AU Speed: Deciphering Rotation Rates

Hello.

Nach Jans Posting zur AU im Düsentheorie-Thread hab ich mir mal meine AU-Bescheinigungen angeschaut. And lo and behold:

The vehicle data indicates a maximum engine speed of 4950 rpm, while the three free acceleration tests (which measure soot emissions) recorded maximum engine speeds of 4980 rpm and 4990 rpm, respectively.

In the second inspection after 5 years, the difference between the actual speed and the free acceleration values was 70 rpm instead of the expected 40 rpm.

Does anyone know what's behind these different billing rates?

hello,
"With the automatic transmission, there's a speed range (from X to Y) for the upshift point. If you like, I can check on Monday what the target speed is for your car (what's your engine code?)."
CU
Gerrit.

Hi Gerrit,

I'm familiar with that window (it's also listed on the AU certificate = 4800 - 5200).

I'm only surprised by the differences between the actual settling speeds and the free acceleration values.

How are they determined differently?

I always thought that "full throttle until the RPMs stop increasing" was always the same thing...

Hi Ulf,

That's exactly what I meant ). It seems your certificate contains exactly the same information... so the interesting question is, WHERE does that number come from in the vehicle's "as-is" data (5043 for me, 4950 for you).

"Yours isn't divisible by 21, which, according to the theory from the other thread, probably means it's not coming from the engine computer. But why is mine so skewed?"

Perhaps someone should eventually (I unfortunately don't have the time right now) use VAG-COM to go through all possible channels of the control unit and see if anything can be read out that is similar or even identical - or does the VAS5051 generate this data based on the engine code from an internal database?

If the throttle control system works in a way that, when the set speed limit (the actual speed) is exceeded, the fuel injection is simply set to zero, I can easily imagine that the mechanical components of the ESP system still have some inertia, which is responsible for those few extra revolutions.

Unlike the idle speed, we're not really seeing any "regulation" here, but rather the initial overshoot (more precisely, its maximum). I suspect that if someone were to be so reckless as to keep the accelerator pedal fully pressed, the speed would eventually settle very close to the specified target value.

We just need one more crazy person to test it.

Best regards,

Jan.

Deleted upon request.

diesel.gustav wrote:

Because: if the speed is measured in revolutions per second and is off by +/- 1 revolution per second, then, calculated, it would be +/- 60 revolutions per minute. When you look at all of this in detail and consider other influencing factors, then Ulf's specified speed range of 4800 to 5200 RPM, as stated in the AU documents, becomes clearer to me.

Hello Gustav.

Yes, of course: Within a target range for measurement results, the permissible tolerances of the object being measured and the measuring device must be taken into account.

And these are not even very high percentages: +/- 200 rpm with an "ideal value" of 5000 is only +/- 4% as a worst-case sum of deviations for the motor and the measurement device .

If the load is distributed 50:50, the automatic unit (AU) may already be malfunctioning if the motor and measuring device have tolerances greater than 2%.

Quote:

. . . Anyway, I wish you all a lovely Sunday!

Thank you as well (for the rest of the day)

Jan6K wrote:

"Yours is not divisible by 21, which, according to the theory from the other thread, probably means it's not coming from the engine computer, but why is mine so crooked?

"

Hi Jan,

ich bin zu faul dafür, aber vielleicht kannst Du ja nen PC rödeln lassen, um in Deinen Zahlen einen größten gemeinsamen Teiler zu finden - das könnte dann die Auflösung des betr. Systems sein.

Quote:

Maybe someone should (I unfortunately don't have the time right now) use VAG-COM to go through all possible channels of the control unit and see if anything can be read out that is similar or even identical - or does the VAS5051 generate this data based on the input of the engine code from an internal database?

If it says Is-value, shouldn't that, based on the name, refer to a single instance, but not, for example, to a general database related to ICD codes...?
And in my second absence from work, the actual value was also different from the first one.

