VF-Engineering addresses misleading F80 M3 & F82 M4 S55 tuning dynos (correction factors) and the difference between consistency and marketing

[Sticky]

Tuners using inflated peak figures to market with is nothing new. Often times on forums people just look at the peak figures on a dyno to determine which tune is 'bes' and this is a very flawed approach. Pushing an engine too hard just to get an impressive peak number may look nice on the graph but how nice will it feel when something let's go?

There are many tricks tuners employ. One that is popular with a certain tuner is to get the car very hot and induce heak soak for a baseline run. Then the tuner waits until the car has fully cooled down and does a 'cool' tuned run to exaggerate the gain from the baseline.

There are many ways to influence dyno numbers such as using different gears or playing with correction factors and VF-Engineering delves into why consistency is important below.

Now that we're a forum sponsor, I just wanted to make it known that we're here to answer questions, offer feedback, respond to constructive criticism, etc...

Keeping with the theme of the original post, independent 3rd party dyno results are right on par with our advertised gains by design. We've all been personally bothered when you see that the advertised gains for some companies products seem to always be just ever soo slightly out of reach. Understanding that a lot of this is marketing and companies trying to put their best foot forward, we've made it a point to remain slightly more conservative and let the product and customer reviews do more of the talking.

Here's a excerpt from something we wrote recently to address how misleading dyno numbers can be within context. (hope you enjoy!)

---

Aside from what people consider to be the more "marketable" numbers, we prefer to use our DynoJet for it's measuring capabilities as it gives us the best ability to overlay multiple runs and look for consistencies or inconsistencies as we push the car in back-to-back-to-back-to-back runs.

The purpose of these multiple run-groups is to track the changes as temperatures increase, or as we make subtle calibration changes for fine tuning. As you can see below, the consistency of our HEX Tuning Stage TWO runs in the sequence below is quite impressive.(5 runs in a row with 30 seconds rest, then a 10 minute rest before runs 6-7.)

Reading the results.While it's pretty common to see single run comparisons, the more run examples the better. The following plots will give you an idea of how a single run on it's own doesn't necessarily mean much without something to compare it to.

Stage ONE - STD Correction | Smoothing 0 - 486HP 529TQ

Stage ONE - STD Correction | Smoothing 5 - 482HP 521TQ

Stage ONE - Uncorrected | Smoothing 5 - 472HP 510TQ

Stage ONE - SAE Correction | Smoothing 5 - 471HP 509TQ

As you can see from the 4 variations of the exact same run above, there is a deviation of as much as 15HP and 20TQ just by changing the correction factor and smoothing.

When Uncorrected, the dyno is purely displaying the power at the wheels asthey are measured by the machine. SAE and STD Correction will adjust the values based on the the relative air temperature, humidity, barometric pressure, etc. at the test site.

EXAMPLE:
SAE is established at 77°F and 0% Humidity.
STD is established at 60°F and 0% Humidity.

On a 90°F day in July with 70% humidity, the Uncorrected horsepower and torque will likely read lower than on a cooler day with less humidity. SAE or SAE Correction will equalize the results as if they were measured at their respective temperature/humidity.

Generally on a warmer day the correction factor will increase the final results, and on a cooler day the results will decrease.

There are plenty of arguments to be made for or against the use of correction factors, but we'll save that conversation for another day. At VF Engineering we prefer to advertise Uncorrected numbers as we feel the results are more accurate when your before and after dyno results may not take place on the same day and under the same conditions.

[kbaldi29]

Awesome info! Getting the customer to understand difference between graphs and not just go by peak numbers is something most performance shops would benefit from. Trying to explain why a tune is better than other can be really difficult nowadays with so many options in the market, and can make you look like you are hating on others work! Lol. Giving the customer all three charts, corrected and uncorrected I feel like it's the best way to do it, and leave it to his discretion which one to show, at the end of the day, they are his results from a service paid to be performed!

[Bowser330]

Car enthusiasts should really unite on one single dyno platform & configuration....so we can have a platform to compare different makes and models.

[Sticky]

Car enthusiasts should really unite on one single dyno platform & configuration....so we can have a platform to compare different makes and models.

We could but tuners would just avoid it as it wouldn't serve their interest.

We'll never have a set standard. The tuners who oblige would be a disadvantage as some other guy will come in and do who knows what to inflate their peak figures.

[VF-Engineering]

Thanks!

If customers learn how to examine/read area under the curve, then the numbers themselves almost become secondary.

