HP differntial from engine to rear wheels

[glickster]

What is the "accepted" difference between HP measurements at the engine (flywheel) versus the rear wheels. I've heard different opinions ranging from 10 - 15% for a manual tranny and drivetrain, to 15 - 20% for automatics.

I'm trying to compute (guesstimate) what 209 rear wheel HP on a manual 968 equates to at the flywheel.

Any thoughts? Thanks.

[Eagledriver]

Those numbers look about right. I use 10 percent for my 915 equiped car. I think it also might be different on a dyno than on the street. There might be more or less friction on the rollers of a dyno because of the shape of the rollers the smoothness of the rollers and the differnt flex of the tires.

-Andy

[Bill Verburg]

There might be more or less friction on the rollers of a dyno because of the shape of the rollers the smoothness of the rollers and the differnt flex of the tires.

No rollers or tire loss, Rich uses a DynoPac that bolts to the wheel hub.

15% +/- pure guess but that would match the #s I got off Rick's Dyno.

Car almost done? Sorry I didn't get over to talk to you guys at LimeRock. Rick's 914 was awesome when I saw it buzzing around the track.

[Jack Olsen]

15% is the most commonly used figure. But of course, the only way to be certain is with an engine dyno.

[glickster]

Hi All,

Bill - I thought I saw you at LRP. Sorry we didn't get to chat.

Planning on a "test and tune" at WGI next week. 209 HP on the dynapak = 246 at the crank, about a 10HP increase over stock 968 spec(236), which is about right for just a header I suppose. In all other respects the engine is stock.

Advanced Watkins weather reports for 5/15 - 5/18, looks wet every day. I'm hoping the weatherman is wrong as it's impossible to "dial in" the car in the rain.

Ciao.

[chris_seven]

I have always been interested in the reported power losses on Chassis Dyno's and being a born again cynic I feel that the bigger the number the 'better value' we all seem to obtain from the rolling road session.

I would think that the basic physics should give a clue but i need to resort to SI Units to try to think it through.

An engine that devlops 209 BHP at the wheels is about 185kW.

If this is subject to 15% losses then somehow we need to have about 28kW of heat rejection.

Assuming black body radiation from the transmission housings and 75% of the losses going into the transmission I would guess that the transmission oil would burn quite quickly.

With Industrial gearboxes used on high speed test rigs typical estimates for single satge spur gearboxes would be 1 to 1.5% per mesh and these types of figure have been checked using either heat to oil or back to back testing.

The traditional bevel box that is used on a four square rig has reported losses of around 2% and most helicopter gearbox test rigs seem to have a prime mover that is rated at around 10% of the rating of the test box. This prime mover usually has to drive a couple of bevel boxes and the testbox.

I tend to think that typical automotive boxes should have losses in the 3-5% region but I am sure that Dana and other manufacturers will have much better data.

I think that 15% is not a loss but is a combination of losses and measurement errors.

Is it true that friction between the tyres and rollers is a loss? as the friction must produce a torque that is measured by the transducer in the chassis dyno. The error is more likely due to inaccurate speed measurement at the roller and the calculated Horsepower would be incorrect.

I think that this type of error is completely indeterminate and just adding 15% is misleading but could create a feel good factor.

A 'roll down' figure may help to remove some of these losses.

i have been reading about a chassis dyno that has been developed in New Zealand where the rear wheels of the car are bolted to the dynos using hub adaptors and speed measuring errors must be eliminated. Seems a good idea.

I guess that there may be other methods used to look at speed measurement errors between wheel speed and roller speed and companies such as Mustang will know far more about how horsepower is estimated.

[Dave at Pelican Parts]

The hub dynos have been around for at least a year or two. I've seen them used as portable dyno setups for various "shootouts" in several magazines.

--DD

[Dantilla]

How much difference is there between a 911 with it's compact drivetrain and a 968 with a torque tube?

Any ideas?

[Mr Beau]

Chris,

My contention is that the main difference is the intertia experienced on a chassis dyno. The SAE procedure calls for steady-state measurements with eliminate intertia, and also stabilize readings. A reading on a chassis dyno will be lower due to the inertia, but also inflated in some respects due to the lack of stabilized temperatures. Then there's the problem of friction depending on the type of chassis dyno.

[chris_seven]

Surely inertia is only a problem during accel and decel measurements.

Steady state measurments should not be affected by inertia and in fact interia is useful in smoothing out the peak to mean torques produced by a typical IC engine so that torque measuring systems can be appropriately sized to give accurate readings.

Low inertia dynamic dynos are will show significant torque change even within 1 revolution depending on the number of cylinders and this may be important during engine development but not when trying to establish baselines for comparison.

A four stroke single cylinder engine running at 20 000rpm will show at least a 200rpm speed variation within a single revolution and the mean to peak torque variation is very significant.

It is common to use a flywheel with an inertia of around 0.1kgm^2 connected to this type of engine to be able to manage this problem.

I agree that the inertia of a chassis dyno is much higher than that of an engine dyno but when this is inertia is referred to the engine through the gearing of the vehicle its influence is reduced significantly.

