[TDIwyse]

This is something that was interesting to me and thought others might find it interesting as well.
Cross comparing different vehicles power/torque output on different dyno's is problematic, as not all dyno's will read the same, and even on the same dyno and the same location it can produce different results depending on the operator, and many other variables. Examples for how and why this is true:

http://www.jkm.org.uk/performance/dynotesting.htm

The belief by many people unaware of the factors laid out in this report is that because the car is run on a Dyno Dynamics rolling road the result should be consistent regardless of location up or down the country? This would be a valid statement if the same operator and cooling arrangements where used in each case. However the truth of the matter is that the results that you see on any Dyno are affected by a large number of variables; many variables are taken care of by the software (if it is told the truth) however a huge influence is the actual rolling road operator and the choices and actions that they make when running the car.

http://www.turbomagazine.com/tech/03...ine_dyno_dash/

[same car, same day, 5 different dyno's, 5 different results]

Turbo Magazine Dyno Dash Comparison Chart
Dyno ..............HP........TQ
Dynapack.......249.4...242.8
DTS ..............257.6...556.5*
Dynojet (Win).235.8...227.8
Dynojet (DOS).243.7...237.2
Clayton...........265.7...240.6
Super Flow......228.9...226.6

*measured at roller without accounting for gear reduction


And then there's the whole issue where some people like to translate the wheel power/torque numbers back to the crank to get a bigger number. Which is hugely problematic. And why would we care about anything other than what gets put to the ground? That's what really moves the vehicle...

http://www.modified.com/tech/modp-10...s/viewall.html

So rather than attempting to convert your vehicle's dyno-measured wheel horsepower to a SAE net horsepower figure using a percentage or a fixed horsepower value, you're far better off accepting the fact that these two types of horsepower measurements aren't easily correlated and forego any attempt at doing so.

http://speedtalk.com/forum/viewtopic.php?f=1&t=27174
[engine dyno to chassis dyno comparisons]

More dyno accuracy stuff: http://www.caranddriver.com/features...-lying-feature

"There's no really accurate way to get engine horsepower from a chassis dyno," says Matt Harwood, marketing coordinator for Mustang Dynamometer, a major supplier of chassis dynos in Twinsburg, Ohio. "

To prove his point, Dinan bolts to his Dynopack one of his 2003 Dinan M5s, heavily tweaked to make a claimed 470 horsepower at the crank (he expects about 415 at the wheels). With the hood closed and no external fan blowing air into the radiator, the car wheezes out just 334 horsepower at the wheels. An LCD data logger on the dashboard reveals the air-fuel ratio from the engine computer. Approaching redline, the BMW's computer richens the mixture all the way to 9.5:1 as the underhood temperatures soar.


Personally, since I care how my car performs on the road, I've been much more interested in dynamic performance on the road and ways to quantify modifications and power based on that. Using things like a GTech device, or the drag strip, or just recently the Android PerfExpert. Here's some comparisons from my car using these various methods.

Drag strip calculation [God's Dyno]:

Published professional reviews for the stock North American 335d
http://www.motortrend.com/roadtests/...ict/specs.html 14.2 sec @ 98.1 mph
http://www.caranddriver.com/reviews/...take-road-test 14.2 sec @ 100 mph
http://www.edmunds.com/bmw/3-series/...oad-test2.html 14.1 sec 99.1 mph

For this exercise I'm using the Android Torque App Dynomaster and it's dragstrip "calculator": https://play.google.com/store/apps/d...nomaster&hl=en

I like this calculator as it allows you to put in the coefficient of drag, frontal area and choose your drivetrain loss %. Which allows you to see how much hp is lost to wind drag as well as choose your loss to get back to an idea of crank power (which doesn't affect the wheel power number because wheel power is what matters on the road). It does need the total weight of the vehicle going down to the track to be accurate, and also assumes flat road (which on a certified dragstrip would be true) and calm winds.
The curb weight of the US 335d seems to be about ~3825 lbs:
http://www.edmunds.com/bmw/3-series/...est-specs.html
Curb weight, mfr. claim (lbs.) 3,825
Curb weight, as tested (lbs.) 3,804

Let's assume with driver (~190 lbs) the total weight is ~ 4000 lbs.
Plugging these numbers into the calculator shows predicted wheel horsepower numbers of: 213, 217, 218.1 wheel hp. All of these numbers are pretty consistent.

