Kniffelaufgabe: Ursachen für unrunden PD-BUK-Kaltstart? | Posts 128+

DieselBär30x wrote:

Of course, it will then be measured with the engine cooled down and (for now) not in a fully warmed-up or error-free condition.

This means that you can't just "quickly drive to the workshop and have it checked."
Either leave the cart there, or borrow a pressure gauge from someone for the weekend .

Yep, "trying it out" is always better because then you know if it really works and, if not, why.
In a workshop, the goal is either to focus on revenue/profit or on "NoCanDo" (meaning state of the art). ...not always, but increasingly often .

Best regards from Munich!

ulf wrote:

If it weren't active, why would the diagnostic data sometimes immediately reach the maximum limit of +/- 3 mg after starting up (as seen in previous log images)?

How many times have we told you not to take everything at face value that comes from the (distorted) diagnostic data?
100 or 200 times?

Quote:

Here's another question about explaining drafts: Can air in the area of the fuel tank filler neck be discharged towards the tank by running the fuel pump for a sufficient amount of time, i.e., by repeatedly turning the ignition on and off?

DiFi tandem pump -> Yes.
Twin-screw pump - ZK - PDE -> No.

Quote:

Or does one need the full flow rate of the tandem pump to flush air out of all angles of the fuel system?

That's right.


"Ulf, a former colleague, drives a BUK in his Ibiza. He doesn't experience any starting problems with the engine when it's cold. If it were a software bug, he would probably have the same issue. And he parks his car outside."


Have fun searching... It's pointless anyway.

My AXR has finally started to crank over "jerky-like." This is the first time this has happened.

Quote:

My AXR has finally started to crank over "hesitantly"

.

Haha, our Ibi ATD pumps so much diesel through the injectors during cold starts that I think it's going to explode. The colder it is, the more powerful it gets. That's how it should be. It doesn't experience any jolts after a cold start. Maybe because of @Rainer.

DieselBär30x wrote:

Of course, this will be measured with the engine cooled down and (for now) not in a fully warmed-up or error-free condition. Best regards from Munich!

The test conditions for this measurement, according to the guidelines, are: coolant temperature must be at least 85 degrees Celsius.
Perhaps you should perform both a cold and a warm pressure test. If the pressure is low when it's cold, then there's likely an air leak somewhere.
The other AXC that I logged back then had the same problem.
We need another AXC unit. I'm almost afraid that this one might have the same problem.
Didn't you have different software installed on it before, Bärli? How did he start it with that?

Hi Steffen!

Sure, I will measure the pressure both at warm and cold temperatures.

Unfortunately, I don't know anyone else who drives a (T5) AXC. You would be closer to the source .
Mich wundert übrigens immer noch, daß kurz vor Start auschließlich die AXC-MStG-SW "128" im PD-Status zeigt... Warum only??

The cold start performance with the old dataset was also poor. However, I attributed this to the faulty ZMS (Zero-Mean Subtraction) and thought that this was why the LLRR (Likelihood Ratio Regression) values were so off, especially when starting from a cold state. But it wasn't the cause, because even after replacing the ZMS, there was no improvement.

Best regards from Munich!

Unfortunately, most of the other T5 vehicles in the fleet have engines with approximately 75 kW. I need to review them all again.
The MWB 18 from AXC indeed behaves differently during cold starts compared to all other 1.9 engines. I don't know what is programmed differently in it and why. Therefore, it would be good to log another AXC to see if it exhibits the same behavior.

Bertil wrote:

ulf wrote: If it weren't active, why would the diagnostic data sometimes immediately after starting up reach the maximum values of +/- 3 mg (see previous log images)? How many times have we told you not to take everything at face value that comes from the (distorted) diagnostic data?

That's exactly what I've realized myself now - by comparing the BUK software with the ASZ.
But it's not necessarily "inactive" -> could it be that goettmann was referring to the LL-SPEED control system, which I wasn't even talking about here?

Quote:

Quote: Here's another question about explaining drafts: Can air in the area of DiFi-ZK-PDE be discharged towards the tank by running the fuel pump for a sufficiently long time, i.e., by repeatedly turning the ignition on? DiFi tandem pump -> Yes. Twin-screw pump - ZK - PDE -> No. Quote: Or does one need the full flow rate of the tandem pump to flush air out of all angles of the fuel system? That's how it is.

Hmm, then I guess I'll have to try the hardware brute force approach: Measure the pressure in the clear hoses AND with tandem pumps - if I can find a pressure gauge.

Quote:

Ulf, a former colleague, drives a BUK in his Ibiza. He doesn't experience any starting problems whatsoever. If it were a software bug, he would probably have the same issue. And he parks outside.

I still believe that the BUK software contributes to the stuttering (details available upon request via private message or email) - AS SOON as the hardware of all cylinders doesn't run synchronously enough after a cold start, for example, due to small air bubbles in the diesel circuit.

If my colleague's BUK (Build Kit) isn't stuttering, it might have fewer hardware issues than mine, and/or it hasn't been cold enough for him yet.

