Drivetrain losses (ad nauseam...)

[DRIFTIN]

the 47kw figure is irrelevant.. regardless of the figure you are going to end up with a situation where, if you say drivetrain loss is constant, that the vehicle can potentially produce no power at the wheels (or a negative amount!). Even if the drivetrain loss on a std WRX is 30kw, and not 47kw, I don't see it as realistic that with that power level it is going to stall and not move at all.

[Denver]

why not?

if a motor isn't producing enough power to drive the driveline, hows it going to move it?..

there was a show on tele a while back, one of those stupid scientist shows where they had to make a snow mobile to move some logs, one of the machines wasn't producting enough power from the motors to move it along, the effect, it stalled and didn't go anywhere despite revving the box out of it..

so its entirely possible..

[Belly_up]

ok, my 2 seconds of thinking says that with 2 meshing gears, the friction between teath is going to be the CoFriction * Applied force, which will be proportional to torque. So increasing torque will increase frictional losses, and assuming gearbox speed is kept the same, increase power losses. increasing power by increasing RPM's will have the same effect, as power is force * distance / time.

Other people have quite rightly pointed out that all these losses are frictional and will manifest themselves as heat. And thus if there was 47kw being lost in your gearbox and diff, they should melt.

The largest electric bar heater you can plug into your normal houshold outlet is 2.4kw and is pretty efficient at turning that power into heat. now imagine *20* of those heaters heating up the 10L (combined diff & gearbox volume, being generous) of oil, and imagine how quick it would overheat. of course that amount of loss is only going to be generated at full power (if i'm correct and it is related to torque), and indeed if you do run cars at full power for extended periods (top speed runs etc) then they do tend to overheat these components.

It would be possible to measure the temp increase over a given time of a gearbox at a specific torque on a dyno (dyno is good as it decreases the cooling effects of air on the 'box), if you knew the mass of the box, and rough make up of the internals (ie X kg of steel, Y kg of ally case, Z kg of oil and a specific heat capacity for each).

for a really rough example, lets assume that a gearbox weighs 35kg (alloy supra box rough weight) and that most of that weight is in the steel internals rather than the alloy case, and also ignore the weight of oil in the case. assuming the steel has a specific heat capacity of about 0.5 J/g/C, and that we are inputing say, 30kw or 30,000 joules / second of energy into our box as heat. so energy / SHC / mass = temperature rise

30000 / 0.5 / 35000 = 1.7 degrees.

we are putting in 30000 joules per second, so our 35kg of steel should rise in temp by 1.7 degrees per second, ignoring losses to the atmosphere.
So after 60 seconds (at full power mind you) the box would be at 100 degrees above starting temp, say 125 degrees (very generously starting with the box at ambient).

I don't know, i guess that seems resonable plausible, doens't it? considering that there would be a lot of heat transfer to the atmosphere, better conduction from the oil and alloy case etc...

so i guess that might be a reasonable figure... but then if you take the accepted figure of 30% drivetrain loss (humor me for a minute), then a 600fwkw car would be putting 200kw of heat into the drivetrain. if 2/3rds of that was gearbox (same assumption i made above), then the same hypothetical gearbox would be at about 200 degrees after 20 seconds at full power. does anyone with high powered racecar experience know how hot the boxxes get?

ok well as you can guess, that ended up being a bit more than 2 seconds of thought, and there was a hell of alot of simplifications going on there, but it does give some rough figures to think about.

All corrections and thoughts welcome

Adam

[Mini_Me]

edit: said what belly up said

[Belly_up]

to add to that, when you look at industrial stuff (or even not so industrial stuff like chains on pushbikes), they are quoted with a percentage efficiency, not a static figure. ie, if something is 97% efficient at tranferring power, then it is losing 3%, as heat and a bit of noise. so if it's tranferring 1kw, 0.03kw will be lost. 1 million kilowatts, 30000kw lost.

[Supercrown]

Drivetrain toss.......

Yep-tyre to roller interface is the culprit.

Even a 1st year engineer should be able to see that claiming %losses through a manual gearbox is ridiculous.

I can just imagine the meeting where a team of engineers submits a new drivetrain design proposal to management that has a power transmission efficiency of 70%...............

[bugle]

i think this is about right.

(as alway, mspaint is king)

I think you'll find the graph would look more like this

[Steve Whitney]

power at the wheels = power at the flywheel - 47

my near stocker cordia puts 80kW at the fronts.... does this mean i have 127kW at the engine?? i wish!

as others have shown, the power loss from engine to wheels would have to be a varying & sliding percentage depending on the car, its mechanical setup & how much power it puts out.

[260BC]

................................

as others have shown, the power loss from engine to wheels would have to be a varying & sliding percentage depending on the car, its mechanical setup & how much power it puts out.

Plus a few other things, like slip and friction on the rollers as others have pointed out, but yes.

Anyway, chassis dynos are not suited to giving power figures than can be treated as nearly accurate. The giving and quoting of such figures, which itself incorrectly gives the impression of accuracy, has developed into a customer driven business. The customer wants a dyno sheet, he will pay for it, we'll give it to him. You want more power? OK, here you go.

You want some sort of accuracy? Stick the donk on an engine dyno.

Wonder what prompted this topic in the first place? My dick is bigger than yours?

[DRIFTIN]

a thread on rexnet prompted this one.. everyone had their own 'correct constant' for loss to the wheels.

[Belly_up]

you could measure it, but no-one would bother.


2. Remove Engine from car
3. Connect motor of known and relatively constant output power to gearbox input (ie, electric motor)
4. Connect hubs to hud dyno
5. Run motor at various input power levels and measure output
6. Input power - output power = losses.

Step 1. would be to calibrate said hub dyno by connecting directly to said motor.

But as richard points out above, no-one really cares how much losses there are in the drivetrain (at least not enough to try and measure it). People who are serious put the engine on an engine dyno, and everyone else is happy to believe that they are loosing 30% of their power to the drivetrain as it makes their penises seem larger.

[tandy ass]

why not?

if a motor isn't producing enough power to drive the driveline, hows it going to move it?..

there was a show on tele a while back, one of those stupid scientist shows where they had to make a snow mobile to move some logs, one of the machines wasn't producting enough power from the motors to move it along, the effect, it stalled and didn't go anywhere despite revving the box out of it..

so its entirely possible..

Think of it this way - If you remove the spark plugs and turn the front of the motor by hand, perhaps with a breaker bar, do you think that would equal 47kw?

I brought up the heat issue in a thread ages ago, again using this hypothetical 47kw, as already posted above, thats a fuckload of heat. The gearbox has what, 2 or 3 litres of fluid in it. How long does it take your kettle, at 2400 watts, to heat up 3 litres of water? 5 minutes? Lets make this 47,000 watts, it would boil the water in seconds.

It is NOT a fixed constant. Period.

Edit - doh! others already posted better explanations!

[Wobbler]

I think you'll find the graph would look more like this

Good to see someone can use paint properly!
:p

[Ben Wilson]

I know it's not really evidence as such, but, at a dyno day here in Canberra last year, an LS1 Commodore and a V-TEC Civic both produced 21% less power at the wheels than they were claimed to produce at the flywheel.

[BigMuz]

Exactly 30 per cent. Always. For every car.

Hope this helps

Muz

(I am with Ustasa fro once :D But in my experience it is easy to change by putting softer tyres on the car to lose much more power, or to use skinny hard tyres to gain heaps. WHO CARES anyways....)