A lot of people use 15-20% drivetrain loss assumptions for automatics. Let's say we choose something near the lower/middle range like 17% drivetrain loss (the temp/altitude only impacts the wind drag results). That turns the Edmunds result into a crank power of 262.8 hp. That's really close to the manufacturer's claim of 265 hp. See attached picture for calculator screen shot with the calculations.

[TDIwyse]

So now let's look at what my 335d was putting down using the dash mounted GTech device (http://www.gtechpro.com/) to measure power. The attached shows a plot of the car stock, with JBD at 0% and with JBD at ~65%. For the GTech to be accurate you must have correct weight input and be on a flat road and have calm winds, and you also need to be able to pick up a signal for the rpm (I used the battery jump start connection under the hood as the 12V accessory power was too well filtered to pick anything up). Care was taken to ensure this was the case (used 4000 lbs total weight just like the dragstrip calculator method above). However, even with the GTech it will still under report power because of wind drag. Using the Dynomaster calculator (or this online calculator http://ecomodder.com/forum/tool-aero...resistance.php) and using the Cd and frontal area for the 335d you can see that for a 3rd gear pull at the peak power rpm the wind drag horsepower loss is ~ 15hp.
Taking the stock peak power number of 209 @ 3769 rpm, and adding 15 hp for wind loss, you get to wheel power numbers of 224 hp. That number is quite close ( Within ~2-3% ) of the wheel power numbers the calculator gives for the 3 online 335d dragstrip results. Interestingly, that's pretty darn close.

[TDIwyse]

Now, let's look at another comparison. This one is from awhile back where the car was running an Evolve remap with some custom stuff of my own and had all emission equipment. See attached for a GTech test showing 271.1 hp at ~4000 rpm, which for 3rd gear at that rpm would have a wind drag of ~19 hp. So GTech data predicted ~ 290 wheel horsepower. The dragstrip result for this configuration produced a run with 12.65 sec at 109.35 mph.

http://www.dragtimes.com/BMW-335d-Timeslip-25556.html
Let's plug those timeslip results into the Dynomax calculator using 3950 lbs (I had less weight at ~ 1/4 tank of fuel for that trip to the strip). Calculator predicts 295 wheel power and (using 17% drivetrain loss) 356 crank hp. Again, surprisingly good agreement between Gtech and dragstrip calculated wheel horsepower results (295 whp vs 290 whp).

[TDIwyse]

So, my GTech isn't working anymore. But I did get the Android PerfExpert app this last week, which is similar to how the GTech works (accelerometer based). Although this app is also smart enough to back calculate wind drag losses when the Cd and Frontal Area are entered correctly. So far, the app is really impressive and is very repeatable. However it doesn't give wheel power [Edit: This is incorrect... it does give wheel power, see later post] , just back calculates to crank power. I'm not sure exactly how they do that, but let's check its accuracy.

Let's take a look at some more recent dragstrip data with the new modifications: Ecotune with deletes plus my custom stuff.

http://www.dragtimes.com/BMW-335d-Timeslip-26508.html

11.84 sec at 119.95 mph.

Using the calculator at 3950 lbs (the car itself is a bit less weight than previously, but I had more than 1/2 tank of fuel for these runs to help with traction on the launch ... so according to my calculations the weight is basically the same at 3950 lbs) we get 370 wheel horsepower and, using the same 17% loss number, 445 crank horsepower.