One of the things that we see quite often is a dyno plot with impressive "Peak Gains", that unfortunately for unassuming customers, is based on a boost/torque spike that is short lived and/or at an RPM range that you never naturally fall into between shifts. Just as drag racers will examine their dyno plots to find the most beneficial RPM to short shift at in order to fall right back into the range of peak torque, general consumers should be looking to see if the range of peak power is in an advantageous area for their driving style and usage.

Here's another excerpt from our article on how to differentiate between Peak Gains / MAX Gains / AUC:

------------------------------------------------------------------------------

Peak Gains

Peak Gains are probably the simplest to understand, and the most commonly referenced.
Peak, just as it's defined, is the highest point at a specific time. For our purposes, this
will be the highest point for horsepower and torque across the entire rev range. Once those
points are identified, the difference between the two runs will represent your Peak Gains.

1) Stock Horsepower: 407.82
2) Tuned Horsepower: 472.20
(for a Peak Gain of 64.38 HP @ 6,150 RPM)

3) Stock Torque: 411.49
4) Tuned Torque: 510.11
(for a Peak Gain of 98.62 TQ @ 4,550 RPM)

The results here show a Peak Gain of 64 Horsepower and 98 Ft. Lbs. of Torque for our
HEX Tuning Stage ONE ECU software on 91 Octane Fuel. As the before and after runs
were performed on the same vehicle, on the same day, with the same fuel, without ever
removing the vehicle from the dyno, it is absolutely appropriate to show Uncorrected
numbers.

From a customers perspective this is in line with our advertised gains of 60-70 WHP
and 90-100 WTQ for Stage ONE. Additionally the Peak Gains are not the result of a
momentary "spike" at a random or odd rpm range, but right in a very usable portion of
the power curve further exemplifying smooth and consistent power delivery.

But why does that matter?
It is possible to see a dyno plot for the same model of vehicle which produces higher
Peak Gains, while unfortunately not fairly representing "usable" or "consistent" power
because this peak occurs at a less than ideal RPM, or is only a momentary spike
before falling right off again.

IE:
A 488 HP and 515 TQ spike for a brief moment is less than ideal than
a 477 HP and 510 TQ peak at the top of smooth curve.

And in referring back to our first post at the top of this page, imagine if those
488 HP and 515 TQ numbers had STD Correction factor and Smoothing 0 ...

The more context you offer, the more you start to see that peak numbers
don't always offer a complete story.

MAX Gains

Max Gains are similar in that we are still focusing on the delta change between the stock
and tuned runs, but instead of looking at the highest point across the curve for each run,
we are looking for the largest delta or improvement at the same RPM.

1) Stock Horsepower: 349.11
2) Tuned Horsepower: 459.02
(for a MAX Gain of 109.91 HP @ 4,800 RPM)

3) Stock Torque: 382.22
4) Tuned Torque: 502.57
(for a MAX Gain of 120.35 TQ @ 4,550 RPM)

Impressive, right?!
The results here show a Maximum (Max) Gain of 109 Horsepower and 120 Ft. Lbs. of
Torque for our HEX Tuning Stage ONE ECU software on 91 Octane Fuel. As you can see,
that is a noticeably higher delta than when examining peak numbers. While Peak Gains are
typically the numbers you see advertised by manufacturers as they represent the "highest"
gross numbers, you'll often hear discussion about Max Gains as they represent the "largest"
net gains.

As with Peak Gains, there is a level of importance to where Max Gains happen within the
rev range. They often do exploit peaks and valleys within the power curve which is why it's
important to note whether similar levels of power are sustained throughout the rpm range,
or if they are short lived due to a momentary spike in tuned power or dip in stock power;
which interestingly enough ties, directly into the next segment ...

Area Under the Curve (AUC)

Area Under the Curve refers to the total volume of power below the horsepower or torque
line as it's displayed in the graph. For our purposes, when examining "before and after" runs
for the same vehicle, we are looking for the Area Under the Curve that represents the
increase in power between stock and tune runs.

Using the exact same dyno plot as we've should throughout this thread, we've separated
Horsepower and Torque so they can be viewed independently in the two graphs below.

Simply put, the visible "blue" area represents your total volume of gains over stock.
The visible red/purple area becomes your baseline so that instead of putting the emphasis
purely on the numbers, you have a visual representation of how the software has influenced
power delivery throughout the entire rev range.

In determining optimum power delivery, the more Area Under the Curve you have across
the widest RPM range, (assuming we are discussing two identically weighted and geared cars)
the more powerful and quicker the car will ultimately be.

We'll skip the calculus lesson that is used to determine the total area under the curve since
fortunately the Dynojet software has already done that for us. The difference in AUC from
our Stock and Tuned can be factored to give us a percentage of total increase in horsepower
and/or torque throughout the full run.

