Mystery Solved: Complete correlation between S55 Dyno's, Simulations, and Real World

[pkimM3r]

Originally Posted by Cl0ud7

in conclusion, we're paying for the same power output but tailored differently on the outside?

yes we have the same car. my butt dyno reads EXACTLY the same but you have higher output in a dynograph.

damn i just drove around my wife's 135i after being in e90m3 for a while and it is pretty torque and quick. cant imagine what the m4 feels like.

[Ricfutures]

Originally Posted by Keto

You had me at hello...

But man this thread pretty much killed whatever "stock value" Insoric may have on these forums.

Amazing information none the less!

[Powerslide]

I am not an engineer, but I nevertheless find highly-technical discussions to be incredibly interesting. The theories raised by many in this thread (Swamp, Boss 330 and others) of average power within a given, single gear over a specific rpm range seems highly plausible, especially as applied to the M4 vs. C7 corvette anectodal example (slightly lower in-gear average horsepower for given rpm spread for C7, which it compensates for by having slightly lower weight than M4).

So - this leads to my next question which those on this forum with more technical expertise than I can possibly answer - do the "formulas" or "modeling" that dynos use (either Dynojet, Maha, or both) necessarily assume a gradual, in-gear power increase as one would expect to see on a NA engine such as the S65 (where in-gear horsepower, whether at the crank or at the rear wheels, can vary significantly between say 6,500 rpm vs. "peak" horsepower at 8,300 rpm)? If the answer to this question is yes, then how can such a dyno possibly yield consistently accurate "corrected" horsepower figures for an FI engine which develops a "constant" amount of horsepower over a very broad rpm range (i.e., S55 with a stated crank horsepower being at a constant 425 hp from 5500 to 7300 rpm)?

To you engineers out there - thanks in advance for the clarification on this incredibly interesting topic (at least I find it interesting - I can read this kind of stuff for hours on end)!!!

[Diver]

Some things on my mind... I don't like relying on car magazine numbers, especially C&D. There is a lack of consistency in how they test, some use roll out, some do not. Some temperature correct and some do not. Then there is the usual thing with BMW turbo motors putting out more horsepower than stated spec. That is nothing to complain about, but a lot of us would like to know why. My personal theory is there is an undocumented overboost function that lasts about as long as a rear wheel dyno pull. I don't care, cause some of these bimmers go like stink!

[Ghostriderf80]

This is good to know. So aside from the cost of car I will need Akrapovic full exhaust, Tune, intake, intercooler if anyone makes it, H&R springs and BBS F1 wheels w bigger tyres. About another 20K on top of the car.

[Mandi90TT]

Originally Posted by Dexterspeed

This is good to know. So aside from the cost of car I will need Akrapovic full exhaust, Tune, intake, intercooler if anyone makes it, H&R springs and BBS F1 wheels w bigger tyres. About another 20K on top of the car.

Why would you 'need' those? Sure, they will make your M3/4 perform better, but certainly not needed, as it will perform quite well all on its on.

I mean, I understand the mod bug, and I planning on doing some of those things to my own M4, but I certainly don't claim to 'need' them.

[sward]

Originally Posted by Boss330

I'm just saying what MAHA also claims...

http://www.maha.de/single-roller-dyn...r-msr-1050.htm



Or, rather it seems I have been mistaken... If it's +/- 2%, then that should mean there could be a 4% variation in total, right? (It could be 2% low or 2% high from the actual number)

I wonder if operator error could be part of that high Pschlepp value on the Sport Auto run?

Isn't it 5? -2, -1, 0, 1 and 2

[P1 Motorcars]

Originally Posted by Boss330

And we have at least two different simulations, using the same software, that shows either engine performance just as BMW claims or shows under rating.

