transmission losses, the Universe and beyond...

I'm currently reading a (relatively prehistoric) book on turbochargers, from the seventies, last print 1984. An era in which there are probably people on this forum that were not even born yet

I had read it years ago, loaned from a friend, and found it badly written and not worth buying myself. But for $12 odd from the States, I thought 'wot the fek' and added it to another order as a filler.

This guy holds a few patents on turbochargers, so he wasn't exactly born yesterday (or even the day before)

I found it interesting how we share some 'pet hates', one of which being RR results. He ended up using accelerometers and vehicle weights as a means of approximating real horsepower, after being disgusted with the accuracy of the dynos at the time. Here's the interesting bit:

Engine out of the bay, measured on dyno Versus Engine measured in the bay on an accurate RR: 4% power difference, a bit more for autos

Not 20, nor 25%, hell no 30% for autos.

His point was that brand new cars (at the time) sold as '220hp' would typically show 150hp (flywheel) on a decent dyno. Harder to sell with such figures though, so everybody shut up, (but everybody knew as well)

Just an idle thought, regarding the 'padding' behind the scenes to make 'stock' cars produce the manufacturer's figures - hey, that proves the accuracy, right?

Not qute sure I'm followng the thrust of this post, John. Are you saying that he measured the power at the road wheels on a rolling road and then compared it to an engine dyno figure with the flywheel attached to a brake and found that the transmission losses were only 4%, or that he calculated the flywheel horsepower using a rolling road and then compared it to the results from an engine dyno and found that there was a 4% error between the two?

No-one seriously uses rolling roads for engine development in the OEM world anyway. Only for emissions, some calibration and engine NVH.

As for fudging power figures, in the good ol' U of K we have something called "power certification". In a controlled, independant test, any new engine design has to get within +/-3% of its stated power at the flywheel. If the engine achieves this, then it gets certified, but importantly, the manufacturer is allowed +/-5% to allow for manufacturing tolerances and "Friday afternoon" engines.

Therfore, if a manufacturer wants to, they can perfectly legally and without any fear of reporach, build an engine that makes 3% below the figure they were shooting for, but claim it has 5% more, so an engine with a headline power figure of 200hp might in reality only be making 184hp.

AFAIK, in the 'States OEMs don't have to do power certfication tests.

He claimed measuring transmission losses on manual transmissions in the region of 4%.

He's an old fool perhaps, heh...

And he would probably be right, the differential, propshaft and most importantly the tires account for the rest of the losses.

Differential, propshaft and tires are in the chain when measuring torque on a RR. How else could it be done? (there were no hub dynos back then)

I don't understand how people buy the theory that a 400bhp car needs 100bhp (or more!) to overcome friction inside the gearbox, shafts and tyres. It is so far off reality that it amazes me...

I don't understand how people buy the theory that a 400bhp car needs 100bhp (or more!) to overcome friction inside the gearbox, shafts and tyres. It is so far off reality that it amazes me...

Yea i agree John, so are people trying to say that if you bolted a Ford 1.3 engine (around 80bhp) into a supra, it wouldn't have enough power to turn the wheels?

Actually some people think that if you bolted a Cossie engine, it would need to got full throttle, full boost to turn the wheels of a well-mapped Supra with a single as long as it is auto. (25-30% of 600-700bhp)

I think a reality check is in order, lol...

I dont quite understand why transmission losses are calculated using %?

I mean say for example, you turn on your aircon, that will need say 5bhp more to drive the compressor, that load will remain the same whether you have a 300bhp car or a 600bhp car?

yes with the transmision you will get a rise in heat etc which are energy losses that will increase but sureley the calculation should be somehthing like, -40bhp & then -2% or 3% for increasing losses?

I dont quite understand why transmission losses are calculated using %?

because that would create even bigger figures for the more powerful cars (whose owners have spent loads more to 'tune' them)

----a cynic would say-----

I mean say for example, you turn on your aircon, that will need say 5bhp more to drive the compressor, that load will remain the same whether you have a 300bhp car or a 600bhp car?

explain that to someone who has just spent ten grand on a performance upgrade of his pride and joy. It's a top bhp figure he's interested in, so he can brag at the pub and sleep peacefully at night.

Torque curves and efficiencies and even 1/4mile figures just get in the way.

yes with the transmision you will get a rise in heat etc which are energy losses that will increase but sureley the calculation should be somehthing like, -40bhp & then -2% or 3% for increasing losses?

I like the percentages, because when the car is overfuelling and running like shite it makes (say) 200bhp on the rollers, which we adjust for altitude (10K feet) and ambient temperature (65C) and end up with 300bhp. Add to this 30% for losses, you get a neat 400 to keep the punter happy. See it as VAT on horsepower

Then when the fuelling is sorted out, it makes 230bhp on the rollers, but the overall 'adjusted' figure is increased by waaaaay more than 30bhp.

