TDI Performance: Real-World Results & Reviews

Hi everyone,

Quote:

lt. diversen Tunern mit reproduzierbaren Prüfstandstests auf modernen Prüfständen (***) war noch kein Exemplar mit mehr als 5% Streuung nach oben oder unten dabei. Wenn D still B. (Docter)

This quote from the "Ruckel" thread is now worth starting a new topic about.

I have to revisit the topic of the "Durchzugsrechner," which, based on the times usually posted here, suggests that performance gains are unlikely to be less than 10% - unless a forum member's vehicle is currently experiencing some issues .

I'm not trying to stubbornly defend the "overly powerful" TDIs, but that doesn't quite align with Docters' quote – so one of the two must be "wrong."
Either the average times posted here, or the calculations in the online calculator, contain calculation or logical errors, or the measurements from the tuners mentioned above are not representative or are incorrect.

Who can provide contributions towards resolving the "contradiction"?

Hi Ulf,

The fact that there is variation is evident from the many contributions here, but we will probably never find out exactly what is causing it, because every measurement has its weaknesses, and we know where those weaknesses lie in the case of test results.

And the subjective factor is also significant... sometimes the car feels different, but on the other hand, it's a question of whether it's the car itself, or rather one's own perception. Especially when trying to diagnose a problem, the psychological factor plays a major role.

Perhaps, sometime in the distant future, we should organize a TDI meet, somewhere near a nice Ulf test track. Then, we could always have two people (one driving, one measuring) at the exact same spot, and ideally use VAG-COM (which I think should be possible) to measure the acceleration times. We wouldn't have absolute values, because the track would still be variable, but we would at least have a comparison that shows the variation, and it would help filter out subjective factors.

For example, I'd like to know what you, Ulf, would say in a direct comparison between the AFN and ASV engines, because I remember your statement that even with the power difference taken into account, your wife's 90hp ALH engine feels slower. It would be interesting to hear your assessment on that.

My suspicion is that the torque curves of the ALH (or AHL, I always forget which) are deliberately designed to perform noticeably worse than the ASV, because if people didn't notice the difference, perhaps fewer people would pay the extra cost for the 110 horsepower, especially those for whom acceleration is more important than top speed.

That's all from me for now... just some additional thoughts to consider.

Best regards, Jan.

Hello!

Try changing each individual value in the airflow calculation software by only a small amount. Then, add a timing inaccuracy of 0.2 seconds, and I would say that the variation between different units is likely smaller than the inaccuracy of the airflow calculation software. This makes it impossible to verify the results.

Subjective factor... to be honest, it's still present for me, even though I now say that, according to the DZR (presumably a measurement standard), I only have about a 5% increase in performance.

Greetings.
Ernst.
(For me, the subjective factor is approximately 1% (at most )).

Hello,
I was just wondering if the inaccurate speedometer reading (which is always a little too high, and probably proportional to the speed) might also have a small effect? After all, you are using the data obtained from the displayed speed!

The specified speed certainly has a significant impact on the inaccuracy of the DZRechner. Whether the speed is entered via the speedometer or calculated based on wheel radius and gear ratio, it makes a difference.

And if you want to determine the dynamic rolling radius yourself: As a first approximation, it is 2/3 r_0 /3 r_S. Here, r_0 is the radius of the undeformed tire, and r_S is the static tire radius, measured at the point where the wheel rests and is compressed.

Whatever you choose, it will always be inaccurate, and the speed has a significant impact.
That's why Ulf, Rainer, and others repeatedly mentioned in old forum posts that the dynamometer should not be used for absolute power measurements, but only for comparison purposes under identical conditions.
Is that still valid?

Hello Ulf!

Have you ever wondered if you might have any duplicate entries? I think I found something.

In literature (as cited by Mitschke in 'Dynamik der Kraftfahrzeuge'), the rolling resistance coefficient (fR) for passenger car tires is given as: 0.01 - 0.025.
It increases with the increasing driving speed, v.
Rolling resistance plays a minor role in calculations related to vehicle dynamics. However, it is important for determining fuel consumption.

I suspect your measurement of rolling resistance included some factors that are already accounted for in the 12% loss.

I've also incorporated the speed dependency into the DZR and emailed it to Rainer. Please take a look and see if it works for you. This results in approximately a 3% increase in efficiency.

(For those interested in more details from the book: In a stationary tire, the pressure distribution in the contact patch is symmetrical. The line of action of the resulting normal force passes through the center of the wheel. When the tire is in motion, there is an increase in pressure in the front part of the contact patch, due to elastic and dissipative deformation resistances. The resulting normal force now acts in front of the center of the wheel. This creates the rolling resistance moment.)

Best regards.
Ernst.

Hi Ernst,

Quote:

That's why Ulf, Rainer, and others repeatedly mentioned in old forum posts that the dynamometer (DZR) should not be used for absolute power measurements, but only for comparison purposes under unchanged conditions. Is that still valid, or?

In "common" practice, yes.
Nevertheless, the basic idea of the DZR, in my opinion, offers a "testing platform for on-the-go use" that, with the now-integrated correction options, should deliver almost perfect results - if vehicle data (especially weight + speed at 2000 rpm) and environmental conditions are entered correctly.

And since the times posted here for comparable vehicles are usually quite close, it seems either the same mistakes are being made everywhere, resulting in an inflated horsepower reading (which I find unlikely) - or the average TDI in this forum actually has a 10% power surplus, or even more.

Using the example of our ALH-Ibiza, I can assure you that the relevant data was determined with extreme precision, and my routine using a stopwatch allows for errors well below 0.1 seconds.
However, the dyno test using a dynamometer shows a "calculated" power output of approximately 103 horsepower instead of 90.

