Looking To Match Gearbox Ratios To Powerband

[smckeown 1]

I've been getting more and more interested in this area. I have read a lot of posts on here and some bookmaterial.

 

What I don't seem to be able to find material on (or work it out myself) is how to actually map a power curve with a set of gearbox ratios to inform you which ones complement the powerband more. So the general consensus is that you'd want a close ratio box that wont drop you too far down the rev range that you are outside the power band. But there also seems to be the area that some engines have larger power bands, and i'm not sure how to make the link to available ratios/FDs to that specififc engine.

 

I was hoping that building a graph where you have both power band and different ratios available to pugs on the same graph would be a useful insight. I've got as far as being able to map the %changes of different ratios and map that to the rev range, but not to include the FD. Maybe I just need to use that as a multiplier to my existing graph. But if anyone has knowledge in this area, then i'd like a good discussion about differing theories and what's good practice in this area etc.

[smckeown 1]

Right I think I might have some answers. Luckily I had Graham Bell with me Didn't realie there is a good section o nthis in his book.

 

The initial theory seems to be that you work out the percentage changes between the gear sets to work out their suitability, the lower the percentage the less revs will drop.

 

Then there's the theory about the changes in actual axle torque when changing gear. This is where you can bring in the final drive to compare actual axle torque comparisons between gears and FD.

 

But what he doesnt explicitly mention is mapping gear seats with certain final drives to torque or BMEP ranges.

 

So, I propose that for an engine that's peaks (i.e. has a narrow bmep range) you would want to go for closer ratios so you dont stray out of the narrow bmep range. However, if you have a broader bmep range then surely you would also want a wider set of ratios, otherwise you don't benefit from the wider power band ?

 

Anyone care to comment on this theory ?

[sutol 0]

Did you ever draw graphs at school ?

Then make up a graph on a piece of paper and after trying the car out and noting power surges and road speeds, fill it in.

It will show you where to change, where the power overlaps are,and if you need to change gearbox.

 

Sometimes a lower diff is needed which brings all the ratios closer together avoiding the need to swap cogs in box.

Edited January 17, 2007 by sutol

[smckeown 1]

Did you ever draw graphs at school ?

 

You obviously haven't seen many of my posts then, my middle names are Excel and graph

 

I have posten on this area before, I understand the concept of axle torque etc. It's the theory behind closer and wider ratios matched to the power band i'm in need of input. I have listed the best place to shift based on axle torque before on this forum

Edited January 17, 2007 by smckeown

[sutol 0]

Yeh, see what you mean.

I'm old fashioned and use a stopwatch on the road to gauge upshifts etc.. This method tells be if I'm using the power to best advantage, crucial in dragracing.

 

Spose technology is overtaking me, untill I get the Mi in that is..

[Rippthrough 98]

A lot will depend on the time it takes you to shift as well as the actually ratios.

[smckeown 1]

I stayed up until 1am last night calculating and graphing the wheel torque differences between different ratios; and then the optimum shift points, resulting in the wheel torque curve through the gears. It was a useful expervice, and at the very end just before I was going to publish the results I realised i've used the standard torque figures rather than my cars..damn!

 

I also read more into a link i've posted before, and he states this:

 

"A more rigorous way of doing this is to graph horsepower vs. velocity in each of the gears. If power in one gear drops below the horsepower of the next gear at a particular MPH, then that MPH is where you should shift, otherwise shift at the redline.

 

I leave as an exercise for the reader the following: predicting shift points based on engine torque, RPM, and gear ratio gives the same results as predicting shift points based on power and vehicle velocity. "

 

So i'll post up both areas of analysis when I get more time

Edited January 18, 2007 by smckeown

[smckeown 1]

These were the figures I went through last night. These are just useful to compare the available pur rations and final drives. It still doesn't get into the theory of matching ratios to an engine's torque/bmep characteristice. I'm assuming the velocity vs bhp does that..

 

 

Edited January 18, 2007 by smckeown

[Hilgie 16 2 Cars]

All looking good.

 

But general rule of thumb is that Naturally Aspirated engines need a short ratio gearbox (because you don't want to drop out of the powerband, max torque etc) and that Forced Induction engines have a wide powerband and can use longer ratio gearboxes.

[Rob Thomson 6]

That graph doesn't tell you anything other than what you already know; a 1.6 box has a lower first gear but taller final drive than a 1.9/Mi16 hybrid.

 

What you really want to do is plot the torque at the wheels for each gear (vertical axis) against vehicle speed (horizontal axis). You'll have five curves on one graph, and where they overlap will show you the best points to change gear for maximum acceleration.

[Craigb 2]

You'll have five curves on one graph, and where they overlap will show you the best points to change gear for maximum acceleration.

 

I've been playing around with My ap22 g meter, and as far as i can see. the best point to change unless you have a massive drop off in torque at the the top end of the scale , is at the rev limiter in every gear .

 

Purely because the lower gear accelerates the car faster.

 

Sean I have emailed you an output from the last run from the AP22 , it shows acceleration and speed against time , you can clearly see the g forces in each gear , the speed curve and the gaps where the gear changes are . Maybe you could post it for discussion .

The way i see it is if you have a narrow power band , you need a very close ratio box, if you have a wider power band the all you have is a more flexible package. Can accommodate mistakes more easily .