But only one of the two uses the real dyno charts, real world conditions, real car conditions, real weather conditions to match the simulation to real world vBox results, then use the real world vBox results to generate a vBox Dyno which (not coincidentally) matched the Dynojet numbers. Our simulations were identical to the vBox numbers. Our vBox numbers were identical to the magazines. And our vBox Dyno showed 420 whp (an aspect of this article nobody has wanted to discuss). We looked at this from multiple angles and got consistent correlation from every angle.

Powerslide]So - this leads to my next question which those on this forum with more technical expertise than I can possibly answer - do the "formulas" or "modeling" that dynos use (either Dynojet, Maha, or both) necessarily assume a gradual, in-gear power increase as one would expect to see on a NA engine such as the S65 (where in-gear horsepower, whether at the crank or at the rear wheels, can vary significantly between say 6,500 rpm vs. "peak" horsepower at 8,300 rpm)? If the answer to this question is yes, then how can such a dyno possibly yield consistently accurate "corrected" horsepower figures for an FI engine which develops a "constant" amount of horsepower over a very broad rpm range (i.e., S55 with a stated crank horsepower being at a constant 425 hp from 5500 to 7300 rpm)?

For the most part, chassis dyno's use a standard physics formula to calculate power: F = M * A (Force = Mass times Acceleration). The "force" is calculated and using a very simple conversion formula is converted to HP. The "mass" is the weight of the roller in the roller dyno. Then you simply measure the acceleration from the car speeding on the roller and out comes horsepower. That's how a Dynojet works. This approach makes no assumption and makes no attempt to calculate crank horsepower or any type of "in-gear power increases." It's simple F=M*A. So the simple answer to your question is "NO" -- chassis dyno's don't do this because it's not needed to calculate horsepower.

The Maha takes it a step further. The Maha first measures the wheel horsepower using a very similar technique as the Dynojet. Once maxRPM is reached, the Maha turns on an eddy current motor and applies a very small load to the roller. The Maha then measures the rate of deceleration of the roller and develops a drive train "power loss" profile in this manner. By re-adding the power lost due to drive train losses, the Maha calculates what they believe is crank horsepower. It's rather ingenious. It's worth noting, the just as the Maha calculates drive train losses by measuring deceleration, the Insoric RealPower module does the same thing by different means.

Originally Posted by Ricfutures

But man this thread pretty much killed whatever "stock value" Insoric may have on these forums.

We weren't trying to. But can you imagine our disappointment after analyzing the product selling for $5200 and realizing it has some significant design shortcomings and may only cost $20-$50 to manufacture?

[04RC51]

Like watching E9X owners swing their purses at dyno operators for returning numbers shaming their current cars. Two months later, even deeper denial.

[Boss330]

Originally Posted by P1 Motorcars

But only one of the two uses the real dyno charts, real world conditions, real car conditions, real weather conditions to match the simulation to real world vBox results, then use the real world vBox results to generate a vBox Dyno which (not coincidentally) matched the Dynojet numbers. Our simulations were identical to the vBox numbers. Our vBox numbers were identical to the magazines. And our vBox Dyno showed 420 whp (an aspect of this article nobody has wanted to discuss). We looked at this from multiple angles and got consistent correlation from every angle.

And by "real dyno charts" you mean the Dynojet that measures 50-60 more WHP than the MAHA...

A MAHA recently measured 453PS (crank)/365PS (wheels) on a stock M4, which is within 5% of factory stated 431PS
On the same dyno a tuned M4 measured 439,1PS/433HP (wheels) and 520PS/512HP (crank). With a tuned M4, the MAHA measures quite similar WHP that the Dynojet does on a stock M4...

Aren't the MAHA results "real dyno charts"?

[solstice]

Look at this E92 M3 Maha dyno. Prad 301 whp. Again close to 60-70hp short of what we see on the F8X MAHA dynos. Calculated crank hp is about 50 hp short. I'm using the MAHA run we have data from, I.e the 465 ps crank run. The 453 hp run is just by what someone said. No data provided for that run.