I didn't actually weigh our cars; instead, I used the unladen weight specified in the vehicle registration documents.
If they were a bit lighter, that could naturally explain some of the "good times."
Again, using the ALH-Ibiza as an example: for the stopped times, 5% more power would be sufficient instead of 15%, if the car were about 120 kg lighter.

However, since most cars are heavier than what's stated in their registration documents, I would like to rule that out as much as possible...

Hi,

Here's a suggestion for the test, without getting too deep into the math:

How about we do a kind of "differential" test? So, a "normal" test, then invite a defined weight (e.g., 100 kilograms or more), and then measure again.

It has no effect on air resistance, in my opinion only a very minimal effect on rolling resistance, but it definitely affects the mass that needs to be accelerated.

The excess velocity would then result solely from the force required to accelerate that mass. And theoretically, very precise calculations should be possible, right?

Best regards,

Jan.

Jan6K wrote:

How about we do a kind of "differential" test? So, a "normal" test, then invite a defined mass (e.g., 100 kilograms or more), and then measure again.

Hi Jan,

I quickly calculated something in the DZR (German vehicle registration document) for our Ibiza.

If it had 90 horsepower, the time would increase by 0.73 seconds with 100 kg more weight.

If it had the (currently assumed) 103 horsepower, the time would increase by 0.61 seconds.

So, only a 0.12-second difference for 13 horsepower .
...while the "normal measurement" still shows a difference of more than 1.2 seconds, which is, in a way, a tenfold improvement in resolution.

That means that with your method, it would be impossible to avoid an "automatic measurement" using a VAG-Com or similar device in order to ensure the necessary precision.
So, at least for the majority of forum participants, it is probably not feasible right now.

Hello Ulf,

With VAG-COM, you can also use tricks to measure more accurately, but you only get approximately 3 measurement blocks per second.
More precisely, it's quite straightforward by examining the slope of the velocity curve with respect to the time axis and then interpolating from the known measurement points. The times of the measurement points themselves are accurate to 1/100th of a second.

Best regards, Rainer.

@Ulf

I weighed my car: 1260 kg, with an accuracy of +/- 20 kg. The manufacturer's specifications state 1239 kg. The vehicle was empty of a driver, but had 40 liters of diesel fuel in the tank.

Best regards,
Uwe

Hi everyone,

Ulfs point is certainly correct that it wouldn't be easy to use this method directly due to the extreme measurement accuracy required. However, we could try it to "calibrate" the DZR. Currently, we are only guessing at the values of the factors, but I believe that this approach would give us a much better indication.

Because the absolute value for the drag coefficient is significantly reduced by all those errors (friction, air resistance), a comparative measurement would have the advantage of eliminating or minimizing these errors to some extent.

Basically, you would need to do the following:

"A series of measurements should be taken using VAG-COM, multiple times, so that any "outliers" can be discarded. Then, the same process should be repeated with an additional 100 kilograms. The accuracy of 1/100 s mentioned by Rainer should provide very precise results, and I don't think much is lost through interpolation."

Then, by analyzing the difference, one could calculate the performance very precisely and compare the result with the conventional DZR (dynamometer reading), adjusting the parameters accordingly. This should be beneficial for everyone, even those who don't have a VAG-COM. Or am I missing something?

Someday I might try using VAG-COM for that... but right now, I don't have the time, especially since I have to drive quite a distance to get to a suitable test track.

Best regards,

Jan.

Hello,

Take a look at this:

http://www.autobild.de/suche/artikel.php?artikel_id=1689&artikel_seite=4&A_SESS=951c2d47c7da4b5421b2be19d4640d82


The article in Autobild discussed chip tuning for the TDI pump-injector engine with 100 hp. The test car had 105 hp and 248 Nm on the Autobild dynamometer. Can you really talk about a significant increase in power with those numbers...? That's a 3-5% deviation.

best regards

Adrian.

MarkusAFN wrote:

Hello, I was just wondering if the inaccurate speedometer reading (which is always a little too high, and probably proportional to the speed) might also have a small effect? After all, you are using the data obtained from the displayed speed!

Hi Markus,

Sure, "bad" errors can creep in regarding the speedometer reading... that's why the rule of thumb is "speedometer - 5%" for those who don't know their gear ratios.
Determining speed using a speed calculation sheet is usually more accurate.

Ernst S. wrote:

Hello Ulf! Have you ever wondered if you might have any duplicate entries? I think I found something. Bitte gib mir den deutschen Text, den du übersetzt haben möchtest. I suspect your measurement of rolling resistance included some factors that are already accounted for in the 12% loss. I've also incorporated the dependency on speed into the DZR and emailed it to Rainer. Please take a look and see if it works for you. This results in approximately a 3% increase in "saved" power.

Hi Ernst,

The 12% losses are significantly lower than what is typically assumed in dynamometer tests, to the best of my knowledge.
Therefore, in my opinion, rolling resistance is not necessarily "double-counted."

Even in "my version" of the DZR, the power requirement for rolling resistance increases linearly with speed.
P = F x v, where F, the rolling resistance force, remains constant.

However, the power requirement due to rolling resistance is not explicitly calculated anywhere, so its "contribution" to the overall calculation might be overlooked.

I haven't compared it to your update yet. It's too late for that today --> the mattress is calling .
See you later!

@Ulf:

I also observed the same deviations (likely due to perceived performance gains) in various gasoline models. This doesn't seem to be limited to TDI engines. The values were calculated using the DZRechner software (modified for gasoline engines, with some measurement times extended for higher resolution).

Perhaps a test in 4th gear on the BAB (Autobahn) Sindvoll would increase the resolution. Of course, it's not always easy to do.