systems are working OK. If the engine runs roughly or inconsistently, then firstly  look at the service items around the ignition system.  

If these steps fail, then a simple voltmeter across the lambda probe outputs will tell you if the correct fuel air ratio is being maintained . An analogue voltmeter is best, but a digital will do at a push, and you need to read a signal of 0 -2v DC across the black and white wires. Meter probes can be forced into the rear of the Lambda connector. With the engine running the voltage should constantly switch between about .3volts  and .7 volts at about 1/2 second intervals , and you may see the odd peak of 1.2 volts once the probes are hot (after 20 -30 seconds). If the voltages do not switch constantly, either the Lambda probe has failed, or the fueling is a long way out, exceeding the lambdas monitoring range, so investigation for air leaks or fuel pressure problems need to take place. Another test is to reset the ECU, and monitor the voltages straight after a reset. The voltages are likely to stay static for a short period (high or low) until the ECU readjusts the basic fueling, and brings it back under lambda control. This should take no more than 30 seconds at most on a warm engine, and proves the lambda feedback is working.  If the ECU is running outside these conditions, the fault codes should be generated and stored within the ECU.

A fairly brutal way of removing the ECU's ability to control its mixture at tick over, is to revert to the non catalyst fuel map by changing the under dash board resistor, and then set the mixture by hand with the setting screw on the side of the AFM. Although it can be set with DC voltages, it is far better to set it with a gas analyser to get it spot on. Alternatively the output from the lambda probes can be measured, and the CO trim value altered until you reach the point the probe voltage switches from high to low or vice versa. This method may damage the catalysts however if the engine over fuels, and is a test at best. If the engine runs perfectly like this, then its likely the ignition system is OK. By manually setting the CO output however, you could end up compensating for fueling or air issues, so simply masking he fault.

Fuel Injection Fault Display (Note Never fitted to Ginetta's)

The Range Rover setup allows an ECU warning light to come on if the correct air fuel ratio cannot be maintained, but is not enabled on all ECU fuel maps ! .It appears that only after market and maps supplied to the American market have this useful function. Failure in one of the following sensors will cause the warning light to illuminate, and the ECU will drop into the "get you home " mode. This causes the ECU to default to a "fixed fueling" map, that can run without sensor inputs. This will limit the cars performance and tends to run a bit rich, until the fault can be cleared.

Failing sensors.

Airflow sensor.

Lambda sensor

Water temperature thermistor

Throttle potentiometer

 To get a full diagnostic a fuel injection fault display has to be plugged into a socket in the wiring loom above the ECU that provides two-digit diagnostic codes.

.

These were available as a hand held unit from Lucas for around £125 but are now obsolete.  There is an aftermarket unit available.

http://www.bespokeintelligentsystems.co.uk

Alternatively the Lucas units  are for sale in the USA (like www.Ebay.com) as all US Range Rovers are fitted with them as standard on the second hand market. Part number is PRC17EM or Range Rover OBD.

There is normally a socket in the wiring loom near the fuel pump relay that the fault code reader plugs into. There are some latter looms however that the wiring and socket has not been included at all so extra wiring to the ECU connector is needed as follows:

ECU pins

Pin 2     Orange / black on the ECU   going to Orange black on the reader.12 volt supply to reader
Pin 30     Pink- to pink on the reader. Fault code data
Pin 31     Black green to black green on reader -- request input
Pin 38     Brown pink to brown pink on reader- Fault code data

  For cars without Lambda sensors the fault codes are limited, but can at least point you in the correct direction. Modern Range Rover test equipment can still read the older units. To reset the unit after a fault has occurred, disconnect it from the battery for some minutes.

If multiple faults exist, the display shows the one that the ECU thinks is highest priority. Higher priority faults need to be "cleared" before lower priority faults will be displayed. A "blank" (dark) display usually indicates there are no faults.

Use this procedure to clear faults:
  1. Switch "on" the ignition.
  2. Disconnect the serial link mating plug, wait five seconds, and reconnect.
  3. Switch "off" the ignition, and wait several seconds.
  4. Switch "on" the ignition. The display should now reset.
Note: It should either show a lower priority fault code or appear dark.


Note: Fault code "02" will appear after a disconnected ECU is reconnected. Simply switch on the ignition to clear the display.