So a 10hp rated difference of these cars keep showing 60-70 hp difference time after time on different dynos. Yet it is continued argued the true difference is 10 hp. Yes, denial does comes to mind.

[Boss330]

Originally Posted by solstice

Look at this E92 M3 Maha dyno. Prad 301 whp. Again close to 60-70hp short of what we see on the F8X MAHA dynos. Calculated crank hp is about 50 hp short. I'm using the MAHA run we have data from, I.e the 465 ps crank run. The 453 hp run is just by what someone said. No data provided for that run.

So a 10hp rated difference of these cars keep showing 60-70 hp difference time after time on different dynos. Yet it is continued argued the true difference is 10 hp. Yes, denial does comes to mind.

Good find

This S65 MAHA dyno run records a 415,7PS/410HP at the crank. Very close to the factory claim of 420PS/414Hp. So, this MAHA dyno run seems pretty accurate as regards crank PS/HP.

We have two MAHA measurements for the S55 that we know of so far, one we have a dyno graph for and one we don't. One has a crank rating of 465PS/458HP and the other 453PS/446HP.

465PS/431PS = 7,8% (34PS more than adverticed)
453PS/431PS = 5,1% (22PS more than adverticed)

The variation between these two dyno measurements are:

453/465 = 2,6%
(Maha operates with a accuracy of +/-2% for their dynos, so this difference is perfectly within tolerance)

INSORIC:

449PS/431PS = 4,1% (18PS more than adverticed)


Apart from the Sport Auto test, they are all within 5% of the factory number. Two measurements (1 MAHA and INSORIC) have on average 20PS more than what BMW claims, while the third (Sport Auto MAHA) has 34PS more than BMW claims. This means that, using the MAHA dyno results, the real world difference in crank PS between S55 and S65 is around 33PS to 45PS (11PS + 22PS = 33PS and 11PS + 34PS = 45PS).

It's only the Dynojets that get 60-70HP (or 15-16%) more than what BMW claims...

But, I'm not 100% sure that we should rely to much on any of these dyno measurements... The difference between the Dynojet and the MAHA is further evidence towards the difficulty in knowing which dyno to believe and surely a measuring tool that has this large variation seems a bit flawed...

[Black Gold]

Originally Posted by Boss330

Good find

This S65 MAHA dyno run records a 415,7PS/410HP at the crank. Very close to the factory claim of 420PS/414Hp. So, this MAHA dyno run seems pretty accurate as regards crank PS/HP.

We have two MAHA measurements for the S55 that we know of so far, one we have a dyno graph for and one we don't. One has a crank rating of 465PS/458HP and the other 453PS/446HP.

465PS/431PS = 7,8% (34PS more than adverticed)
453PS/431PS = 5,1% (22PS more than adverticed)

The variation between these two dyno measurements are:

453/465 = 2,6%
(Maha operates with a accuracy of +/-2% for their dynos, so this difference is perfectly within tolerance)

INSORIC:

449PS/431PS = 4,1% (18PS more than adverticed)


Apart from the Sport Auto test, they are all within 5% of the factory number. Two measurements (1 MAHA and INSORIC) have on average 20PS more than what BMW claims, while the third (Sport Auto MAHA) has 34PS more than BMW claims.

It's only the Dynojets that get 60-70HP (or 15-16%) more than what BMW claims...

that is not correct, because you are making up the drivetrain losses. no one knows what they are for this car.

like solstice said, this car is putting down 50+ whp more to the ground relative to the e9x on basically every measurement we have seen. whether that is all from underrating, or some from underrating and some from lower drivetrain losses is something we cannot quantify at this time.

BUT, I think its fairly clear that its not all due to less drivetrain losses.

[Boss330]

Originally Posted by KennyPowers

that is not correct, because you are making up the drivetrain losses. no one knows what they are for this car.

like solstice said, this car is putting down 50+ whp more to the ground relative to the e9x on basically every measurement we have seen. whether that is all from underrating, or some from underrating and some from lower drivetrain losses is something we cannot quantify at this time.

