Alfa Boxer Tuning | Cylinder Head | Coolant System | Fuel System | Ignition System | Engine Block | Exhaust | FAQ's | How Do I... | Photo Page | Additional Photos | Ancillaries | Favorite Links and Contact Page
Ignition System
The standard setup


The standard ignition system used by the Alfa boxer engine in the twin carburettor Sprint, AlfaSud and 33 is an electronic Bosch unit, which uses a part mechanical distributor with a hall effect reluctor trigger, and a conventional coil.

Inputs to it are:     +12 volts switched supply from ignition switch

                                trigger sensing connection from distributor

The coil also has +12 volts supplied to it, and a low tension connection from the electronic ignition box. In addition, it outputs a signal to the Tachometer on the drivers instrument panel. This wire is green with a yellow tracer on it, and is common with the low tension connection.

A standard HT king lead then feeds the centre of the conventional distributor cap, and a conventional rotor arm then feeds the spark voltage to the individual HT leads. The rotor arm is fitted with a spring loaded moveable contact, which acts as a centrifugally operated rev limiter.

The advance curve of the distributor is pre-set, and only vacuum advance adjustment is available.

There is not much that can be done with the standard ignition system, and for most intents and purposes it functions very well. As with any ignition system, the key is to ensure that all the components are kept in good condition. I have found that this setup is susceptible to deterioration of the leads, and to spark deposits on the contacts inside the distributor cap. Beyond this, it is a trouble free system.

It may be necessary, however, to consider the use of harder plugs if you have seriously uprated your engine. The best way to determine the need for this is to see what performs best on a rolling road session.




Alternative ignition systems


In order to best deal with tuning the ignition system, an understanding of what it needs to do is useful.

Why do we need the spark? Quite obviously, unlike a diesel, the petrol and air mix in the cylinder needs something to initiate combustion. But just igniting the mix is not sufficient. If you ignite the spark too early, the burning mix in the cylinder tries to stop the piston on the compression stroke, rather than pushing the piston dow on its power stroke. If it is far too early, it will ignite the mixture as it is coming into the cylinder with the valve open. If it is a little too late, the burning mix is chasing the piston down the bore, rather than pushing it, again wasting power. This leads to overheating and burnt valves.

So why not just have the distributor fire at a set point? Well, that would work perfectly if the engine was kept at a constant speed and constant throttle setting. The trouble is, the mix will burn at a certain rate, and as the engine speeds up,  the mix has to either burn faster, or more practically, you need to start burning it earlier. That forms the beginning of the advance curve. So all you have to do is have some centrifugal weights that change the advance as the engine speeds up, right?

Well, no, not quite. If the cylinder is full of a good mix of air and fuel, like when the throttle is wide open, it will burn nice and quickly when ignited. If the cylinder only has a little air and fuel, like when the throttle is closed, the mix burns slowly. So a means is needed to advance the ignition even more when the throttle is closed and cylinder filling is not so good. This can be compensated by something that senses when the throttle is closed. The vacuum advance capsule does this, as vacuum exists in the inlet manifold when the throttle is closed, but reduces as the throttle is opened. In addition, the vacuum acts a a free mechanical force to pull the distributor mechanism to more advance against a spring.

So, the standard distributor comes with a centrifugal advance, and an additional vacuum advance mechanism. By repeated testing, and using different springs and weights, a distributor is modified to produce a curve that similar to the ideal that an engine wants to see. The manufacturers are then able to be certain that for a given engine and specification,  their curve will be satisfactory for all engines produced to that specification.

The mechanical curve is fairly linear, as are the advance requirements charecteristics of production standard engines. Unfortunately, lumpy cams and improved breathing characteristics change what advance is required.

The trouble is that speed and load are not the only factors influencing the ignition timing requirements. A weak mixture burns slower. A higher compression ratio makes the mix burn faster, and vice versa. Higher temperatures also make the mix burn faster. Certain conditions will cause pre-ignition.

Look at what a lumpy cam does. Due to the increased overlap, at low speeds, some of the air and fuel that goes into the cylinder goes out of the exhaust valve, some of the exhaust gas pushes the intake mix back up the inlet tract, and if running carburettors the low speed fuel mix can be considerably off the ideal. The result of this is that the effective compression ratio is very low, you may have a poor fuel to air ratio, and the quality of the charge in the cylinder is contaminated by exhaust gas, which does not want to burn. You therefore need considerably more advance so that the mix has ample time to light off. Unfortunately, you have stepped on the throttle to accelerate, so you have reduced the vacuum advance. Poor running characteristics at low speed are therefore made even worse.

