DZR Explained: More Details and Insights

Hi everyone,

I'm starting this as a new topic to avoid mixing wild guesses (like this one) with substantiated facts .

I just looked at my diagram from yesterday, specifically the second one, because the increase remains more linear there, and I calculated the actual pulling force.

If I'm not mistaken (it's been over 10 years since I studied physics in school), then I'm getting a real acceleration of 2.33 m/s^2 at the bottom, which corresponds to a pulling force of 2891 N.

The ventilation calculator indicates a value of 2880N, which is very close to the actual value.

If you look at the curve more closely and overlay the virtual curve from the computer, you'll notice that the actual increase is significantly steeper than what the computer shows. According to the DZR (presumably a regulatory body), about 25% is "lost," which isn't as drastic in my curves. And because of this, the errors become increasingly pronounced.

Perhaps the root cause of all of this (including the excess power of the TDI engine) lies in the fact that the torque curve drops off much more gradually "on the upper end" than we previously assumed.

What do you think about it?

Best regards,

Jan.

Hi Jan,

If the torque curve were "flatter" at the top end, starting from the standard/official 235 Nm, then more Nm would be present at the point of maximum power, and after the multiplication "kW = Nm x rpm / 9549", a higher kW value would inevitably result.
Which seems to be confirmed by various recorded times.

I recall my post from yesterday, where the analysis of your average(!) acceleration resulted in maximum deviations of 1.5%, while the uniform torque curve in the DZR showed an increase of 14%.

Okay, maybe I made a mistake in my calculations, even though I've checked most of the results multiple times (and even found a few errors before I wrote the post).

My question is more about why the curves are "strongly" different in the first place, rather than getting stuck on a "seemingly ideal" trajectory.
--> Measurement errors, evaluation errors, a significant overshoot in charging pressure, a slight downward slope, or a strong tailwind towards the end of the measurement...??

As long as multiple measurements are available, it is my opinion that, for methodological reasons, one should base calculations on average values – or, if necessary, even exclude extreme results from further consideration.

I have, in fact, only calculated it for that one curve, but the other one is, in my opinion, steeper in the upper part (it looks steeper than the DZR would indicate ).

Of course, I might also be mistaken, because your calculations yielded something different, and right now I trust your knowledge of physics more than my own .

Regarding the discrepancies between the two curves, I believe the cause lies in the unfortunately very coarse sampling rate. The measured values are only recorded every slightly more than one second, especially when logging three data blocks. While this can be interpolated to some extent, as the curves have sufficiently straight sections, it does indeed have a negative impact on the curve shape.

"You can see it clearly when you compare the speed and the RPM curve. I didn't disengage the clutch, so they should follow each other closely, but they don't (due to the kinks). The timestamps are written separately for each of the three blocks, which shifts the extrema and makes the analysis more difficult, especially since outliers might also be caused by the coarse grid."

The conditions were exactly the same, which I didn't mention yesterday, and both times we were heading in the same direction (with no wind and a flat road). There was no other way to go at that point; I don't want to drive at 110 km/h into a residential area.

Someday I'll repeat all of this on my "usual route," but for now, it's quite a distance to travel. However, there's plenty of space there for many measurements in both directions, all while complying with traffic regulations. I'd also like to try it with just one measurement block at some point, which should roughly triple the sampling rate.

Here's another possible explanation: In the "better" attempt, the boost pressure was slightly higher, but this is countered by the fact that the overshoot is more noticeable in the other attempt.

"It's interesting to see that the 2 bar pressure only reaches that level shortly before 2000 RPM - is that also the case for you, Ulf? You can see this on the LDA. Based on my diagrams, I would say that the speed range of 30...35 km/h was just enough."

Well... it is and remains a mystery .

Best regards,

Jan.

Hi Jan,

In fact, the two extremes of your four measurement runs, with the specified engine speeds in the "high-speed" range, differ by almost 12%, specifically around 4000 rpm.
It's no surprise, then, that the curves look different in steepness.

On the other hand, it's understandable because the "slower" measurements have a higher RPM range (3885 to 4242 instead of 3612 to 4011), and as is well known, thrust decreases noticeably above 4000 RPM.

Finer measurement intervals would therefore certainly provide more precise individual results.

Quote:

Overall, it's noticeable that the 2 bar pressure only reaches its value shortly before the 2000 RPM - is that also the case for you, Ulf? You can see this on the LDA - based on my diagrams, I would say that the speed range from 30...35 km/h upwards was just enough.

I started the test with full throttle at similar speeds, and the pressure then reliably stabilized at slightly above 1 bar, possibly after some minor overshoots, reaching up to 2000.
"However, I reach 1 bar a bit earlier, around 1700 rpm, if I remember correctly (with one eye on the tachometer and the other on the road - in that case, the LDA [Linear Drive Assist] shifts slightly out of the center of my attention )."

Hi Ulf,

2 attempts, not 4

Of course, I also gave it my all. However, we probably shouldn't overinterpret that increase; it's likely also due to the gaps in the sampling.

"I think we should increase the sampling rate threefold. I'm going to log only block 11. This will reduce the data volume, but we'll still retain information about engine speed, target boost pressure, actual boost pressure, and the boost pressure control valve."

6 would give us speed and the position of the pedals (which would be useless ), and three would finally give us the RPM, the air mass, and the EGR position, which also seems likely to be pointless.

7 could also be interesting, that's the intake air temperature, but I don't think you'd see much difference in those 6 seconds.

So, 11 by itself... what do you think?

Do you have any other logging requests? I mean, if I'm already out there with all the equipment, it's no problem to record more data or run additional tests.

Best regards,

Jan.

Hello!

I briefly had the opportunity today to talk with a professor of vocational education about the German vocational training system and additional benefits.