In digging a little deeper, the shape of the curve represents the manner in which the car builds
power. Ideally a tuner will strive to create the widest and flattest torque curve possible. This
creates linear and very predictable power delivery. Based on the size and efficiency of the turbo
this isn't always possible so compromises will come into play. As a generalization, smaller turbos
tend to spool quicker and build more low end to mid range power, while larger turbos will require
more time to build boost and deliver mid range with greater power at the top end.


In summary, Peak Gains alone don't offer nearly enough information on their own,
while Max Gains will offer a bit more insight, but ultimately Area Under the Curve
is what you need to focus on to really see the bigger picture!

[John@Dyno Spectrum]

All nice points, but I'm not such a fan of Area Under the Curve as a sole or even main metric. The counterpoint to area under the curve is that it is a good way for a tuner to sell torque gains when they are running out of turbo, and sometimes it is done when the power gains are meagre, and to be fair this is touched on above. The downsides are that if you have traction, the car initially feels eager with that fat midrange torque, but the reward in feel and engine note to rev it like an old NA M engine is further eroded. At worst it can feel like a TDI. If you are traction limited, the torque delivery in the midrange compared to the top end loses finesse because it is less consistent across the engine speed range. If you are octane or cooling limited, the massive midrange can mean that in hard use the performance ends up dropping closer to OEM levels. So it is possible to subtract from the overall balance, feel and consistency of the car by overdoing the midrange. Beefing up the area used when you are in gears that are not traction limited is nice, if the car can keep doing that repeatedly without pulling back the torque due to thermal efficiency, real or model EGT, knock.

In numbers, the consistency argument: range of torque from 3000 to 5000 RPM = 380 to 410lbft. Tuned example: 360 to 510lbft. The tuned car has five times the variation in torque in the midrange where you might be feathering the throttle in low and intermediate gears to find traction, assuming you aren't trying to get angle The tuned car has 25lbft increases in only 100 RPM in this range. Flat topping the tuned car at 450 lbft would give a nice "wet" option on a map switch, joining the rest of the tuned curve from 5400 to 7400 RPM, or perhaps exceeding it if you measured output over three back to back pulls to 180mph.

The German tuning approach is different, not better but arguably more mature but also staid. They take a more OEM approach of consistent power so they will quote a power curve that the car will achieve on repeated pulls. Competition between OEMs seems to be eroding this as they use more of their turbo capacity and rely on an initial sprint before pulling back. Previously they were holding back on the initial sprint so they could keep the power consistent for the distance.

I would be impressed to see data that shows that x back to back pulls result in the same curve and show the AFR and knock. Then tuners might get their dyno cooling as good as the black dyno and might tune for consistency a little more than a best curve and then let the ECU pull it back again.

[CobraMarty]

I have always been a proponent of AUC. I wish it was reported on every dyno graph. If the software calculates the AUC, why not post it.

AUC allows comparison between a very peaky high HP engine and a lower broad HP engine.

So post your AUC.

[VF-Engineering]

I would be impressed to see data that shows that x back to back pulls result in the same curve and show the AFR and knock. Then tuners might get their dyno cooling as good as the black dyno and might tune for consistency a little more than a best curve and then let the ECU pull it back again.

I posted a dyno plot with "back-to-back-to-back-to-back" (you get the idea) runs in our original article.

We log IAT's, EGT's, Lambda, Ignition Timing, etc... when we are doing development work, but typically not when just doing measuring. (we have two in-house dyno's, so calibration testing is done on our Mustang AWD-500, and measuring for comparables is done on the DynoJet as it's the favorite for bragging rights.)

That being said, our intention is not to be the leaders in the power gains by the numbers ... and what you see below is pretty darn consistent and repeatable which is what we prefer to strive for.

I'd rather have 90 ft. lbs. of torque over stock after five or more pulls, than 120 ft. lbs. that fades after two.

I believe I have AFR's from the sniffer for the following runs, but no logs unfortunately. That being said timing correction resulting in loss of power would be visually apparent in this many repeated pulls if it were indeed prevalent.

[VF-Engineering]

I have always been a proponent of AUC. I wish it was reported on every dyno graph. If the software calculates the AUC, why not post it.

AUC allows comparison between a very peaky high HP engine and a lower broad HP engine.
So post your AUC.

It's always different strokes for different folks.
I've never seen anyone post too much data, or provide soo much info that people say "ok, that's enough! We get the idea".


And what we consistently tell people is that dyno's are for measuring changes in power, not increase in performance. The only real way to measure performance is on the pavement with timing instruments. That's where it matters most.