Welcome to the North American Subaru Impreza Owners Club Saturday August 30, 2014
Home Forums WikiNASIOC Products Store Modifications Upgrade Garage
NASIOC
Here you can view your subscribed threads, work with private messages and edit your profile and preferences Home Registration is free! Visit the NASIOC Store NASIOC Rules Search Find other members Frequently Asked Questions Calendar Archive NASIOC Upgrade Garage Logout
Go Back   NASIOC > NASIOC Technical > Normally Aspirated Powertrain

Welcome to NASIOC - The world's largest online community for Subaru enthusiasts!
Welcome to the NASIOC.com Subaru forum.

You are currently viewing our forum as a guest, which gives you limited access to view most discussions and access our other features. By joining our community, free of charge, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content and access many other special features. Registration is free, fast and simple, so please join our community today!

If you have any problems with the registration process or your account login, please contact us.
* Registered users of the site do not see these ads.
Reply
 
Thread Tools Display Modes
Old 01-08-2009, 01:46 AM   #1
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default Optimal Shift Points and Choosing Mods Smartly

This is not one of my usual in-depth, critical looks at a new part or tuning option. This is simply me working through the data I have collected and thinking about that data in terms of how to get the most out of what I have now and how to make beneficial changes in the future.

My initial intent was simply to calculate the optimal shift points for the data I had. It was practice, really, because the dyno data I have is for a hardware setup I don't even have on the car anymore. But, I haven't dynoed the new setup yet.

Ok, so the first thing that I had to do is to figure out, as a percent of the maximum torque, how much torque AT THE WHEELS was being made as I rowed up through each gear. So I made a pretty graph (click to make bigger):


While it's interesting to look at how much torque you lose with each shift, that graph doesn't tell you anything useful about when to shift, so it has to be drawn with respect to wheels speed:


Now we're getting somewhere. Optimal shift points are where you see one line cross the next. What the graph is showing is that, at that point, you can upshift to the next gear and have the same amount of torque at the wheels, which is what you want. the lack of intersection between the 1st and 2nd gear plots means that the gear ratios are so far apart that redline comes before the ideal shift point. For my car, at the time it was dynoed (and remember, your car and my car now could be completely different), the optimal shift points work out as follows:

1st: Redline
2nd: 6200RPM
3rd: 6100RPM
4th: 6100RPM

Mission accomplished! Since I had slogged through all the numbers to make that happen, I decided to see what other interesting bits of data I could mine out of it. One of the things that I don't like is having to wring the engine all the way out to the fuel cut to get anywhere. Not to mention, in the kind of driving I do (hill country, TSD, RX, etc), you're rarely able to just pin the throttle down and hold it there for long enough for the tach to sweep from 3k to 6k. What's the implication of running in the 3k to 5k rev range (which is what's MUCH more likely to be used in real life). Presuming I upshift at 5k, I'm taking a 30-50% hit compared to shifting at optimal, depending on gear. Ouch.

Now we get to the fun part. What can I do about this? Well, first of all, anything I can do that just linearly adds more power (say more displacement, N2O, etc), isn't going to change the situation at all. Anything that adds torque above peak is just going to make the situation worse. Anything that subtracts torque from the top end will move my optimal shift points lower, but it'll do so artificially by making the top end suck more and making the car slower overall. The productive options are to move the peak toque lower without reducing top end torque or to boost torque below peak torque without sacrificing the top end (or at least not much).

I started exploring in that direction specifically because I have an interest in the GrimmSpeed Intake Manifold Spacers. The IMSes lengthen the IM runners by 8mm. In theory, this should boost the torque below peak torque, not move peak torque at all, and reduce the torque above peak torque. Now, the increase below peak torque sounds like a good thing but the decrease above peak torque worries me. If the low end increases enough that my optimal shift point moves below the onset of the high end reduction, I'm golden. However, if I have to run up into the reduced area a bit to hit my optimal shift, things get A LOT more complicated. At that point, you need to find the difference in the area under the curve. I'm sure that phrase brings back nightmares about Calc 1 for a lot of you. Sure does for me. Basically, if the change can boost the low and mid enough for a wide enough rev range, you'll still end up being faster even though your high end has fallen off (assuming you didn't end up with an extra shift in there). The advantage to this is 1) explosive torque whenever you touch the throttle even at normal driving around RPM and 2) you don't have to flog the engine to get somewhere, saving a bunch of wear and tear on the engine.

