D-Jetronic Fuel Injection

The Bosch D-Jetronic fuel injection system was used by VW in the Type III in 1968. This was the first electronic fuel injection system to be fitted to a production car. The Bosch Fuel Injection Systems 1970-1979 Shop Manual published by CARBOOKS, INC. gives the following list of cars using the D-Jet system:

I've also heard that 1975-1980 V12 Jaguars used a D-Jet system, with the ECU (Electronic Control Unit) built by Lucas, and using Bosch parts otherwise.

The following text is Bill Lewis' explanation of how the D-jet system works.

Many injection systems worked on this princple, and some still do, although Bosch has decided it is an inferior way of measuring the quantity of air in a cylinder. Just changing the tailpipe on a car can throw this system out of whack. Small vacuum leaks make it really difficult to tune at idle.

The D system works in an extremely clever way. Remember that this was built first in 1968 on the Type 3. Transistors were still expensive then. The main input to the system is this vacuum/pressure transducer. It's about the size of a baseball. It is a sealed chamber, with a port connected to the intake manifold. Inside it is a metal bellows like you'd find in a barometer. One end of the bellows is attached to the housing of this thing, and the other end is a prong about an inch long and about 1/8 inch in diameter. This prong is actually the core of a transformer that is also inside the housing. As the vacuum in the chamber changes, the bellows extends or shrinks which pushes or pulls the core in/out of the transformer windings. This changes the inductance of the windings. Everything about this gadget is designed to match the engine characteristics.

The ignition distributor has an extra set of two breaker points (two because the injectors are fired in pairs). When the points close, or maybe when they open, I forget exactly and I haven't looked at the schematic, it triggers this circuit called a blocking oscillator. When the points close, it turns on a flip flop which puts a voltage across one of the windings in that vacuum gadget. The current through the inductance of that winding rises with time. This creates an increasing magnetic field in the core. As long as that field is increasing, a voltage is induced on the secondary side of the transformer. After some time has gone by, the core saturates, the current in the primary stops rising, the magnetic field stops increasing, and the voltage on the secondary decays. When that happens, the flip flop switches back to the other state. When the flip flop is in the normal state the injector is off. In the active state, the injector is squirting. The length of time for this whole thing to happen is set by the position of the core withing the windings, and that by the vacuum. The whole thing is cleverly and carefully built so that the time for this to happen is also the correct injection time.

There are also cylinder head and/or intake air temperature sensors (thermistors) which are used to modify (lengthen) the time period to give a richer mixture during cold starting and warmup operation.

There is also a micro switch on the throttle to put it in a special idle mode, and sometimes one for full throttle. The is a wiper arm on the throttle that wipes across a zigzag pattern on a small circuit board to give a bunch of pulses when the throttle is moved a large amount. This is used in the same fashion as an accelerator pump in a carburetor.

In 1974, the next Bosch system, L-Jetronic, appeared on 1.8 litre 914 engines. It uses an airflow measuring device with a flap in a box rather than the vacuum system. This air flap box adds more restriction to the intake system and produces less horsepower as a result. It also uses integrated circuit technology and a mix of analog and digital technique. D is a totally analog system.

../Bill Lewis

So, as you can see from the above, the vacuum sensor itself forms a large part of the D-jet system. The 'computer' box is mostly an interface to the vacuum sensor and injectors, the only 'computation' that happens is for temperature compensation (probably just a variable threshold for the voltage on the vacuum sensor secondary, but I won't know that for sure until I completely reverse-engineer a unit).

Kjell Nelin wrote the definitive guide to diagnosing and setting up a D-jet system, which was hosted on http://www.914fan.net before it disappeared. Fortunately Jeff Bowlsby saved a copy.

What I would like to do is to map the curve of the D-Jet system, i.e. injection time v.s. manifold vacuum and rpm, also the temperature compensation curves for the two temperature sensors (engine temp and inlet air temp). The next step would be to build a digital FI box that incorporates these curves and uses all the stock hardware except for the manifold pressure sensor. I'm planning to use a more modern sensor, one that gives an analog voltage. This is all set to happen as soon as I get the time for it... :-).

• Jeff Bowlsby's 914 Technical Notebook.
• Julian Bunn's page.

Back to Wouter's page