[TheAssassin]

So I posted this in my build thread http://www.d-series.org/forums/showcase/157028-my-d16y8-turbo-build.html but I thought people would want to read it without having to go through my whole thread.

Beating a Dead Horse: The Great Torque vs. Horsepower Debate

So this thread by ddd4114 made me want to calculate the average horsepower that my car makes since average horsepower means more than peak horsepower.
Calculating average horsepower would have been easy if I had the formula for the line on the dyno plot because then I'd just take the integral of that line to find the area under the curve and divide that value by the rpm range (in my case 7200-2000=5200). Except I didn't have the formula for this line and didn't know of any programs that could generate it. Since I was just wanting a rough estimate of the average horsepower I decided to just transcribe the dyno plot onto a sheet of graph paper and enter in the horsepower values into Excel with their respective rpms. I used the trapezoid rule as opposed to the rectangle rule, just so I wouldn't be too far off.
So the trapezoid rule:

∫f(x)dx≈(b-a)*[(f(a)+f(b))/2]
or
(2200-2000)*[(32+37)/2]=6900

I went up by increments of 200 starting at 2000 (where the dyno run started) up to 7200 (where it ended).
Using Excel's handy autosum feature I was able to quickly calculate the sum of all these values: 691600
This obviously isn't my average horsepower lol I needed to divide this number by the rpms, in this case it was 5200
691600/5200=133 average whp from 2000rpm to 7200 rpm
If there's an easier way to do this, cool, but this is how I did it and it makes sense. I think it would be interesting if people would list their average horsepower along with their peak horsepower. In addition, you could find your average horsepower for any rpm range, such as the rpm range that you're in while drag racing e.g. 4800-7200.
Tell me what you guys think, have any of you ever calculated your average horsepower?

 

[transzex]

start with average torque.

 

[ddd4114]

start with average torque.

For once I sort of agree.

You'll have a lot more accuracy if you take points on the torque curve. However, you'll need to associate an rpm with each torque point to calculate power afterward.

 

[JBrettHowell]

It's all pretty simple if you think about it in NASCAR terms:
Oversteer is when you hit the wall with the rear end of the car.
Understeer is when you hit the wall with the front end of the car.
Horsepower is how fast you were going when you hit the wall.
Torque is how far you moved the wall when you hit it.

 

[TheAssassin]

Dan, I did it your way (and bone's way as well) by starting with torque. I did the same thing, entering all the torque values and rpms going up by 200. I found the average torque to be just over 145 ftlbs. Then for each torque value I calculated the horsepower for their respective rpm. The new horsepower values were very similar to the ones I first used. Average horsepower in this case was 132.455 which is very close to 133!

 

[christbone20]

why are you calculating average hp..?

and if your going to calculate it you should be picking the intervals of what rpm you drop to in gears to redline. take your hp at redline and your hp where your gear starts and divide by 2.

picking 2000 rpm is as arbitrary as picking 1000 rpm and redline or simply dividing your peak hp by 2.

 

[TheAssassin]

read dan's thread about torque versus horsepower, average horsepower or average torque (over a given rpm range) is much more important than peak hp or tq.
I could calculate it by using any rpm range, that's in the first post. For drag racing I would use the range from redline down to where gearshifts take me. The reason why I used 2000 rpm was that was where my dyno plot started as well as most plots I have seen. Most tuners I guess can't go WOT from idle on the dyno, that's just a guess.
And dividing your peak horsepower by 2 does not get you average horsepower. You could have two cars with identical peak whp (or torque) but very different averages. For any given rpm range, the car with the higher average (all else equal) will be faster. A car that produces 50 horsepower until 5000 rpm and peaks at 400whp will have a lower average than a car that produces 150whp until 5000 rpm and peaks at 400 whp.

I don't know if you read dan's thread but it's very informative. What I'm trying to say is that peak horsepower/torque doesn't matter as much as averages over whatever rpm range you drive in. For me, I'm not always drag racing so I used the entire dyno plot, 2000 rpm to 7200 rpm. I could calculate my averages for whatever range. When I drag race I may never be below 4800 rpm so I'll use the range 4800-7200.

Again, not flaming or anything, but taking the sum of the starting whp and the ending whp and dividing by 2 does not give you the average.

E.g.
Range: 5000-8000
Car A: 200whp-300whp
Car B: 200whp-300whp
Adding and dividing by 2 gives an average of 250 but this means nothing. Why? Well lets add more rpm points:
Range: 5000-6000-7000-8000
Car A: 200-225-275-300
Car B: 200-250-280-300

Car B would have the higher average horsepower despite having the same beginning and ending horsepowers.