Quote:

If the aggregate functions in such a way that when the overrun speed (that actual speed) is exceeded, the injection amount is simply set to zero, I can easily imagine that the mechanics of the ESP still have a certain inertia, which is responsible for those few revolutions above the set speed. We are not observing a real "regulation" here, unlike the idle speed, but rather the initial overshoot (more precisely, its maximum). I suspect that if someone were to be so reckless as to keep the accelerator fully pressed, the speed would settle very close to the specified target value.

However, if the same system is capable of compensating for deviations, at least in a 30 Hz cycle, in a controlled and precise manner , then, in my opinion, it should also be fast enough to prevent overshoots during the shutdown process . . .

Quote:

We just need one more crazy person to test it...

What does "verrückt" mean? In every AU (Automatic Update), the engine is tortured multiple times.
If the urge strikes me, I might actually try it and log the data using VAGCOM... of course, with the aforementioned unavoidable 21 rpm blur.

Hi,

@Gustav: Of course, we're aware of all these aspects of the measurement possibilities. We're just wondering where this peculiar "actual value" comes from, as it doesn't seem to have any connection to the high-speed tests described later - the final speeds of which, as you mentioned, are likely determined in the way you described.

@Ulf: It's very interesting that the actual value varies for different control units in the same car... I only have one log file, so that's not evident from it. This suggests that the value is likely being input or calculated externally.

It doesn't seem like there's a common factor, because 5073 and 5082 are only 9 apart, so it would have to be 9 or a divisor of 9, meaning 9 or 3. However, neither of those works because 5053 is not divisible by 9 or 3, and neither is 905, which is the idle speed.

If you go ahead with the measurements, I'm very curious to see the results!

If the AU manages to determine the maximum value relatively accurately, then the differences in the measurements from the three trials, with sampling at different points of the overshoot (which is only a few percent), can be explained, along with what Gustav wrote. Only the actual speed... that remains hidden }.

Best regards,

Jan.

Jan6K wrote:

There doesn't seem to be a factor, because since 5073 and 5082 are only 9 apart, it should be 9 or a divisor of it, so either 9 or 3. However, neither of these works, because 5053 is not divisible by 9 or 3, and neither is 905, the idle speed.

Hi Jan,

And if we assume a resolution of at least 13 bits? This would allow for a precision of up to almost 8200 rpm, for example, with a resolution of 1 rpm.

Quote:

If the AU (Automatic Unit) manages to determine the maximum value relatively accurately, then the differences in the measurements from the three trials, with sampling at different points of the overshoot (which is only a matter of a few %), can be explained, along with what Gustav wrote.

With the thought of abrupt overshoots during deceleration, while on the other hand having a very precise and completely non-oscillating idle speed control (except in some cases with PDs ), I cannot come to terms with it for the aforementioned reasons.

I don't believe that larger nozzles necessarily favor overshoot (although this is actually more relevant to the other thread, but since the aspect is being discussed here), because the mere flow rate of the nozzles, in my opinion, should not be considered one of the "dynamic" parameters (such as reaction speed or phase shift) that determine the settling behavior of a controlled system.

Quote:

Only the actual speed... that remains hidden

First of all, yes . But maybe someone could talk to a TÜV inspector, or there might be a resource online with the details of the diesel emissions test.

Hi Ulf,

ulf wrote:

And if we assume a resolution of at least 13 bits? This would allow for a precision of up to almost 8200 rpm, for example, with a resolution of 1 rpm.

At least for the exhaust gas measurement device, this could very well be the case; perhaps they even use a full 16 bits, which makes calculations easier in the CPU ).

Regarding the ASV, I just looked at one of my old logs and started examining some sample speed values. And... guess what: I found that factor of 21 there too!

Therefore, the "actual speed" likely doesn't come from the engine control unit.

It is also interesting to note: the idle speed was measured here as 903 (which is divisible by 21), while on the AU test report, it is listed as "actual" 904, and as "actually measured" 905 ).

Therefore, it again suggests using an external measurement method for the inspection, rather than relying on the engine control unit (at most, using its sensors), because, from a digital perspective, these speeds appear identical.

ulf wrote:

With the thought of abrupt overshoots during deceleration, while on the other hand having a very precise and completely non-oscillating idle speed control (except in some cases with PDs ), I cannot come to terms with it for the aforementioned reasons.