 

Fixing gear ratios in my opinion is based around what top speed you want to be achieving at full revs and working out from there , what suits the speed discipline/ track you are competing in .

 

Theory says a 4.8 for me (road rallying) is the best max acc / speed (119) , but to be honest the 4.4 i have , i find to be a better compromise (129)

[smckeown 1]

That graph doesn't tell you anything other than what you already know; a 1.6 box has a lower first gear but taller final drive than a 1.9/Mi16 hybrid.

 

I disagree. Yes it's similar to the ratios difference grahp; but it shows the 'actual' power at the wheel differences. Plus it includes the shift points and leading to what power is achieved in each gear. This is more useful for comparing different cars (with or without diff ratios)

 

What you really want to do is plot the torque at the wheels for each gear (vertical axis) against vehicle speed (horizontal axis). You'll have five curves on one graph, and where they overlap will show you the best points to change gear for maximum acceleration.

 

Yes as stated I need to graph velocity and power onto the same graph to bet something useful. That's next week's job

 

 

Sean I have emailed you an output from the last run from the AP22 , it shows acceleration and speed against time , you can clearly see the g forces in each gear , the speed curve and the gaps where the gear changes are . Maybe you could post it for discussion .

 

Cheers, I'll have a look at that when it arrives, my email is playing up at mo.

 

 

Just to confirm, what i'm trying to learn about is to be able to prove mathmatically and via graphs what gearsets are more suited to a particular engine's power band. I've never seen such a graph to date.

Edited January 19, 2007 by smckeown

[Rob Thomson 6]

I disagree. Yes it's similar to the ratios difference grahp; but it shows the 'actual' power at the wheel differences. Plus it includes the shift points and leading to what power is achieved in each gear. This is more useful for comparing different cars (with or without diff ratios)

 

The graph doesn't tell you any more than glancing at the different ratios for three seconds.

 

I find this a bit hard to explain because I find this really obvious (physics, generally, is a piece of piss) but you need to completely forget about changing gear at peak torque, which is what your graph shows.

 

Let me try a couple of analogies...

 

1) Diesels. Every bugger and their armies are forever raving about the 'immense torque' of their diesel engine and how petrol engines are old-skool and past their sell by date because of this. What they forget is that flywheel torque is a meaningless pile of old bollocks, and they need to think about torque at the wheels. And because diesel engines turn so slowly they need very high gearbox ratios which neatly screws that up. What really matters is that they have a very good spread of power.

 

On the other hand, a high-revving petrol engine will have sod-all torque but much lower gear ratios that overall will generate just as much torque at the wheels, but will need more gear changes to keep it in its useful power band.

 

The flywheel torque figures for comparable diesel and petrol cars will be infinitely different, but the power figures will be the similar.

 

2) Cycling. You're into bikes, right? I couldn't give a f*** about mountain bikes, but do like road bikes. A couple of years ago the best road racers were Lance Armstrong and Jan Ulrich. Armstrong is a little bloke who pedals all day at cadences of 120 or more. Jan Ulrich is a big bastard who pedals at 80 rpm. In terms of torque, there's no comparison. Ulrich could almost certainly twist those cranks much, much harder than Armstrong because he's got big bulky leg muscles. But Armstrong made up for that by spinning faster and using lower gear ratios for a given speed. In terms of power they're very similar riders.

 

And that's what I'm trying to get across, that flywheel torque is a meanless figure by itself. You have to think about revs, and when you think about flywheel torque and revs you're actually thinking about power. It's the spread of power that accelerates a car quickly.

 

Plot power against vehicle speed and it'll tell you everything you need to know.

[niklas 1]

In my opinion the following aspects defines what gear ratios to opt for:

 

1. Power band. In what range do you develop usable power? Say between 4000-6500, then you should opt for having each gear drop the revs by approc 2500 rpm.

 

2. Track. This is the hardest part and specific for each track.

a. What is the highest speed you'll ever use on that specific track? There's no point in having ratios that can get you past that top speed, that is just a waste of performance..

b. What are the corners like? You have to opt the ratios to always be able to have a usable gear for each corner. This is often a compromise with point 1.

 

Rob: I fully agree with you regarding power/torque. Torque is one of the most mis-used words today. Torque != broad powerband!

I can easily produce 1000 Nm of torque with a 1m bar (yes, I'm that fat; 100kg * 1m = 1000Nm), but I still can't accelerate very fast at all

Edited January 20, 2007 by niklas

[bales 1]

I might be looking at this too simplistically but I have the A. G. Bell on 4 stroke tuning like you have, and have read all the section on gear ratios.

 

Surely it is a matter of recording your torque readings at specific intervals through the rev range from your mapping graphs. Then where ever peak torque (or peak BMEP) is or the rev range that has the highest average values of this e.g something like 4900rpm - 6500rpm the from your upshift points you can calcualte what the engine will drop too and make sure that the engine doesnt fall out of the specified range?

 

I guess that it is more complex than that but that seems a relatively simple way of getting a good gear set.

[smckeown 1]

i made a mistake in graphing that you shift at peak torque, silly as I know this is incorrect. I have posred before that you shift more like when the current gear generates more wheel torque than the next gear would if you were to change. This works out more or less at peak BHP.

 

G Bell does mention hweel speed (velocity) does need to be considered but he doesnt explain it very well.

 

I'll spend more time mid week looking at this when im bored in my hotel room...