By the way, mine started stuttering again yesterday afternoon after being outside for almost 10 hours in fairly windy conditions at around -4°C, which means my software bug fix attempt apparently doesn't address all the underlying causes.
However, based on my current understanding, it alleviates the problem to some extent.

Quote:

My AXR has finally started to crank "jerky"-like. This is the first time.

Welcome to the club .

Dieselbär30x wrote:

From conversations with Steffen_W, I know that another AXC (a cheap exotic version in the T5) also has significant runout problems after starting at temperatures below 0°C. Here too, the LLRR values are incorrect, and the PDE statuses were "128" before the start (which seems to be the normal startup state for the AXC only??), and not least, here too the PDE switching times were firmly in the negative range.

Ha, I had completely forgotten about this approach!
Strongly negative values usually indicate the presence of air in the pressure differential equipment (PDE).
While the values on the BUK unfortunately remain at zero for a longer period than the stuttering lasts... how long does it take for the AXC to reach normal PDE switching times? Longer than the stuttering, or approximately the same...?

Hi Ulf!

Apparently, the ping times for the PDE server at AXC remain within the normal range of around -50.
(Are you looking at the logs I uploaded ?)
Even swapping the DFI chip didn't make any difference.
Similarly, regarding Steffen_W's comparison log with another AXC.
Therefore, it seems that this is normal for the exotic PD with its PD start status of "128" .

Best regards from Munich.

PS: How much time does it take to read your software bug documentation?

DieselBär30x wrote:

The PDE response times appear to remain within the normal range of around -50 on the AXC.

So, have you ever observed any noticeable changes in the data at the moment when the engine stabilizes?

Quote:

PS: How much reading is required for your software bug documentation?

If you have enough basic knowledge about the EDC15+P, it's described in a few sentences.

Again: Please see the logs that have already been uploaded .
The timing values change only slightly between an unstable idle and normal operation – at least not significantly. PDE number 2 shows somewhat different values in both LLRR and MWB18, both when the engine is cold and warm.

Okay, then let's not use private messages or emails .

Best regards from Munich!

But don't be fooled by the term 'not active' - it's possible that goettmann was referring to the LL-SPEED control system...


No.

I meant the LLR (Leerlaufregelungs- regulation), which is the idle speed regulation in your statements. The last expression doesn't actually exist. Although it's included in all workshop documents, it's not referred to as such during development.
We've already been talking about the same thing.

Once again:
The LLRR (to make it clear ), which you can read from the diagnostic interface, are purely calculated values. It should be obvious to anyone who has some experience with control engineering that these values may not provide meaningful results at the moment of start-up and for about 3-5 seconds after start-up. The LLRR can only send meaningful values via the diagnostic interface when the control processes have stabilized, i.e., when a stable idle state is present. These values are also used for correction purposes in this way. If there is no established regulation, the LLRR is deactivated. However, display values are still provided.

Hello Göttmann!

*Ulf-Mode: ON*
Does this mean we can never determine the cause or identify the actual cylinder(s) that are "misfiring" during a cold start using MWB13?
*Ulf-Mode: OFF*

If that is correct, should the LLRR data only be evaluated (or can it only be evaluated) when the PDE status is "0.0"?

Best regards from Munich!

DieselBär30x wrote:

Hello Göttmann! *Ulf-Mode: ON* Does this mean we can never determine the cause or identify the actual cylinder(s) that are "misfiring" during a cold start using MWB13? *Ulf-Mode: OFF* If that is correct, one might only be able to evaluate the LLRR when the PDE status is "0.0".

Roughly speaking, I would interpret it as follows:
As long as the MWB13 data are not translated into actual differences in quantities, one only sees what the regulation wants to show, presumably mainly during the phase where the fluctuations increase, because no success is achieved.
Once the load balancing system is functioning correctly, the regulation will recognize that its requirements are being met and can stabilize, or it may not necessarily continue to increase -> rather, fluctuations will tend to decrease or remain constant.

However, a self-stabilizing engine operation (e.g., when air has been bled from the fuel system) could also reduce the MW13 fluctuations...

goettmann wrote:

... If no specific regulation is in place, the LLRR will be deactivated. However, display values will continue to be provided.

I understand it like this:
The closed-loop control system is currently disabled, the fuel injection quantity is being controlled, and therefore values from module MWB13 cannot be evaluated until the control loop stabilizes.
"...but then the question remains: why is it always the same specific memory location that malfunctions during and immediately after startup, rather than a random one?"

Best regards from Munich!

DieselBär30x wrote:

I understand it like this: The LLRR is deactivated, the fuel injection amount is controlled...

Therefore, this cannot apply to the "total quantity" of all cylinders, otherwise the LL-speed would not remain relatively stable immediately after starting, even with varying engine load (turning on the lights, accelerating without giving gas...).
At most, the fuel distribution (LLRR) between the cylinders could be controlled, but even that, in my opinion, would likely increase the tendency for misfires rather than provide uniform fuel delivery to all cylinders.

Quote:

...then the question remains open as to why, during and immediately after startup, always the same MW of the LLRR deviates from the norm, and not – randomly – some other one?

Because that is the reaction to the measured speed differences between the cylinders - and hardware-related strengths and weaknesses of the cylinders usually remain where they are.