BUT, I think its fairly clear that its not all due to less drivetrain losses.

I am not making up drivetrain losses... I use the MAHA calculated crank numbers

And as we have seen, the MAHA seems to have a dyno model dependent drivetrain loss figure (PSchlepp). For instance, on the Sport Auto run Pschlepp was 120PS at the most, while the tuned M4 only had 82PS on that MAHA dyno.

This difference seems to come from the method they employ for measurements and where it seems the weight and rotational force of the drums are included in Pschlepp. I am the first to say that this is something I am not an expert on. But it seems CanAutM3's observations on the wide rage of Pschlepp values between different MAHA models must be correct after all. I mean, there is no way that the F8x drivetrain can range from 80-120PS in drivetrain losses due to factory tolerances... This means that the Pschlepp figure on the MAHA must include more than simply drivetrain losses.

[Black Gold]

I'm referring specifically to your dynojet comment.

You can't make a comparison from dynojet whp to crank without applying a drivetrain loss factor.

IMO it's best not to even try.

But, the point remains that some combination of more power and probably less drivetrain loss relative to e92 results in the f8x putting 50+ more whp down than the e92 on a dynojet. This has been shown on no less than 8 different dynojets so far.

I can't comment on maha because they are not common here and I don't have experience with them

[solstice]

There could very well be a 465 - 415 = 50 hp difference between the S65 and S55 at the crank but 70 hp at the wheels. One of the reasons could be that the S55 make peak power earlier in the rpm range where losses are less another could be less losses overall. 50 and 70 hp are both very significant. A +30 hp under rating is very significant 465ps - 431ps. In every scenario there is significant under rating and most significant is the hp that matters the whp. The S55 is a beast and the F8X even beastlier

[Boss330]

Originally Posted by KennyPowers

I'm referring specifically to your dynojet comment.

You can't make a comparison from dynojet whp to crank without applying a drivetrain loss factor.

IMO it's best not to even try.

But, the point remains that some combination of more power and probably less drivetrain loss relative to e92 results in the f8x putting 50+ more whp down than the e92 on a dynojet. This has been shown on no less than 8 different dynojets so far

Ok, sorry about the confusion on my part and a reply that didn't really answer what you pointed out

Here's how I assumed the 60-70hp higher dynojet readings:

If we use take the factory rating of 425hp and use a drivetrain loss of between 10-15% (assumption based on historical data and previous info from manufacturers, but NOT F8x specific so it can be off). The OP (P1) also uses a 10% drivetrain loss for his CarTest simulations. So at least that's consistent with the OP's assumptions.

425 x 0,85 = 361,25WHP
425 x 0,90 = 382,5WHP

The Dynojet measures 420-430WHP (AFAIK), which is around 40-70WHP higher than the THEORETICAL whp number if the engine had 425HP at the crank. And the OP's CarTest simulations also end up with 420-430whp. Add the 10% drivetrain losses (that the OP used as a input in his CarTest simulation and which he also used to make a input on crank hp in CarTest) and we end up with a crank HP of around 470-475HP.

So, depending on drivetrain loss we (and the OP) end up with a crank HP of:

470-475HP (10% drivetrain losses as per OP's assumptions and simulations)
495-500HP (15% drivetrain losses, which probably is on the high end here)


The OP has discussed why he choose 10%. If you have further questions about that, or why he used drivetrain losses to prove his point, those questions should be directed towards him Unless of course, you also question the OP's simulation results (which relies on a assumed 10% drivetrain loss to create crank hp from whp measured on the Dynojet).

[P1 Motorcars]

Originally Posted by Boss330

And by "real dyno charts" you mean the Dynojet that measures 50-60 more WHP than the MAHA...

That's not what I meant. You were comparing two different simulation results. I was pointing out that one of us used theoretical data with theoretical conditions. We used real data and real conditions. Our simulations also came up with the correct real world results.