Everyone's happy.

lol fair enough, glad im not the only one confused by their calculations then.

Good articles here and here. The guy agrees with John but reckons if you have to express it as a % then it should be in the order of 15-17%, or 10%+10bhp. Note that he qualifies this by saying it is for cars of 100hp to 200hp. For more powerful cars, presumably the factor decreases further, down to 7-9% for cars with 400hp ish.

For info, SCA used 8% for Chis's recent 600hp dyno run.

Cheers Digsy, that'll make good bedtime reading.

Those figures sound much more realistic and as you say, should be decreased with higher bhp cars.

Just use other perfomance cars for benchmarks.. for example, an Enzo or McLaren F1, 650 bhp, weighs near to nothing (compared with a Supra) does a 1/4 mile in 11.0 @ 133mph secs.... now you can look at Supra dynos and the times/speeds these cars run and have a guess at the figures.

"Those figures sound buch more realistic and as you say, should be decreased with higher bhp cars." - I agree with this

Just use other perfomance cars for benchmarks.. for example, an Enzo or McLaren F1, 650 bhp, weighs near to nothing (compared with a Supra) does a 1/4 mile in 11.0 @ 133mph secs.... now you can look at Supra dynos and the times/speeds these cars run and have a guess at the figures.

 

"Those figures sound buch more realistic and as you say, should be decreased with higher bhp cars." - I agree with this

Oooh, ooh, I got a 129mph terminal - my times are worse though due to weight, tyre width, suspension, ability, etc.

-Ian

Terminal speeds at the 1/4 mile, now that's a decent indication of horsepower.

Oooh, ooh, I got a 129mph terminal - my times are worse though due to weight, tyre width, suspension, ability, etc.

 

-Ian

Do you have the times to go with that, dude?

...OK, I got off my @rse and found Chis' results in the TOTB thread.

This is interesting: I've banged his 1/4 mile figures into a (you guessed it) quick spreadsheet that tries to work out flywheel horsepower based on terminal speed, distance, time, average acceleration, rolling resistance and assumed transmission losses. If I have the maths right it ought to give you the power at the flywheel the moment the car crosses the line.

The big assumption is acceration being linear, which of course it won't be, but it's the best we can do with 1/4 mile figures.

Using Ian's TOTB results, and then fudging the tranny losses to make the flywheel horsepower equal his dyno run, the losses come out at 6.1%

Check it out.

Power spreadsheet - IanC 0.25mile.zip

Both G-Force and Surrey rolling road told me that you lose 10% through the tyres alone...

Thats interesting. So If I take their assumed 18% loss for manuals which gave me a spot on figure for my car, and minus the 10% loss for tyres, that gives me an 8% loss figure for a hub dyno maybe?

Which works out to be around 335 bhp flywheel) for my car from a reported 309 hub figure, which is far more realistic a figure than the 360 originally projected for a stock boost car with just a cat back exhaust.

Fascinating stuff

I like the way all these calculations are suprisingly close to the dyno results

I don't know why people are so intent on having 20%+ transmission losses despite all these numbers coming out If I went for 20% I'd have 725bhp lol - I don't think so

-Ian

Where have you had your car dyno'd before Ian?

Try the figures I suggest for losses, the 18% for a rolling road and 8% for Thor? See how they compare.

Damn - just as I thought we were getting somewhere.

Firstly, I made the schoolboy % error on my spreadsheet - luckily at low losses it doesn't make much of a difference. I've corrected that and re-uploaded it.

The second thing is that a dyno figure will have to account for both the losses through the transmission AND the losses through the tyres. The figure that the spreadsheet spat out is for driveline losses ONLY, as it calculates the rolling resistance for itself, based on the vehicle speed and mass. This means that you can't directly compare the 8% that SCA used on Ian's car with the 6.1% that the spreadheet is now spitting out.

What it does mean is that the TOTAL losses are looking more in the region of 10% for a car of Ian's ilk. Maybe still a tad on the high side.

But still a lot lower than Rolling Roads have calculated.

Damn - just as I thought we were getting somewhere.

 

Firstly, I made the schoolboy % error on my spreadsheet - luckily at low losses it doesn't make much of a difference. I've corrected that and re-uploaded it.

 

The second thing is that a dyno figure will have to account for both the losses through the transmission AND the losses through the tyres. The figure that the spreadsheet spat out is for driveline losses ONLY, as it calculates the rolling resistance for itself, based on the vehicle speed and mass. This means that you can't directly compare the 8% that SCA used on Ian's car with the 6.1% that the spreadheet is now spitting out.

 

What it does mean is that the TOTAL losses are looking more in the region of 10% for a car of Ian's ilk. Maybe still a tad on the high side.

What are his figures and from what rolling roads?

No I have had figures of 17% from Rolling Roads