Why not just set the distributor to give more advance? Well, as your engine accelerates, and climbs onto the cam, suddenly everything starts working wonderfully. Remember that compression ration increase to make the cams work? That means that you need less advance. The valve overlap that previously was spoiling the intake charge is now helping to pull more air and fuel into the engine, increasing the volumetric efficiency of the cylinder and increasing the effective compression ratio even more. The carbs are on song now, so the mix is good quality. Everything is saying 'less advance'.

Combine the need for more advance at low speeds and less at high, and it can be seen that a steeper curve is needed, which can be altered by changing the spring rates inside the distributor. In the good old days before electronics, these requirements where met by companies like Aldon, who modified distributors to achieve this. This worked to a degree. The trouble is that our lumpy cams and mods have not finished giving us trouble yet.

All of the mechanical advance mechanisms assume one thing; a more or less linear change in ignition requirements. Unfortunately, our tuned engine is nothing but. Different engine speeds make different things work. At 2500 rpm, the engine may be off cam, with lots of pulse and resonant effects doing their damnedest to spoil the quality and quantity of the intake charge. You need lots of advance. At 3000 rpm, the engine comes on cam, and all those pulse effects are stuffing tons of air and fuel into the cylinder. You need less advance. At 3500, a pulse length, or resonant effect suddenly starts working against the flow (you would see this as a dip, or flat spot in the power curve - very common on modified engines). You suddenly need more advance again.  In addition is the possibility for pre-ignition at high load, which requires backing off the advance.There is no easy way for a mechanical distributor to do this.

Depending on the budget available, or the need, it is possible to introduce a completely different setup that completely gets around this. This is a distributorless fully electronic solid state system.

The advantage of this system is that you are always going to have the spark occurring at exactly the moment that the engine wants to see it, regardless of the type of cam, exhaust, fuel system etc. A mechanical distributor follows a pre-set curve which is determined by springs and centrifugal weights, with correction for low load by means of a vecuum connection. It is crude, but for a production engine works well enough because the manufacturer tailored the particular weights and spring rates for a standard of engine. The advance curve is a compromise, but a closely matched one. For a modified engine though, that compromised curve is not going to be as well matched, simply because you have altered the requirements of the engine.

A fully electronic system can be set to give the spark at exactly the right moment whatever the engine whatever the conditions.

This setup comprises the following components:

Twin double ended coil unit

Electronic Control Unit

Water temperature sensor
Air temperature sensor
Throttle position sensor
Crank angle / position sensor
Crank speed sensor

The coil unit is a solid state unit comprising two parallel coils, each with two outputs. When triggered, each coil outputs two sparks. By connecting these coils to plugs in the cylinders which are at top dead centre at the same time, the cylinder on the firing stroke is ignited, and the cylinder on the exhaust stroke simply wastes the spark generated. This is the wasted spark principle.

As a result of this setup, each coil must be individually triggered, and the timing of this is controlled by an electronic 'brain'.

The 'brain' is a mappable elecronic control unit (such as the Emerald M3D ECU, as I used on my 16v), which can also control the injection. This takes inputs from the sensors listed above. The crank sensor enables the ECU to determine at which angle the crankshaft is at, and at what speed. Using these inputs, the ECU is then able to look at RPM, throttle demand, water temperature and air temperature to calculate exactly what the correct advance is. It does this by looking at a three dimensional 'map', which plots advance against throttle load and RPM. Additional correction factors for temperature and cold start requirements are then superimposed on these values.

Once the ECU has determined when the spark is required, it then triggers the appropriate coil. The coils are connected directly to the spark plugs, so no HT king lead or distributor is required. It is a simple system to physically install.

The 'map' referred to earlier is one which is determined by actually running the engine on a dynamometer, either an engine dyno, or more frequently a rolling road. During this, the engine is run at different speeds and loads, and the exact ignition advance is set by changing the timing until the engine is giving the best power. The system can also be changed to sort out dips in the power curve, prevent pre-ignition, pulse tuning effects, off cam running etc. In addition, it can be used to auto stabilise idle running, and act as a soft-cut and hard cut rev limiter.

Other benefits of this are:

it can easily be reprogrammed to suit further mods to your engine, or transferred to another engine

There are no moving parts to wear, or lose adjustment



I decided to fit a fully solid state electronic ignition system to my 16v.


Physically installing the distributorless system

The assumption is that you are fitting this equipment to a 16v engine. For 8v engines please refer to footnotes.

Please refer to the fuel system section for installation of the ECU.

What you will require are

A suitable Electronic Control Unit (ECU). I chose the Emerald M3D unit

m3d ecu.jpg (26345 bytes)

The Emerald M3D ECU

A dual double-ended coil ignition pack, probably from a Mondeo or Escort zetec engine (mine was a 1800 MkV escort 16v unit). REMEMBER to get a piece of the wiring loom complete with connector!! Ford does not sell this separately.