How do we know that the idle speed control doesn't overshoot by 20...50 revolutions for a few milliseconds during the initial adjustment, or "undershoots" during the decrease? I don't think that's noticeable... and the fact that it then regulates perfectly doesn't say anything about how it behaves in the fractions of a second after the initial adjustment.

You can get that kind of over-rev, for example, if you start off in neutral (without giving it any gas) and engage the clutch a bit more abruptly – that can cause the injection system to "spike," and that could easily exceed 903 RPM before it settles down. At least in my log, I see the sequence 903 -> 861 -> 924 -> 903 multiple times, and some of the numbers are even repeated. Of course, it doesn't really tell us much, because after almost a year, I don't remember exactly what I was doing, but it could very well be a normal idle start without giving it any gas.

You could try giving it some throttle in neutral using the log function, and then let it return to idle, I actually expect it to briefly drop below 903 before it stabilizes.

Quote:

First of all, yes . But maybe someone could talk to a TÜV inspector, or there might be a resource online with the details of the diesel emissions test.

Let's see what comes up...

Best regards,

Jan.

Hello.

I found something: under.
http://bundesrecht.juris.de/bundesrecht/stvzo/anlage_xia_163.html
You can view a regulation text regarding the approval process (including for diesel engines, e.g., No. 3.2).

According to that, at least four times full throttle will be applied.
The first time was for "conditioning" (whatever that exactly means, but I suspect that's where the baseline data comes from), and then three times for soot measurement - during which, of course, other speeds could be obtained as part of the overall range of possible values.

imo:

The engine speed during the AU (Abgasuntersuchung - emissions test) can be measured in two different ways. First, you can use the half-waves of the engine's signal and measure their frequency. If that doesn't work (due to a 'saturated' battery), you can try applying power to the battery to measure the frequency. If that also fails, you can use a piezoelectric sensor connected to the fuel injection line from the ESP (Electronic Stability Program) to the first injector. Since everything ultimately connects to a PC, and we've already discussed various interface problems regarding timing, etc., some tolerance in the measurement is likely and normal. If you're interested, I'll try to get my personal 'white coat' (expert) to explain it in more detail soon!

joergs wrote:

If anyone is interested, I might try to get my personal "blue coat" (referring to a position of authority or privilege) squeezed out soon!

Hi Jörg,

That would be nice:
I would be particularly interested in the worst-case tolerances of the speed measurement performed by TÜV (German Technical Inspection Association).

Let's see, maybe it will work out in the next few days, as I have to go to the vehicle inspection station anyway because of the speed limit issue with my caravan! If the inspector has time, then... !

So, I went to see my mechanic today (who, by the way, also drives a TDI and now has a brand new A6 with all the options, a V6 TDI with an Ottinger chip, putting out about 210 horsepower). Unfortunately, he was a bit grumpy and irritable today.

'I specifically addressed the issue of larger nozzles and the air intake control valve speed with him. He believes that larger nozzles are likely to cause more soot-related problems than the maximum engine speed. He said that for the 1.9 TDI engines, the maximum engine speed can be between 4800 and 5300 RPM. This is also how it's configured in the diagnostic tools.'
He said that the limits are far enough and that there shouldn't be any problems due to the larger nozzles.
'And if it were possible, there would be enough tricks to bring the engine within tolerance when it's running at high RPM. He himself had once considered doing this with his 110hp TDI, but gave up due to lack of time. However, he said it wouldn't be possible without changing the turbo boost, in terms of emissions.'

joergs wrote:

and if that were the case, there would be enough tricks to bring the engine within tolerance when it's running at high RPMs. but he said that it wouldn't be possible from a soot perspective without changing the boost pressure!

Hi Jörg,

Do you have any idea from what nozzle size increase the TÜV inspector starts suspecting excessive soot emissions?
If the automatic version gets the ALH 216 engine instead of the 154, I can imagine that happening more easily than if it already comes with the 205 engine from the factory.

Besides that, I'm also curious about how he plans to bring a motor that's running too fast within the tolerance range without losing any power.