Originally Posted by Boss330

The OP (P1) also uses a 10% drivetrain loss for his CarTest simulations. So at least that's consistent with the OP's assumptions.
...
And the OP's CarTest simulations also end up with 420-430whp. Add the 10% drivetrain losses (that the OP used as a input in his CarTest simulation and which he also used to make a input on crank hp in CarTest) and we end up with a crank HP of around 470-475HP.

There is a handful of misunderstanding here.

When inputting wheel horsepower into CarTest (like I did with Dynojet), drivetrain losses are ignored by CarTest. The 10% losses I discussed were specific to the Maha crank (not wheel) dyno numbers I placed in CarTest for the second (Maha) simulation. I pointed this out in my article.

Our simulations didn't "end up with 420-430 whp" -- they STARTED with the dyno chart (which was ~420 whp). Before CarTest can run a simulation and estimate 1/4 mile and 60-130 results, it must subtract the drivetrain losses. Since the Dynojet results already account for drivetrain losses, there's no extra calculations required in CarTest. The drivetrain losses are ignored.

Originally Posted by Boss330

The OP has discussed why he choose 10%. If you have further questions about that, or why he used drivetrain losses to prove his point, those questions should be directed towards him Unless of course, you also question the OP's simulation results (which relies on a assumed 10% drivetrain loss to create crank hp from whp measured on the Dynojet).

10% drivetrain loss isn't added to CarTest and then subtracted for Dynojet. 10% drivetrain loss is ignored all together when you use Dynojet results in CarTest. Therefore none of the results for Dynojet rely on this 10% loss.

I hope this clears this up.

Originally Posted by Boss330

INSORIC:

449PS/431PS = 4,1% (18PS more than adverticed)

Insoric uses the same physics formulas as CarTest, and the vBox Dyno. I say that because if you google search the subject, you will see there's really only one set of formulas to use. Given the missing data in Insoric, the results won't be nearly as complete as the vBox Dyno. In fact, due to the missing calculations in Insoric, the horsepower output will only go up and get higher if they start to include them. So why not discuss the vBox Dyno results?

The vBox Dyno should be far more accurate because it doesn't miss any of these calculations. Here's a side-by-side comparison between Insoric and vBox Dyno:

Calculations:

• WHP due to Acceleration
• WHP due to Slope (absent in Insoric)
• WHP due to Friction (rolling resistance) (Insoric missing minor elements used in these calculations)
• WHP due to Aerodynamic drag (wind resistance) (Insoric missing key elements used in these calculations)
• WHP due to wheel weight (equivalent mass) (Insoric missing key elements used in these calculations)

Inputs:

• Tires: Size as MMM-RR-DD
• Tires: Pressure PSI (absent in Insoric)
• Tires: Number of tires (currently all wheels/tires are assumed the same...DING)
• Tires: Tire weight
• Tires: Tire weight distribution (absent in Insoric)
• Tires: Wheel weight
• Tires: Wheel weight distribution (absent in Insoric)
• Vehicle: Vehicle weight
• Vehicle: Driver weight
• Vehicle: Gasoline weight
• Vehicle: Weight distribution (front/rear) (absent in Insoric)
• Vehicle: Frontal area (absent in Insoric)
• Vehicle: Drag coefficient (absent in Insoric)
• Weather: Temperature
• Weather: Pressure
• Weather: Humidity (absent in Insoric)
• VBox: Velocity
• VBox: Height/Slope (absent in Insoric)

Look over that list and you can see Insoric calculations are low because they don't include all of the forces acting on the car (or don't include them correctly). If they added these missing forces, the estimated power would only go up

[Boss330]

Originally Posted by P1 Motorcars

The vBox Dyno should be far more accurate because it doesn't miss any of these calculations. Here's a side-by-side comparison between Insoric and vBox Dyno:

Calculations:

