This entry – very short is about how the stabilization of the idle for petrol engines has evolved. I will mention exactly the smoothing out of the cylinder efficiency.
From the manufacturer, as usual, there is zero information; accordingly, the numeration is my personal interpretation. All information described below is gathered by observing the performance of the DME. For example, I will mention the 6-cylinder petrol engines, but identical principles are used for the engines of the same generation with other configurations.
Generation 0. A long time ago, for example, in M20 (both Jetronic and first Motronic), the performance of the cylinders was not smoothed out. If some injector was “more active”, the appropriate cylinder was performing more actively. The engine was slightly/permanently vibrating, but – there was no solution. If only – replacement of the injectors or trying to sort out the injectors with similar activity.
Generation 1. Starting with the popular MS41/42/43 management systems, the smoothing of the idle started. By measuring the combustion cycle of each cylinder (using the crankshaft sensor), the performance of the cylinders was compared. During a longer time (system reaction time – minute/two), DME measured the average performance of each cylinder and slowly, step by step, corrected the injection time (portion of the injected fuel) in each cylinder.
Generation 2. In N series engines (starting with N52), the smoothing of the idle becomes more complicated. The principle of the energetical model is introduced. Valvetronic is introduced (it significantly complicates the management of the idle). In addition to the “simple” smoothing of the efficiency, DME now predicts the mechanical load, planned for the future when the other obstacles change (load of the AC, alternator, power steering, etc). The smoothing of the cylinder performance is still happening, comparing the average mechanical load of each cylinder during a longer time and slowly compensating for noticed differences from the ideal.
Generation 3. This generation I mark in N53 series engines. It is interesting that in the E series, these engines use significantly different algorithms for injections in Homogenous mode and Stratified charge. For Homogenous mode, still the previous generation slow-performing algorithms are used. Instead, in Stratified charge mode, DME reacts much faster. Typically, during several (2 .. 5) seconds, the mechanical efficiency of all cylinders is smoothed. In addition, the combustion cycle of each cylinder is separated into several segments. Evaluating the efficiency of each cylinder in each segment, the amount of injected fuel in each of several injection portions is modified, and the ignition moment for each cylinder individually is adapted. In addition, in each of the modes (both Homogenous and Stratified charge), DME performs a range of specific injector tests (tests of cylinder chemical efficiency) – the fly-time of the injectors is measured, the leaking and the atomization quality too. All these tests are used to predict the performance of the injectors under variable regime conditions. DME plans the performance of the injectors “forward” – the idle should stay perfectly smooth even when the AC is turned on/turned off when the steering wheel is turned or the load of the alternator changes. The principle of cylinder smoothing stays the same – DME watches the average mechanical efficiency of each cylinder for a longer time and slowly brings it closer to ideal (relatively slowly).
Generation 3A. For N53 series engines in F series and next N series DI engines, relatively fast working algorithms are used – similar to ones previously used in E series N53 series engines in Stratified charge mode. DME relatively quickly (reaction time – several seconds) reacts to the differences from the ideal of each cylinder’s average mechanical energy and compensates for these differences step by step.
Generation 4. A new quality jump can be seen in B series engines. In B series engines, relatively fast-working average mechanical efficiency smoothing methods are used – similar to the ones used in the last N series DI injections. But there is one more significant addition. New DMEs do not limit themselves to the smoothing of the average mechanical efficiency (measuring that during quite a long time – in several seconds). New DMEs react immediately, already in the actual combustion cycle! How does it happen? In addition to the existing average efficiency smoothing method (its job is to prevent significant and permanent/stable differences between the mechanical efficiency of cylinders), a new additional method is used. If DME notices that some cylinder has been a bit “lazy” during the existing combustion cycle, the next cylinder (by firing order) is “activated” to immediately compensate for the “laziness” of the previous cylinder. In such a way, the short-term/undiagnosable fluctuations of the engine performance are masked. What is necessary to realize this new additional method?
a. DME needs a very powerful management processor;
b. the average fuel mixture should be kept quite lean, which means Lambda is around 1.01;
c. additional adaptation maps for the “vector” reaction of each injector should be created.
Short comments regarding each nuance:
a. from DME, an immediate reaction is required – this time, we don’t talk of response for several seconds. The injection corrections should be calculated and executed within thousandths of a second! In addition, these calculations have to be performed parallel to other “basic tasks”, which are not canceled;
b. if the average fuel mixture is slightly lean (Lambda 1.01) – each cylinder “throws” in the exhaust around 1% of air. For 6 cylinder engine, it means – if it’s necessary, the fuel mixture of some (one) cylinder can be enriched even for +5%, but still – the CO catalytic converter will correctly burn all leftover the fuel;
c. I named this additional method of smoothing the idle – the “vector” method. I chose such a name because graphically, this correction looks like a vector (direction) to the zero (ideal) value of the mechanical efficiency. For the reaction of each cylinder to be correct, DME creates the reaction adaptation maps. DME observes how the exact cylinder handled the correction of these vectors in previous times and adapts the “power” of the vector correction. If previously the cylinder solved the problem incompletely, the power of the vector (value of the adaptation map) is increased. If the cylinder happens to have overcompensation – the value of the vector power coefficient is reduced.
Note: this additional vector method is used only in cases when DME works in full functionality mode (DME and connected systems have no problems, and/or any error messages are recorded). The efficiency of this method (amount of the correction) is limited and works efficiently only if the performance of the engine is stable enough already without using this method.
What is the gain from this additional method?
Previously (for N series DI engines), a short-term mechanical efficiency difference of +/- 5 .. 8% from the ideal was considered acceptable. Yes, the engine was working smoothly, but specifically, capricious users could feel that it was working. Till the B series, it was expected – there is a 300+hp powerful aggregate! Its performance SHOULD be felt!
B series engines can maintain the mechanical efficiency of cylinders with precision of +/-1 .. 3%. The engine runs so smoothly that it can not even be felt inside the car!
Here is how the correctly performing idle of B58 looks:
Differences in the cylinder performance from the ideal is till +/-1 .. 2 units (Axys 2) or +/-1.2 .. 2.5%. Something previously unattainable.