Now, BRR just posted a thread about his experiences with the IMS. Sadly, the way he collected the data makes it impossible to draw any hard conclusions from it, but it certainly at least provides a platform for speculation. The data he posts shows a clear and increasing gain from about 3.7kRPM on down and a clear and increasing loss from about 4.8kRPM on up. So the low end gain looks strong but the high end hit looks like a real mofo. That would definitely be in "area under the curve" land for wither or not it was a net loss or gain. It would also move the shift point down a lot. In fact, it might be enough to cause enough extra shifting to cause a net loss even if the area under the curve were bigger. However, there's an interesting thing being overlooked here. My peak torque is MUCH higher than his; 4250RPM for me vs 3600-3700RPM for him. If the pattern for my change is the same, and the theory says it should be, I should see a boost to a wider range of lower-than-peak RPM as well as pushing the worst of the loss higher. It might be enough to make it a clear win for me. Sadly, I'd actually have to buy the parts, install them, and dyno them to be sure.


Again, this wasn't a rigorous analysis of the hardware to be sure. However, I thought it might be interesting to you guys to see the work and thought I put in before I decide on a mod.
* Registered users of the site do not see these ads.
williaty is offline   Reply With Quote
Old 01-08-2009, 03:56 AM   #2
chazly413
Scooby Specialist
 
Member#: 151079
Join Date: Jun 2007
Chapter/Region: MAIC
Location: College Park, MD
Vehicle:
2005 2.5RS RBP
motec system exhaust-RIP

Default

Great data and a good write up. I'd like to see it, though, on a stock or relatively stock engine.
chazly413 is offline   Reply With Quote
Old 01-08-2009, 04:03 AM   #3
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Get your car on a dyno and I can help you work out the numbers.
williaty is offline   Reply With Quote
Old 01-08-2009, 05:37 PM   #4
V8Slayer
Scooby Specialist
 
Member#: 111565
Join Date: Apr 2006
Chapter/Region: W. Canada
Location: Winnipeg
Vehicle:
1999 Legacy 2.5GT
2007 Legacy GT Spec-B

Default

Great wright up Williaty!! Thanks for the info.
V8Slayer is offline   Reply With Quote
Old 01-08-2009, 05:51 PM   #5
Back Road Runner
Scooby Guru
 
Member#: 60082
Join Date: Apr 2004
Chapter/Region: MWSOC
Location: Minnesota
Vehicle:
2004 Forester STI
Silver

Default

Yep, this is an important issue when modding a car.

I think most folks have little concern for it on a car of near stock trim. In many cases, redline or near redline is generally the best point. However, it heavily depends upon the engine. For example, think of a turboed engine. You look at the stock WRX or similar, and you get massive torque around 3k rpm that rolls off deeply on either side.

Picking on Wall of TVs:


The torque curve is vastly different then the NA engine. Shift points will vary with it, especially since so much of the torque is focused around 3k rpm with heavy drops on either side.
Back Road Runner is offline   Reply With Quote
Old 01-08-2009, 06:04 PM   #6
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

That is a completely hilarious dyno plot!
williaty is offline   Reply With Quote
Old 01-08-2009, 06:43 PM   #7
sniper1rfa
Scooby Specialist
 
Member#: 141040
Join Date: Feb 2007
Chapter/Region: NESIC
Location: Designing things
Vehicle:
07 2.5i wagon
UGM

Default

What the....

I don't know a whole lot about engines, so i'll ask:

What would cause such an enormous drop in torque like that? Turbo is way too small?
sniper1rfa is offline   Reply With Quote
Old 01-08-2009, 06:45 PM   #8
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

Turbo is optimized for quick spool is unable to flow enough to maintain engine loading on the top end. A bigger turbo wouldn't drop off like that on the top end, but would take forever to spool up, leading to lower torque in the places you actually use it in a street car.
williaty is offline   Reply With Quote
Old 01-08-2009, 10:22 PM   #9
Back Road Runner
Scooby Guru
 
Member#: 60082
Join Date: Apr 2004
Chapter/Region: MWSOC
Location: Minnesota
Vehicle:
2004 Forester STI
Silver

Default

Yes, it is hilarious but also amazing at the same time. It's pretty neat what he accomplished. This was on E85 though, so it could be pushed a bit more then pump gas.

Yes, the reason why it spikes early and drops so fast is because it's the stock, tiny TD04 turbo. It's actually sort of on the small side for a 2.5L engine, however, it provides great mid range torque.

The point is, torque comes in different shapes. As one tunes their car, they will modify the torque band. With this comes the need to readjust shift points.
Back Road Runner is offline   Reply With Quote
Old 01-09-2009, 08:58 AM   #10
LetItSnow
Scooby Specialist
 
Member#: 38222
Join Date: Jun 2003
Chapter/Region: Tri-State
Location: Cicero, NY
Vehicle:
'10 &'12 Foresters
another Miata convert

Default

A lot of it depends on what you're doing with your car when your foot's on the floor, too... If you spend your time in second gear in autocross, you don't want all your power to show up late. I wonder what that curve would look like in a gear-to-gear graph like the one williaty showed above?

