In the first part of this story (here), I mentioned a strange crash that B58 of my car experienced. I found out that the misfires were the culprit. 

What was done about it?

a. the cleaning procedure of the spark plugs was performed. The quality of the idle improved; the popcorn was not so loud, but – such crash repeated itself;

b. the spark plugs were replaced by new ones (BMW). Crash repeated again;

c. I introduced snubbers for all ignition coils;

d. I found that the unstable performance of cylinder No.2 is the consequence of the ignition problems of the 6th cylinder. I replaced the ignition coils in cylinder No.2 (this ignition coil suffered due to misfires in this cylinder, caused by cylinder No.6) and No.6 (this cylinder/coil was the cause of the problem). Operation instability of these cylinders disappeared, but the crash – continued. 

It’s a very interesting defect. Everything is perfect for the cold and partially warm engine – even if the engine has been idle for a longer while – no problems. For a warm engine – no problem on the go or short idle. But, if the warmed-up engine is idle for a longer while (20 .. 30 minutes), be ashamed in front of the entire district. 

For the last time, these crashes were truly impressive – smelly smoke, loud “screeching” from the engine room, popcorn from the exhaust. 

After these performances, I had to conclude that in the DME error message memory, there are only error messages regarding the amount of the inlet air or its plausibility. Excellent!

My patience was done. I drove to a remote place, turned on ISTA Expert mode, and waited. After 20 minutes, the “show” started. 

I marked the cylinders – culprits with bold:

As we see, in the 227th-second cylinder, No.5 fails slightly (not a lot, only some -6 .. 8% from efficiency). Then, on the 231st second, a group of misfires of cylinder No.5 started. 

Only several cycles later, the misfires of cylinder No.2 started, too. What happened to this cylinder? Just now, the ignition coil was replaced! It looks like these are the consequences of idle stabilizing algorithms. My version: in the situation when the performance of some cylinder should be corrected strongly (and this is exactly the case) with the same “polarity,” the “further” (by firing order) cylinder should be compensated to avoid unstable idle (“waving” or so-called subharmonics of the transition process). So, if suddenly the performance of the 5th cylinder turned out to be significantly reduced, DME tried to reduce the efficiency of the 2nd cylinder, and all this led to misfires of this cylinder, too. 

After a second, the situation got better; misfires in these cylinders stopped. But still, there is no sign of a stable idle. What happened?

If DME correctly identifies misfires (for example, MDS80/85/87), in case of misfires, it immediately switches off the injector of the damaged cylinder and starts to dry the spark plug. Switches DME to the open-loop mode. After drying the spark plug, the injector is turned back, and the fuel control is switched to closed-loop mode. Yes, for a moment still, the engine’s vibration can be felt, but after 10 .. 15 seconds (typically, if the engine runs idle), the correct performance of the engine is restored. In the DME error message memory, a clear record – cylinder X had misfired; we turned it off. 

For these DMEs, a correct identification of misfires is not happening; the injectors are not turned off. Unburned fuel gets in the exhaust. There, it floods the Lambda probe, gets in the catalytic converter, and then, heated up, is injected out via the exhaust pipes. Smoke, stench – everything!

In this image, we see the situation after 25 .. 30 seconds. Unstable idle became more and more chaotic till the whole car started to shiver. Subjectively – not some (or even two) turned-off cylinders, but chaotic misfires in all cylinders! Rough-run data indicate exactly that: misfires already in ALL cylinders! Total chaos!

From the moment of the crash, 40+ seconds have passed. Starting with the 282nd second, the most interesting part starts – DME suddenly starts to register misfires! Here, the most misfires are in cylinder No.5, then cylinder No.2 follows. See the chaos in the cylinder performance! Would DME be able to conclude such data from the exciting live data? No! Of course – no! These misfires are “pulled” from the initial crash event. Bosch is cheating! As (formally, due to patent war with Siemens) DME can not identify misfires from these Rough run data, Bosch pretends that it is not doing that. But then – after 50 seconds – it suddenly “remembers” the culprits!

Here, a minute after the crash, complete chaos continues. The car shivers and agonizes. The blasts can be heard in the exhaust. Smoke is still coming from the exhaust; the exhaust stinks. We see that DME continues to “see” the misfires in cylinders No.5 and No.2. 