It is quite common that F1 engines to be tested for steady state performance with low speed water brakes and a flywheel. There is usually a speed reducing gearbox also present in the driveline.

Even the more modern AC dynamic dynos being used now have a driveline with around 20kgm^2 of inertia and when simulating down changes torque injection is used to simulate vehicle inertia. (F1 car about 50kgm^2 total)

Typical max torque of last seasons 3.0 litre F1 engine used to be around 300Nm. Torquemeters used on F1 engine test beds are typically rated at almost twice this value to ensure that they survive. This over rating clearly affects measurement uncertainty and hence accuracy.

With regard to friction I am not 100% sure that this is a 'loss'.

A race car typically has around 2-3% wheel slip when there is maximum grip. The prime mover must be able to supply sufficient force to drive this wheel slip and the torque measurment system must react this force.

The dyno is after all absorbs power and is a brake.

I still think that speed measurement inaccuracy is the main issue.

[Wil Ferch]

How 'bout we turn the question on its head ????

Who cares what the motor HP is? ... it's what gets delivered to the "ground" that matters...( think "system" vs. "component"). And guess what...voila...we have chassis dyno's that do just that.

IMO....chasing down what the motor makes is simply "bragging" rights when guys belly up around the barbeque and trade lies....

Ahem....one man's opinion...

- Wil ( flame suit on)

[Mr Beau]

Chris,

Some good points and lots of information. I think your post supports my contention that it's impossible to do a straight percentage loss from from engine to chassis dynos. You seem to be well connected in the industry. Might I ask what your day job is?

Will,

I know what you're saying, but it can make a difference. Lets say that Joe Blow is trying to set a new land speed record. He dynos his car on a Dynojet, and it makes 100 HP. He then puts on lighter wheels and replaces his steel flywheel with an aluminum version. The car is then dynoed and indicates 110 HP.

This same effect is why a car will have different power readings on a Dynojet depending on which gear is used. The power loss between the different gears in a transmission is essentially the same, so what you're seeing is the effects of inertia not being eliminated.

[randywebb]

you ignored convective heat losses

also, not all frictional losses in the gearbox will appear as heat during your test run

then there are fluid dynamic losses -- maybe people are calling that friction tho

I don't see trying to mechanistically account for every factor resulting in the engine dyno vs. road dyno power losses - just too complicated.

[beepbeep]

15% is exaggerated figure often used by dyno operators to make people happy.

Real world numbers are around 9% for typical 911. As chris_seven already explained, 15% of losses would mean heaps of heat to be dissipated from gearbox casing. And yes, all losses would be converted to heat (what else).

Frankly, I don't understand the eternal fascination with corrected hp figures. What reaches the wheels is only thing that matters. Otherwise you can say that you have 1000hp engine unfortunately tied to gearbox with 80% losses.

There is a very good dyno bench called "Rototest". Measures are taken by bolting hydro-brakes on wheel hubs. That way, tire-roller losses are eliminated.

It also allows you to "lock" the RPM's at certain setting and thus tweak the engine for static loading.

Classical "pulls" might give wrong results if pull itself is done too quickly.

[lateapex911]

The Dynopak is turning out to have potential as the best choice in "chassis dynos" , as it helps follow the #1 cardinal rule in any measurement: Absolutely consistant.

Eliminating the tire-->roller interface is key.

It should yeild higher numbers than a Mustang, for instance, but the actual loss vs crank numbers is variant on the drivetrain layout, the heat of the drivetrain, the oil in the drivetrain, the bearing type and condition, and so on.

A chassis dyno is oviously very useful for testing all those factors.

It's too bad a baseline run with the old exhaust wasn't conducted first.

Finally, never use the numbers that are "calculated" by the "roll down" technique favored by the dyno operators, as they are fundementally flawed.

[randywebb]

"all losses would be converted to heat (what else)."

- you are thinking in equilibrium terms not in terms of transients -- eql. will take a loong time to reach -- reread my comment above

Unless you plan to do calorimetry on the whole car, you won't be able to ascertain the heat loss.

- I'll agree re the 9% for a 911 tho.

Also agree re the odd fascination. It's one thing to use engine dynos to compare engines - I'm fine w/that.
Also they avoid the conversion approximations to most road dynos. But the hp at the tires is definitely the real metric of interest.

[Mr Beau]

I agree that HP at the tires is what really matters most, but with the intertia type dynos (i.e. Dynojet) that most people use, you will get different HP for each gear (lower in lower gears, higher in higher gears). So what's the real output?

SAE J1349 calls for steady state RPM with temps stabilized for 2 minutes. I've never heard of anyone doing this on a chassis dyno so immediately it is flawed to compare chassis to engine dyno figures.

[89911]

Nice article along these lines:


http://www.caranddriver.com/article.asp?section_id=4&article_id=8020&page_number=1

[afterburn 549]

I was say what Will F. said but in my own crude way..the only reason to know crank power is in a out of chaises run up ( say you are a engine re builder -tuner ) then you can get a base line to work from...what ya get to the ground is what counts.....ltmo

[deathpunk dan]

209 rwhp from a 968 is roughly ~240 or a bit more @ the crank. This is roughly what a stock E36 BMW M3 makes as well.