The PerfExpert data taken a few days ago (it doesn't give wheel power, just crank power) predicted 435 crank horsepower for a 3rd gear pull [see attached]. So these two methods seem to correspond pretty good as well.

[TDIwyse]

Now let's look at a chassis dyno test. This was on a stingy Superflow dyno with the biggest rollers they make so big dually pickups can be run without having to remove their outside wheels. This test was done without any fans blowing on the car (worst case type of stress test … see the caranddriver dyno lying link where Dinan demonstrates hp impact when no blower is used on the BMW) and this dyno measured ~370 wheel hp on the first pull (a bit less on the 2nd heat soaked pull ... again, no fans at all being used). Again, really close to the dragstrip calculation method.

http://www.e90post.com/forums/showpo...&postcount=874

But remember from the 5 dyno comparisons in the Dyno Dash link, the Superflow was the stingiest dyno. And the place I dyno'd has a well known stingy dyno compared to other dyno's based on people I've talked to who have done pulls there.

So ... it seems to me that real world dragstrip data can compare reliably to handheld device measurements (Gtech, PerfExpert) and to stingy chassis dyno numbers.

However, there's a lot of people that seem to get overly inflated dyno numbers that DO NOT agree with real world performance on the road/track. For example, this place claims a 410hp tune on the 335d but it barely broke into the 12's.

https://www.youtube.com/watch?featur...&v=1OPOlelOOSc

12.94 sec at ~107 mph.

Let's use the dragstrip performance to do a sanity check on the claimed dyno results. Using the calculator (assuming a hefty 4000 lbs to give a bigger hp number, and the same 17% loss) predicts 280 wheel horsepower and 337 crank horsepower. That's a long way from the claimed dyno predicted power numbers.

I think a lot of hp claims people throw around online come from over inflated dyno's. If it's not backed up by other methods (especially on the dragstrip), it would be a good idea to take such claims with a grain of salt...

[floydarogers]

My head is spinning faster than your wheels.

[Stu In Blackpool]

A Dyno figure really is just that, a figure.
It needs to be backed up with other data to validate it and this date includes atmo pressure, humidity, intake temp, ambient temp, tyre pressures and strapping tension. There should also be three runs, back to back, within a few lb and bhp of each other.

Any Dyno that can't provide the above should be offering refunds.

The problem with trying to equate 1/4 mile times to horsepower is again, your comparing apples and oranges. A fast car is a car with a lot of area under the torque curve. It's perfectly common for a car to gain a lot of top end horsepower and be awful low down, or vice versa, lots of midrange torque but no major gain in top end power.

Again, data is everything, without all the facts, the numbers are of little use to you.

Example.
335d with peak torque of 580ftlb.
But is it only 580ftlb for 100rpm and under 450 everywhere else? Or is it above 550 almost everywhere. The later example is the faster car... But the original quote of "580ftlb" told you nothing about the area under the graph.

[TDIwyse]

Correct, a peaky engine horsepower curve would not be as accurate with the dragstrip calculation method. But this is about the 335d. It has a broad, nearly flat power curve over the rpm's where the transmission is shifting in a 1/4 mile run. So the power calculations off the dragstrip results are coming in nice and consistent and accurate with this powertrain. As verified with multiple different power measurement techniques.

It's best when theory can be verified with multiple independent approaches as opposed to relying solely on one method.

[TDIwyse]

Forgot one other aspect of the PerfExpert app that's really neat. It uses your smartphone and location to find the atmospheric conditions and uses this info to correctly alter the readings to normalized standards.

- Compute corrected power and torque according to the correction norm of your choice (taking into account atmospheric pressure, ambient temperature and air humidity). So you can compare measurements on different days. Available standards : DIN (Europe), SAE (America), JIS (Japan), CEE (Europe) and ISO (International).

[Yozh]

Thanks TDI, I know now how I shall be verifying my d's numbers when I do a retune. It's always nice to know "around" where the power of your car is.