Zetec coil unit annotated.jpg (38310 bytes)
the Zetec coil unit

A three way male/female electrical connector

Either 4 off No1 cylinder plug leads from a 1991 1800 Escort, or a pack of plug lead adapters (a photo of these is below)

zetec coil plug lead adapters.jpg (23425 bytes)

the Zetec coil adapters required for the HT leads

You will also need to supply a switched +12 volt supply and some wire to connect the ECU to the coil.

The zetec double ended coil is fitted so that at the same time as cylinder no 1 is firing on its firing stroke, the spark plug for cylinder no 2 will be firing on its exhaust stroke, and ditto for cylinders 3 and 4. Before anyone writes to me to tell me that it should be 1 and 4, and 2 and 3, remember that the boxer has an unusual firing order of 1-3-2-4, rather than the normal 1-3-4-2.

The +12v line is connected to the centre wire of the coil connection. The ECU pin for the no.1 cylinder is connected to the wire on the side of the coil connector marked 1 and 4, and the remaining ignition coil ECU pin is connected to the remaining coil wire.

A separate feed goes to the tachometer, but you can connect the green/yellow tacho wire from the existing system direct to the ECU tacho output. It works fine with the standard tacho fitted to the car.

The ECU will automatically sense the crank position from the crank sensor, so you do not have to worry about setting the timing in the conventional sense. Once the car is running, it will be adjusted using the PC software.

Oh, by the way,  when you remove the distributor, remember to blank off the hole!


That's the theory, now does it work....?


After all this is the interesting task of trying to get the car to at least fire up.

Once all this injection and distributor-less ignition hardware had been fitted, I had to try to make it all work together. I attempted to use the instructions that come with the EMS to get the engine running.

The EMS came programmed with a basic, safe fuel and ignition map from a similar sort of engine, perhaps from a 1.6 or 1.8 ford Zetec. The engine did not want to run using this, point blank. After many hours of trying, I decided to take a step back and simplify the system in order to attempt to diagnose why it wouldn't run.

I had previously stated that I might run into problems, and that as a contingency I could stick on the carbs and the distributor. Well, that proved to be my saviour, or at least some of it did. I could not very easily stick the carbs on because it would mean a change of fuel pump, plus devising a means of attaching the throttle pot to the carbs. The distributor, however, could simply be bolted back into place, and the connections remade by re-attaching the spade connectors.

This I did, and with the help of a can of 'Easy Start' and the lap top, we tried again. Squirting the Easy Start in while cranking showed that the engine wanted to run. By richening the mixture VERY considerably, it finally burst into life. It seems that the fuel required by the engine was considerably more than he engine was getting using this map and fuel pressure.

However, once I had the fuel mapping at a point where the engine was happier, I attempted to refit the distributorless system. The engine would now run, but was a pig to start, and it was obvious that my crank sensor was not helping. It seems that the output signal from the sensor was not very strong at low speeds (below 1000 rpm), and the ECU was not able to consistently read from it. As a result, the speed site information was switching between 0 and site 2, with the result that the injection and ignition was being switched on and off. This results in an erratic idle and poor running characteristics at low speed.

The ECU sensitivity can be increased to compensate for this, but will have to be done at Emerald in London, at the same time as the rolling road mapping is performed. As an interim measure, the distributor ignition has been refitted and the car currently runs using this.




Just a few comments regarding general ignition related issues.

Firstly, spark plugs; the ones fitted are the Golden Lodge multiple electrode type. I have used both these and standard electrode plugs on Alfa Boxer engines with absolutely no apparent difference in the way the engine runs. Any comments or findings from anyone who has some experience of different plugs in 16v engines would be appreciated.

Secondly, the plugs on a 16v are a bastard to get at. No ifs, no buts, they are buried well into the heads, and the chassis rail stops you getting full access to them. They are impossible to see with the engine in the car without the use of a mirror. Getting the plug leads on and off them is also a thankless task. The rears are even worse than the fronts, and the fronts are pigs. My recommendation is to change or check the plugs at every opportunity that the engine is out or dropped down.

You will need to devise some fiendishly clever bit of tooling to get them out, just in case you need to. I used a Halfords plug socket, plus a 3 inch long extension bar and a slimline ratchet or tommy bar. This was satisfactory, but I will be making a purpose built tool as access to the rear plugs is difficult even with this.

If you are fitting all this kit to an 8v engine, you will need to install either a 36 - 1 or a 60 - 2 toothed wheel for a crankshaft trigger sensor.

You will have to fit a water temperature sensor, although this is easy as both inlet manifolds are tapped to accept sensors. Just fit a 16v sensor in the manifold with the blanking plug.

If keeping carburettors, a throttle potentiometer will be required. Fitting kits exist, and are easily available.