 Fault codes

 02 ECU Supply has been disconnected

 Normal code to get after ECU has been reset.  May indicate supply problems if ECU has not been reset recently.  Code will be cleared if the ignition is turned on for 30 seconds or so 

03 Data corrupted. Corrupted data in ECU.  Reset ECU and test drive again.

 12 Airflow meter out of range Possible faulty airflow meter or connection/wiring fault.

 14 Coolant thermistor out of range. Possible faulty coolant thermistor or connection/wiring fault.

 15 Fuel thermistor out of range. Possible faulty coolant thermistor or connection/wiring fault.

 17 Throttle pot out of range. Possible faulty throttle potentiometer / incorrect setting or connection/wiring fault.

 18 Throttle pot output too high with low airflow. Major air leak between airflow meter and intake plenum.  Possible faulty airflow meter or throttle potentiometer or wiring/connections to either. 

 19 Throttle sensor output too low with high airflow. Possible faulty airflow meter or throttle potentiometer or wiring/connections to either.

 21 Tune resistor out of range Tune resistor has become disconnected / damaged.

 23 Low fuel pressure. Possible faulty fuel pump/ blocked filter or faulty fuel pressure regulator. VALID FOR CAT CARS ONLY.

 25 Misfire at full load. A misfire has been detected while engine under heavy throttle and during high airflow meter readings.  Check coil/plugs/leads/ distributor and ignition module.  Code means that lambda sensors have detected a rich condition under load.  Also see codes 40 and 50. VALID FOR CAT CARS ONLY.

 28 Air leak. Air leak around intake plenum.  Check all hoses/ injector seals etc.

29 Checksum error. ECU fault.  Try resetting ECU and test driving again.  If fault code returns ECU may be faulty.  NOTE – any other codes generated should be ignored.

 34 Fuelling fault in nearside injector bank. For cylinders 1-3-5-7 Possible injector, lambda sensor fault or wiring / connection fault to either.   Blocked injector(s) or air leak at injector /inlet manifold.  VALID FOR CAT CARS ONLY.

 36 Fuelling fault in offside injector bank.  For cylinders 2-4-6-8. Possible injector, lambda sensor fault or wiring / connection fault to either.   Blocked injector(s) or air leak at injector /inlet manifold.  VALID FOR CAT CARS ONLY. 

40 Misfire on nearside bank. Misfire on 1-3-5-7 cylinders only.  Nearside lambda sensor has detected a fault (or fault in lambda sensor circuit). VALID FOR CAT CARS ONLY

44 Offside lambda sensor out of range. Possible faulty lambda sensor or wiring/connection fault.  See section 5 below for check. If in conjunction with code 45 then suspect lambda heater circuit.  VALID FOR CAT CARS ONLY 

45 Nearside lambda sensor out of range. Possible faulty lambda sensor or wiring/connection fault.  See section 5 below for check.  If in conjunction with code 44 then suspect lambda heater circuit.  VALID FOR CAT CARS ONLY 

48 Stepper motor fully open below 500rpm or fully closed above 750rpm. Possible faulty stepper motor or wiring connections.  Stepper motor needs cleaning / is jammed.  Incorrectly set idle adjustment screw.  Incorrectly set throttle butterfly.  Misfire or rough running because of other faults.      

 50 Misfire on offside bank Misfire on 2-4-6-8 cylinders only.  Offside lambda sensor has detected a fault (or fault in lambda sensor circuit). VALID FOR CAT CARS ONLY

 58 ECU cannot distinguish between codes 23 and 28. Fault maybe either due to code 23 or 28. 

59 Same as 58 or fuel thermistor out of range. Documentation seems to vary here.  Code either means same as 58, or faulty fuel thermistor or wiring / connector fault.

 68 Road speed sensor too low at medium rpm and high airflow. Possible speed sensor or wiring/ connection fault.

 88 Power-up check / purge valve fault. Sometimes shown on power up (CAT and NON-CAT CARS ) or could also indicate purge valve fault with carbon canister system.  (CAT CARS ONLY)

ECU urban myths.

Over the years various myths have risen up about this ECU, and not all of them are true.

Examples are:

You need to drive for 20 minutes to stablise the mixture after a ECU reset.

Untrue- it takes place in about 15 seconds at tick over.

You need to reset the stepper motor if you remove it.

Untrue. The stepper is reset every time the ignition is cycled off, but sending 200 pulses into a 180 step motor, so it will always reach "home".

You need an ECU reset to clear fault codes and reset the mixture.

True - The ECU contains no volatile memory, so a 30 second power off (as in unplug it or disconnect the battery ) will remove any fault codes, and any stored fuel base settings. These will be re established  once then engine is running at tick over.

 The fuel map “stops” at 5400 rpm. 

On the standard Range Rover fuel map does not increase the amount of fuel above 5400 rpm, but simply carries on fueling as if the engine was still at 5400 rpm. This can be overcome with aftermarket maps, where one of the alterable parameters in the fuel map, allows the entire fuel map to be moved up or down the RPM  Range, so the upper limits can be moved to cover the higher RPM range.