I also wonder how the skill of the driver will affect the calculations of your real world results?

[TDIwyse]

Originally Posted by Yozh

Thanks TDI, I know now how I shall be verifying my d's numbers when I do a retune. It's always nice to know "around" where the power of your car is.

I also wonder how the skill of the driver will affect the calculations of your real world results?

The handheld devices are just so easy to use and makes testing things so quick and easy ... as long as they're repeatable and you follow a consistent protocol (all my Gtech data and the PerfExpert data are done on the same stretch of road and on days with minimal wind) you'll be able to quantify impacts to the mods you've done. And the PerfExpert's ability to use weather conditions to self correct to set standards is a really neat feature.

The driver can impact the dragstrip method. As well as if you can't hook up your wheels (traction issues). But with an automatic car with a broad power band there's not a lot of skill needed at the dragstrip.

[Puerto Rican 335d]

Thanks Tdi impressive

[Yozh]

Hey TDI, what correction norm are you using, cos mine only plots speed and G (acceleration) vs time. Your graphs on the PerfExpert look much more usable.

[Yozh]

Oh I think I got it, you use a dyno run instead of a timed rum. Gotta try that one.

[TDIwyse]

Originally Posted by Yozh

Oh I think I got it, you use a dyno run instead of a timed rum. Gotta try that one.

Dyno mode, correct. And I was doing 3rd gear for the pull, which isn't as good as being in 4th or 5th which are closer to 1:1 gear ratio. But you'd be really, really speeding using anything above 3rd on a road...

[Yozh]

Awesome, thank you. Now the challenge to find a good straight empty road...

[GottaBeKD]

Great read and info.

It would be nice if you can set the final drive ratio in the app to account for the different gearing. Should be a simple calculation and can be done in software.

Maybe you can email the app owner to add that in.

[TDIwyse]

Originally Posted by GottaBeKD

Great read and info.

It would be nice if you can set the final drive ratio in the app to account for the different gearing. Should be a simple calculation and can be done in software.

Maybe you can email the app owner to add that in.

When you set up the car in the app you need to enter the final drive ratio and the transmission ratio's of the gears you're doing testing in. Also need to enter your tire size, weight, frontal area and coefficient of drag. It uses all that to get the rpm info/wind drag/etc. to make the dyno chart.

It's quite a slick app.

[iaknown]

Very interesting info tdi.

I have a couple questions for you sort of related to this. What software/program do you recommend for tuning like you are on top of the Ecotune? I'm specifically referring to logging not your secret setup :-)

Also, do our cars have a pre-turbo EGT sensor or just post? I haven't gotten a chance to look into that yet....

[TDIwyse]

Only post turbo EGT's. The DPF canister has 3 of them if I'm remembering correctly. Pre DOC, Pre DPF, post DPF.

I really like the Bavarian Technic tool for detailed data analysis. It can see all kinds of stuff that's useful. Things like requested and actual boost, requested torque and actual torque, the EGT probes, injector timing, opening times, the 2 pre injection opening times, the main injection opening time, the 2 post injection opening times, the rail pressure requested and actual, Intake Air temps post intercooler, DPF stuff, EGR stuff ... Lots and lots of fantastic data!

You can also use it to engage the lift pump in the fuel tank, which I use when changing fuel filters, to flush the filters and prime the system to remove air so you're not running the HPFP dry when trying to start the car after a filter change.

And when logging info during testing runs, if you're not logging too many variables, it's quite fast at 3-4 samples per second resolution.

This is hands down my go-to setup for serious data collection and analysis. But it's kinda clumsy and more involved to setup with the laptop than the next option below.

