In my opinion, one of the fundamentally underrated tasks in the performance of modern DME/DDE is the advanced management of the injector of each cylinder.
Usually, if we talk about piezo injectors, the service specialist knows that the injectors have encoding data. Though most of them think that these data are insignificant, say – the injector is an injector, just install it – and done. More advanced specialists know that when input these data, the corrections of the injector parameters from ideal are input. When I say that the injectors are additionally measured in all conditions of the engine performance, usually, no one believes me.
Ta the best case, I manage to convince someone that idle DME corrects the opening time of each injector. In the case of DME, there are several impulses: the first containing the compensation of the injector leaking, but the second – compensation of the injector delay. For ”normal” people this sounds like rocket science.
Even harder to persuade someone that the flow rate of injectors, leaking, and quality of the atomization are measured in all driving conditions of the engine. Because of such “schizophrenic” statements, I’ve even been banned from the forums of the industry “specialists”.
To convince the skeptics, I recorded the live data of the injectors.
Here, the sample video:
The first 3 lines are the opening time of one of the injection impulses of each injector.
Lines 4 to 6 – the flowrate coefficient of the injectors. As you see, in some modes, these flowrate coefficients are changing. Both the total flowrate “power” and data of each injector vs. others are changing!
As I’m already have mentioned in other blog entries, part of the MSD80 DME information regarding injector tests are available here: ../F5/Shift+F6/F3 and ../F5/Shift+F6/F5; I think now it’s time for skeptics to check these menus and my blog – in descriptions, what is shown in these menus.
In this blog entry, the most accent is made that this flow rate correction of the injectors is NOT a fixed value but a multidimensional map.
Here, the injector encoding data, input in the DME:
In this table: injector flowrate coding data input in the DME, and their difference from the average value.
Although the flowrate coding data allow inputting +/-17% difference from ideal, as we see – in this case, the injector data in laboratory conditions have differed for only +/-1.5%. In the lab, the injectors have worked perfectly!
Based on the data from the previously shown video, I made a straightforward Excell chart:
In four conditions of the engine performance mode, I recorded the injector live data; calculated the average flow rate and the relative difference of flow rate of each injector from the average.
Here, how the graphical presentation of these data looks like:
I believe – no one now has the slightest doubt: the flowrate data of the injectors are NOT fixed values! As you see, flowrate data are changing, depending on the required torque and RPM.
The largest changes are for injector No.5: from -11.8% to -2.3%. Here I have to remark that this injector is completely new (resource: around 3 .. 500km). But also other injectors are not so far away, for example, No.2: from 0.4% to 2.8%. Or No.4: from -4.1% to 3.4%. Also, injector No.4 is completely new! In addition, as we see, the data of No.4 and No.5 ARE NOT changing in tandem, but their changes are completely opposite! At the moment, when No.4 starts to get “lazy”, No.5 gets more “efficient” and vice versa.
And yet we see that in the flowrate data of the injectors are NO correlations with the encoded flowrate data. So – DME did not calculate these “new” flowrate data from the ones already input, but they actually were measured and applied.
To make the situation more “interesting”, I will mention that the injector flowrate adaptation maps contain the following dimensions:
a. required torque;
b. RPM of the engine;
c. injection mode: Stratified/Homogeneous lean/Homogeneous (only N43/N53);
d. engine temperature (5 segments).
And additionally, to each bank, there are Multiplicative type LTFT, which have the following dimensions:
a. required torque (5 segments);
b. temperature (5 segments);
c. impact of the EGR;
d. nonstandard modes: warming up CO catalytic converters; desulfation; regeneration (last 2: only for N43/N53), etc.
In the current example, you can see that the true performance of a completely brand new injector can differ from the input data (laboratory conditions) for even 5 .. 10%. What does it mean? If the injector is unmeasured or incorrectly measured, the fuel mixture of the current cylinder on the go will differ from the ideal. The difference will be exactly these 5 .. 10%. That, instead, means that these 5 .. 10% of fuel and air will be burned in CO catalytic converters. The temperature of the catalytic converters will be significantly higher than it is supposed to be, and they will be prematurely damaged. Yes, of course, the fuel consumption also will be increased, but in this case, the damage of the CO catalytic converter is the largest problem.
In the end, I would like to remark – for DME/DDE to measure the injectors and these measurements would be recognized as reliable (and would be recorded/confirmed in the DME/DDE memory in the injector adaptation maps), the engine should run in full functionality mode. What does it mean? DME/DDE has to receive confirmation that all systems of the engine work as they are supposed to. If at least one of them (it would be, for example, VANOS, or some Lambda probe, or MAF, or any other discrepancy od the data, for example – required and the true amount of air) does not work correctly, DME/DDE switches off this feature and switches to much simpler work modes. As a result, you will get a vibrating engine with a low-quality performance and swiftly damaged exhaust. Or, in the roughest case – even damaged piston group.