In Austria, a motor vehicle is allowed to have a power output that varies by +/- 5% from the value stated on the registration document. At BMW (where the person I'm talking about used to work), they adhere to this regulation, but almost every vehicle has a power output that is +5% higher than the stated value, and none have a power output lower than the stated value.
However, other manufacturers often have larger (illegal) power increases at low nominal power levels (around 66 kW), which can also amount to approximately 10%.

Regarding the DZR: The professor confirmed the approximate size of the mass moment of inertia. The term 'rolling resistance' should rather be called 'driving resistance' and include all types of friction. The air resistance is correct, but ideally, the total resistance should be determined through a controlled test involving deceleration from a high speed under perfect conditions. Of course, this is not possible, as the DZR needs to remain compatible.
The losses in the powertrain would then be too high. The order of magnitude is approximately 6%, but this varies depending on the vehicle.


So: Rolling resistance = Weight * 0.15..0.25 (but measuring would be better).
Power loss (drivetrain transmission) = 6% (pure efficiency; resistances are already included in the rolling resistance).

If a TDI engine can produce 10% more power, that's correct. But everyone whose name isn't Ulf or Rainer should double-check the values they entered.

Greetings.
Ernst.
P.S.: Perhaps I can have the DZR (Diplom-Arbeit/Thesis) reviewed in more detail in the next semester.

Hi Ernst,

Using my values (the correct ones, meaning the accurate translations like "Golf 3"), the correct 185/55/15 tire size, and the correct 1240 kg weight, the currently downloadable version results in a time of 7.3 seconds. However, measurements taken the day before yesterday yielded times of 6.18 and 6.31 seconds. As mentioned above, I will repeat this test again with a higher sampling rate, but the measurement method itself should not have any errors related to the timing, as VAG-COM provides the time with an accuracy of 1/100th of a second.

If you like, I can also provide you with the precise measurement data from the new tests for analysis. There would be approximately 20 measurement intervals spread throughout the DZR test duration, for which we would obtain the exact times. Perhaps this will help advance the discussions...

Best regards,

Jan.

Ernst S. wrote:

Regarding the DZR: The expert has confirmed the approximate size of the mass moment of inertia. "Rolling resistance" should rather be called "driving resistance" and then include all types of friction. The air resistance is correct, but it would be best to determine the total resistance through a coast-down test from a high speed under perfect conditions. Of course, that's not possible, as the DZR needs to remain compatible. The losses in the powertrain would then be too high. The order of magnitude is approximately 6%, but this varies depending on the vehicle. So: Rolling resistance = Weight * 0.15..0.25 (but it would be better to measure it). Losses (drivetrain translations) = 6% (pure efficiency, resistances are already included in the rolling resistance)

Hi Ernst,

With these "minor" losses, the purported average performance gain of TDIs would actually shrink to around 5%.

Then I wonder how, for example, "pp," who is clearly a "practitioner," arrives at an overall efficiency between the motor and the test bench of 0.74 and 0.83...
In detail, he specifies losses due to friction between the gear teeth in the power transmission at approximately 7%, oil churning losses at approximately 2%, and rolling resistance of the tires at approximately 7-20% of the engine's power output.

Even with the "most favorable" assumption of test bench specifications, resulting in 17% overall losses (for which the rolling resistance in the DZR is set to 0), for example, Jan's Ibiza would need to have approximately 127 horsepower to achieve the measured 6.2 - 6.3 seconds in the acceleration test.
Unterstellt man die größten o.a. Verluste mit einem Gesamtwirkungsgrad von 0,74, so bräuchte man für die o.g. ZeiOver 140 horsepower?!?
That would certainly be a bit extravagant, but even the less powerful 127 horsepower instead of 110 isn't bad either.

I see, we should revert the DZR back to its original purpose, which is comparative measurements, such as before and after filter changes, tuning modifications, etc.

Hello Ulf!
Too bad, now that we're so close, you want to quit .

And yet, what the expert said sounds so reasonable. Moreover, your measured rolling resistance (driving resistance) accounts for approximately the excess in the losses. And the resulting values are quite realistic, as I said .


Just one more question: Tire rolling resistance of 7-20%... isn't that quite a big difference? Wouldn't it be 7% for more powerful cars and 20% for less powerful ones, simply because it makes a bigger difference in those cases? Is it perhaps not even efficiency, but simply rolling resistance?
In a previous post, it stated: 'Efficiency of the gearbox-axle drive: 0.91-0.95' --> which corresponds to the 6% mentioned by the professor, and everything else is friction and no longer efficiency. This is what is measured on the test stand...
Those being tested probably don't care much about what rolling resistance is. Your dynamic resistance model is certainly much more complex and physically accurate.
Best regards.
Ernst.

Ernst S. wrote:

Too bad, now that we're so close, you want to stop

.

Well, I'm currently at a loss as to the difference between the explanations given by your professor and the explanations provided by the testing facility staff .

I don't want to stop completely, but we're going on vacation next week... and I don't want to start any major new discussions .

Quote:

Just one more question: Tire tread wear of 7-20%... isn't that quite a significant difference.

Sure, an uncertainty factor of almost 3 is quite significant...

Quote:

Doesn't it perhaps amount to 7% for more powerful cars and 20% for weaker ones, simply because it makes a bigger difference there.

Is it perhaps not even efficiency, but simply the rolling resistance.

I cannot rule out, but also cannot confirm

Quote:

. :In a previous post, it stated: Efficiency of the transmission-axle drive is 0.91-0.95 --> which corresponds to the 6% mentioned by the professor, and everything else is driving resistance and no longer efficiency. What is being done at the test stand...

}

Just now: Theory versus practice, and us stuck in between - see above.