My "everyday" application is Torque Pro on my Android device with a Blue Tooth OBD dongle. With it I can see EGT's, coolant temp, OBD speed (which is different than the speedometer), intake air temps at the MAF (haven't been able to find a way to read the post intercooler temps), EGR, Throttle position, etc, and you can customize your gauge cluster to your liking. And it's super easy to use. I get in the car, put phone in the holder, and hit the "Torque" icon and it pops up in the gauge display and automatically connects to the OBD device. One button push and it's up and running. See link below for how it's mounted and the gauge cluster.

http://www.e90post.com/forums/showpo...45&postcount=5

You can also do some "power" and "torque" and 0-60, 1/8th, 1/4 mile stuff with it as well, but I've found it to be far from good for that stuff. The update rate isn't fast enough to catch the rapidly changing data stream for some of those measurements and repeatability isn't good. The PerfExpert is far superior in this from my experience.

Hope that was helpful.

[iaknown]

Originally Posted by TDIwyse

Only post turbo EGT's. The DPF canister has 3 of them if I'm remembering correctly. Pre DOC, Pre DPF, post DPF.

I really like the Bavarian Technic tool for detailed data analysis. It can see all kinds of stuff that's useful. Things like requested and actual boost, requested torque and actual torque, the EGT probes, injector timing, opening times, the 2 pre injection opening times, the main injection opening time, the 2 post injection opening times, the rail pressure requested and actual, Intake Air temps post intercooler, DPF stuff, EGR stuff ... Lots and lots of fantastic data!

You can also use it to engage the lift pump in the fuel tank, which I use when changing fuel filters, to flush the filters and prime the system to remove air so you're not running the HPFP dry when trying to start the car after a filter change.

And when logging info during testing runs, if you're not logging too many variables, it's quite fast at 3-4 samples per second resolution.

This is hands down my go-to setup for serious data collection and analysis. But it's kinda clumsy and more involved to setup with the laptop than the next option below.

My "everyday" application is Torque Pro on my Android device with a Blue Tooth OBD dongle. With it I can see EGT's, coolant temp, OBD speed (which is different than the speedometer), intake air temps at the MAF (haven't been able to find a way to read the post intercooler temps), EGR, Throttle position, etc, and you can customize your gauge cluster to your liking. And it's super easy to use. I get in the car, put phone in the holder, and hit the "Torque" icon and it pops up in the gauge display and automatically connects to the OBD device. One button push and it's up and running. See link below for how it's mounted and the gauge cluster.

http://www.e90post.com/forums/showpo...45&postcount=5

You can also do some "power" and "torque" and 0-60, 1/8th, 1/4 mile stuff with it as well, but I've found it to be far from good for that stuff. The update rate isn't fast enough to catch the rapidly changing data stream for some of those measurements and repeatability isn't good. The PerfExpert is far superior in this from my experience.

Hope that was helpful.

Yes, very much. Thanks.

You have inspired me to keep my stock turbos and run meth lol....but I plan to run the AEM FIC as a piggyback to control everything. In theory I will have full control over fuel, boost or RPM based, meth, and even a safety if meth weren't to spray (have to do a little more research on this but the AEM has the ability to run 2 different maps based on a voltage input).

Hopefully I will get the time to implement all of this in the future.....

[TDIwyse]

Originally Posted by iaknown

Yes, very much. Thanks.

You have inspired me to keep my stock turbos and run meth lol....but I plan to run the AEM FIC as a piggyback to control everything. In theory I will have full control over fuel, boost or RPM based, meth, and even a safety if meth weren't to spray (have to do a little more research on this but the AEM has the ability to run 2 different maps based on a voltage input).

Hopefully I will get the time to implement all of this in the future.....

Looking forward to hearing about your progress. Please keep us posted. That AEM FIC looks really powerful. Think it was TurboEddie (sp?) who had been doing something similar with his 335d and the FIC.

If you go with the Bavarian Technic you'll need to open the correct engine control module to view parameters. I use the DDE7N47 module (they don't have a DDE7 M57 module). It's not obvious that's the one to open, and it doesn't have full compatibility, but it does give you a massive amount of visibility to the operating parameters on our vehicle.