This does raise reason to ask "Where do I get the claimed +5hp from this part?" If it's at 5800RPM, and it costs you power up to 3500RPM, it might not be what you want.

This same thing applies to suspension. I don't think a lot of people work out what they're going to get with their aftermarket springs, struts, and swaybars. Popular part combinations are popular for a reason, but sorting out how your car will end up in turning balance, resonance, and ride are all factors that can be worked into choosing a custom setup.
LetItSnow is offline   Reply With Quote
Old 01-09-2009, 09:25 PM   #11
Back Road Runner
Scooby Guru
 
Member#: 60082
Join Date: Apr 2004
Chapter/Region: MWSOC
Location: Minnesota
Vehicle:
2004 Forester STI
Silver

Default

Thought I'd toss in my car here. I plotted "Unmodded" stock (just I-Speed flash) and "Modded" (TBS, IMS, I-Speed).

The plotted range is from 2450 rpm to 6050 rpm in 225 rpm increments, convenience of the dyno graph labeling mainly.



One thing you should notice is that the stock plot has zero overlaps. This means you must shift near redline every time not to drop in power. Realize I only plotted to 6050 and not 6500 or 7000 which would extend the curves further. Where the lines stop is 6050 rpm. I just want to be clear on that. So, in stock forum, you're pretty much shifting near redline always. With the modded setup, there is overlap on all of them, indicating that I would have to shift a little under 6000 rpm every time to stay optimum.

Also note where the power is higher and lower in the rev range in each gear. Where does the modded setup show benefit? Where does the stock setup show benefit?

This is something I've stated to automatically do as I got accustomed to the new power band. I can just see on paper now why I do it.

So is this overlap bad? Well, sort of. If you are drag racing and winding through the gears, you'd like to maximize what you can. In this sense, you want to make peak power near redline and always shift at redline. My modded setup would be a poor drag setup, and I would fair better stock. Notice I never take advantage of the low end torque banging through the gears.

For normal driving or other sports like auto-x/rally-x, it's a completely different story. Here, you want a wide power band that is broadly useable. You may run from 2k to 6k in 2nd gear over 90% of the course. You want good power everywhere and not a need to shift into a lower or higher gear. Normal driving is also done mostly between idle and 3000 rpm, so this is where you want the main brunt of your power band, basically where you use it more. Watching shift points gets important if you step to extremes and could potentially make the stock gearing a bad fit for the engine power band.

I'll quickly do up Wall of TVs' old graph for fun and show that shift behavior.



His looks a bit different. I was expecting a lot of overlap but rather quite the opposite. Now this is only to 6050 rpm again, so feel free to extend the lines mentally to 6500 or 7000 rpm. I guess one thing to note with the peaky nature, at least where the peak lies is to notice what part of the power band is being used. Compare relative % of torque at various speed points. Compared to the broad, flat response of the NA, the relative torque is lower through the gears as that engine uses the drop off range of the torque band through most of the reving.

In this sense, you want to see something different, a very peaky top end. You want most of the torque at the upper end so you maintain high relative torque and area under the curve through the gears. This is why big turbo cars work at the drag strip. When you're only using between 5000 and 7000 rpm, you kind of want all your power right there.

I'll whip up a peaky setup to show this in a sec...

tino's GT40R tune:
http://forums.nasioc.com/forums/show....php?t=1470621



Notice the high relative torque used in each gear. When you focus towards the top, that's what you get. However, in this specific example, you don't see clean shift points. He basically has to run to redline (7000) and then accept the step drop in torque at the next gear. Now I will state, this isn't accurate. I am using the 5sp gearing in this example. The 6sp would show different and offer closer gearing. This is really just for reference and how power band affects shifting/peak power output.

Last edited by Back Road Runner; 01-09-2009 at 10:04 PM.
Back Road Runner is offline   Reply With Quote
Old 01-27-2009, 02:03 PM   #12
LetItSnow
Scooby Specialist
 
Member#: 38222
Join Date: Jun 2003
Chapter/Region: Tri-State
Location: Cicero, NY
Vehicle:
'10 &'12 Foresters
another Miata convert

Default

What happens if we add speed as a function to reduce those curves? If <this> holds any truth, ...