• WHP due to Acceleration
• WHP due to Slope (absent in Insoric)
• WHP due to Friction (rolling resistance) (Insoric missing minor elements used in these calculations)
• WHP due to Aerodynamic drag (wind resistance) (Insoric missing key elements used in these calculations)
• WHP due to wheel weight (equivalent mass) (Insoric missing key elements used in these calculations)

Inputs:

• Tires: Size as MMM-RR-DD
• Tires: Pressure PSI (absent in Insoric)
• Tires: Number of tires (currently all wheels/tires are assumed the same...DING)
• Tires: Tire weight
• Tires: Tire weight distribution (absent in Insoric)
• Tires: Wheel weight
• Tires: Wheel weight distribution (absent in Insoric)
• Vehicle: Vehicle weight
• Vehicle: Driver weight
• Vehicle: Gasoline weight
• Vehicle: Weight distribution (front/rear) (absent in Insoric)
• Vehicle: Frontal area (absent in Insoric)
• Vehicle: Drag coefficient (absent in Insoric)
• Weather: Temperature
• Weather: Pressure
• Weather: Humidity (absent in Insoric)
• VBox: Velocity
• VBox: Height/Slope (absent in Insoric)

Look over that list and you can see Insoric calculations are low because they don't include all of the forces acting on the car (or don't include them correctly). If they added these missing forces, the estimated power would only go up

And yet, the INSORIC has consistently been very close to manufacturers claimed HP, on a wide variety of vehicles. Regardless of what any of us think about the S55 being underrated or not, it seems even more unlikely that every manufacturer states the wrong hp numbers. The real world tests for the INSORIC so far indicates a system that consistently get's a very good estimate of whp and drivetrain losses.

As regards your list of inputs and calculations:

I was in contact with INSORIC a month ago, and they replied to me on my question if aerodynamic losses also are accounted for (Cd and area)?

This is the reply I received:
It's calculated from the temperature and air pressure, air density = humidity. CW / drag coefficient, acceleration

So, according to INSORIC they use both density/humidity and the cars aero drag to calculate aerodynamic losses. Taking away some key elements of the claims against the accuracy of the INSORIC system.

[ASAP]

Seeing some new races on video, I can now nearly guarantee a stock M4 isn't running 1 hp over 400 WHP (dynojet).

[Boss330]

Originally Posted by ASAP

Seeing some new races on video, I can now nearly guarantee a stock M4 isn't running 1 hp over 400 WHP (dynojet).

Are these runs on YouTube?

I agree with you, and here is an example that I think also must question the accuracy of the Dynojet numbers:

I looked at the Boss 302 Laguna Seca yesterday. That has very similar 1/4 mile times and trap speeds as the F8x (12,3s @ 114-117mph).

Edmunds online had that on a Dynojet and measured 416whp (Ford claims 444hp at the crank). 416whp would mean a drivetrain loss of just 7,4%...

With it's slightly lower weight and higher average power, the F8x should surely have given the Boss 302 LS a good beating if the F8x really has 420-430whp :

[ASAP]

Originally Posted by Boss330

Are these runs on YouTube?

I agree with you, and here is an example that I think also must question the accuracy of the Dynojet numbers:

I looked at the Boss 302 Laguna Seca yesterday. That has very similar 1/4 mile times and trap speeds as the F8x (12,3s @ 117mph).

Edmunds online had that on a Dynojet and measured 416whp (Ford claims 444hp at the crank). 416whp would mean a drivetrain loss of just 7,4%...

With it's slightly lower weight and higher average power, the F8x should surely have given the Boss 302 LS a good beating if the F8x really has 420-430whp :

Here you go...

FBO Protuned N54 135i (6mt running on Euro 102 Ron gas which here in the states is 96 Octane lol). That 1 series isn't 1 hp over the 400 whp right there. What happens to the M4 with the DCT and it's 7600 rpm redline and larger turbos? lol