At that moment, I got tired of torturing the car. The continuation is predictable (I have researched it in other scenarios) – read more here. In short, in each of the “damaged” cylinders, DME counts 95 misfire events and then switches off the injector; in the DME error message memory, the error messages regarding misfires in this cylinder and prophylactically – regarding problems of “several” cylinders (this time, there are really several cylinders, not even a lie). As I did not wait to count 95 (fake misfire) events, it was not a surprise:

We see that in the DME error message, memory is only error messages regarding inlet air – its plausibility and lean fuel mixture. Why lean fuel mixture? I assume that unburned fuel “flooded” the Lambda probe. DME detected it as a rich fuel mixture and made it leaner, more leaner… Then (after X seconds, according to the work algorithm) overwritted LTFT and… When the probe “recovered”, it turned out that the fuel mixture actually is very lean. Yes, these DMEs have a very robust reaction in such a situation (unexpectedly incorrect fuel mixture) – they are not able to “restore” normally. 

This time, the situation tuned out relatively positive – after restarting, the engine worked correctly. In more harsh cases, they were not even repeatedly restarting help. Then, the readaptation of the engine should be performed. 

Let’s return to the crash. In the attributes of the error message – correct inlet depression. Correct atmosphere pressure. Correct Rail pressure. Yes, the fuel mixture adaptations – “to the moon and back”. 

And here – we see a correct fuel mixture proportion, correct Lambda. And again – incorrect fuel adaptations. The situation is more interesting because Bosch DME is not correctly separating Offset and Multiplicative type adaptations, even more – in the Expert mode, even STFT is not visible (in the newest ISTA version, STFT are as if visible in the Lambda probe test). 

In ISTA, I chose the Repair plan. More precisely, what happens in these tests is described here. Without options – this time, too, no defects were found.

ISTA offers to give feedback. Here, let’s see the list: 

And here – the second part of the list.

So, I have the option to choose a defect mentioned in the list or – admit that no problem is found. No, the situation is different – the problem IS understood/found, but is it not on this list! And I don’t have the option “other” or leave a comment!

I admit that I was wrong. Just after getting familiar with B58, I had a feeling that this was a simple and boring engine. No, thanks to this Bosch DME, it gets more and more interesting! And how many desperate users of these engines can we expect in the future?

Is Bosch cheating in this situation? Without a doubt! Why is such cheating possible? There are several options, and only the involved know the answer. 

Siemens may have no complaints (or, more exactly – the time and financial expenses to the lawyers are overreaching the potential gain), taking into account the sad result gained by it.

It is possible that there were consultations with the lawyers or even court cases, but these identified misfires are not recognized as directly related to the crash event (taking into account that there is more than a minute between both events). Siemens may be allowed to cheat in this case by receiving some other “service” in return. It is possible that there is pressure from the BMW AG, to which – from one side, it is possible to ensure (even if minimum) some self-diagnostics; from the other side – it is important not to lose an important business partner. 

Remember the notorious exhaust saga which affected a range of large car manufacturers? Who supplied DDE to these vehicles? It is clear that Bosch is “only” a subcontractor and only performs the technical task. From the other side – it is clear that such a subcontractor is very well aware of what he is doing. The saddest thing is that the user of the car does not gain anything in this situation. Unfortunately, the user of the vehicle is only a loser. This time, too – the vehicle owner will lose a lot of time and money and will replace all possible components one by one. Everything that suggests ISTA. Everything that suggests the dealer and other service centers. Why? Because the self-diagnostics of the engine in this misfire situation is worse than years 20 .. 25 ago!

What can I say to the user of these engines if similar symptoms appear?

a. in case of such inadequate error messages regarding intake air flows – don’t believe in them;

b. use only high-quality ignition coils and spark plugs;

c. after a crash, leave the engine running for at least 5 minutes. Only then the error messages (also, but not only) regarding the ”guilty” cylinders will be recorded;

d. if you see the cylinders No.5 or No.6 in the list – immediately replace the ignition coils for them! Where is the problem? Read the 3rd part of this story (here)!