Quote:
...AeroDyn's GaryEaker points out that horsepower requirements to reach a given speed go up with the cube of velocity. At 200 mph it takes eight times as muchpower to push the car through the air as it does at 100 mph. If your car needs 80 rear-wheel horsepower to run 100 mph, for it to see 200 mph requires 640 hp.
After all, we leave the dyno from time to time, yeh?
LetItSnow is offline   Reply With Quote
Old 01-27-2009, 02:14 PM   #13
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

While airspeed will affect actual acceleration, it doesn't change torque at the wheels at all. So it's entirely irrelevant to this discussion.
williaty is offline   Reply With Quote
Old 01-27-2009, 02:57 PM   #14
Back Road Runner
Scooby Guru
 
Member#: 60082
Join Date: Apr 2004
Chapter/Region: MWSOC
Location: Minnesota
Vehicle:
2004 Forester STI
Silver

Default

Aerodynamics is a big factor in energy usage, but yes, it doesn't have a factor between the engine and tires. However, if you're looking at acceleration and time and energy involved, then yes, you can look at aerodynamics as being decently important. The question comes down to how much we can improve it. For example, my flying brick of a Forester has a frontal area around 29-35 sq.ft.(seen more then one value for this) and a drag coefficient of 0.37. The area is tough to reduce without some chopping, but drag can be reduced a bit. The Impreza sedans are more around 26 sq.ft. and a drag coefficient of 0.33. For reference a Z06 Corvette is 22.3 sq.ft. and 0.33 (from aero package for downforce) or 0.28 base C6.

If you're curious:

Drag Force = 1/2*MassDensityFluid*Velocity*Velocity*DragCoefficie nt*FrontalArea

Mass density of air of 1.3 kg/m^3 or 0.0805 lb/ft^3

Velocity is ft/sec, so 1mph = 1.47 ft/sec

Let's say my Forester is traveling at 65mph. What's my drag force?

Fd = 1/2*0.0805*95.6*95.6*0.37*35/32.2 (32.2 ft/sec^2 converting lb -> lbf)

Fd = 150 lbf.

For reference, an Impreza will only have about 100 lbf applied to it at 65mph, so a 33.3% savings of that type of drag do to it's physical design.

Basically for ease:

Fd = 0.00125*Velocity*Velocity*DragCoefficient*FrontalA rea

That way you just plug in the car speed, the coefficient of drag and frontal area.

Also note that frontal area and drag coefficient are linear properties. Half one and drag force halves. Also note that velocity is squared. Half velocity, and drag force is cut to 1/4, i.e. driving slow saves money.

Also note that
Work = Force*Distance

Well, the Force is our DragForce (Fd), and the distance is how far we drive. This is energy continuously being eaten up.

I'll make a small note that this isn't the only sources of drag and work loss. All forms of friction including internals of the machinery (engine, transmission) and even the tires rolling along the ground are sources of drag and energy loss. Aerodynamics is simply a part of it.

We can convert actual torque values at the various gears back down to the lbf applied at the tires. We can then determine the theoretical maximum speed of the car when the drag force becomes high enough to equal the push force by the engine at the tires in contact with the road, essentially terminal velocity of the system, in a simplified sense because we ignore other forms of loss.

Last edited by Back Road Runner; 01-27-2009 at 03:02 PM.
Back Road Runner is offline   Reply With Quote
Old 02-05-2009, 01:41 AM   #15
williaty
Scooby Guru
 
Member#: 71092
Join Date: Sep 2004
Chapter/Region: MWSOC
Location: Delaware County, Ohio
Vehicle:
2005 2.5RS Wagon
Regal Blue Pearl

Default

So, I just got the dyno results back for the EL headers. Hop on over to that thread to see the difference in actual plot.

Interestingly, the only thing that changed, shift point wise, was to drop the shift out of 3rd 200RPM to 5900RPM. 1st remains above redline and 2nd and 4th remain at 6100RPM.
williaty is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Is there an optimal ignition point for STi? bboy Factory 2.5L Turbo Powertrain 16 12-02-2007 05:25 PM
WTB: Pivot PIVSSL-P Shift Lamp and Sparco Smart pedals Modifying Private 'Wanted' Classifieds 2 05-31-2006 02:39 PM
Help! Optimal shifting for WRX to clock < 6sec vulcan Factory 2.0L Turbo Powertrain 8 10-17-2001 12:56 AM
optimal shift points for wrx imprezario Factory 2.0L Turbo Powertrain 2 06-26-2001 08:47 PM
('93-'01) What are the optimal shift points for saving gas? myko011 Impreza Forum 4 05-17-2001 12:36 AM


All times are GMT -4. The time now is 04:36 AM.


Powered by vBulletin® Version 3.7.0
Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Powered by Searchlight © 2014 Axivo Inc.
Copyright ©1999 - 2014, North American Subaru Impreza Owners Club, Inc.