 

[snm95ls]

why are you calculating average hp..?

and if your going to calculate it you should be picking the intervals of what rpm you drop to in gears to redline. take your hp at redline and your hp where your gear starts and divide by 2.

picking 2000 rpm is as arbitrary as picking 1000 rpm and redline or simply dividing your peak hp by 2.

Assuming the power or torque curve is perfectly linear, that is fine. The reality is that they are not as a general rule.

Using a calculus based approach is far more accurate.

 

[TheAssassin]

Assuming the power or torque curve is perfectly linear, that is fine. The reality is that they are not as a general rule.

Using a calculus based approach is far more accurate.

Precisely, it would be so much easier if I had some way of generating the equation for any particular lines- maybe Excel can do that? Regardless, the trapezoidal rule will suffice since I have no equation to integrate. And as we see so often, torque and horsepower curves are hardly ever perfectly linear.

 

[snm95ls]

Precisely, it would be so much easier if I had some way of generating the equation for any particular lines- maybe Excel can do that? Regardless, the trapezoidal rule will suffice since I have no equation to integrate. And as we see so often, torque and horsepower curves are hardly ever perfectly linear.

Trapezoidal rule is perfectly acceptable in this case.

You could always increase your resolution, but then you run the risk of inducing more reading error ASSuming you are reading the values off of a dyno plot print out or some other visual representation.

Data collected directly by the dyno would be better obviously.

We aren't sending man to the moon here, but some amount of accuracy is certainly desired.

Props for actually taking the time to even do such calculations.

 

[TheAssassin]

exactly, instead of going up by 200 rpm increments I could go up by 50 or 10 assuming I could get precise and accurate readings off of the printed dyno plot. I agree with you 100% I wanted it to be accurate enough while accepting that there will be human error involved.
And thank you, at least I know one person appreciates my efforts! lol

 

[christbone20]

I agree with you both that using calculus to calculate it is more accurate. The reason I chose a linear way was backing up that calculating it is arbitrary. Im not saying your wrong and I totally agree with more average horsepower is better, however, why are you calculating it? You use a dyno to tune a car. If a change makes power keep doing it until it doesnt make any power. Its simply a device to see if your gaining or losing power. I dont see the point in calculating average horsepower. Every dyno is different so you cant compare to other dynos, as well every sized turbo is going to yeild different results. A bigger turbo wont have as good midrange as a smaller turbo. I could see using this calculation for choosing a turbo (to see if a smaller turbo or bigger turbo would end up having a better average hp.) But unless you have more than one turbo and access to the same dyno, calculating this is going to say what?

Im not bashing at all, Im just wondering what your using this information for?

 

[snm95ls]

It is a good way to compare different setups with different power outputs IMO.

:neutral:

 

[TheAssassin]

right, I don't have any reason to calculate this, but ever since I read dan's thread, I wanted to see how easy it was to obtain average hp/tq. I actually think that average hp/tq is a better comparison than peak hp/tq. And just as people say "211 whp at 6800 rpm" (that's me lol) people could say "133 average whp from 2000-7200 rpm" or if you wanted your range to be only while drag racing then you could say "182 average whp and 160 average wtq from 4800-7200" In my opinion it gives whoever a lot more info about your cars output than just peak. Even better, say both peak and average. I think it'll show people you've put a little thought into your builds output (now I'm not saying you haven't given your builds thought! I'm sure you've given them more than I have mine). I just think it paints a better picture to include more than just peaks. And it really didn't take long to find all this using Excel, I can change the rpm range easily. If you want, you could email a pic of your dyno plot and I could figure it out for you if you're at all interested.

 

[ddd4114]

I think you're overcomplicating the concept a little bit, but it's good to see that you're using it. If you're set on calculating this, I would just take points every 500 or 1000 rpm and average them. Just keep it simple.

As mentioned, different dynos read differently. If you're not paying a lot for tuning, you're probably just using a cheap Dynojet. In that case, the shape of the curves will be relatively consistent, but the magnitude will probably be worth fuck-all. In this case, the value in average horsepower is seen when doing back-to-back part comparisons that dramatically change the shape of the torque curve. In those cases, a part change might give you more peak power but make the car accelerate slower.

I never actually calculate average power. If it's not critical, I just compare the shapes of the torque and power curves and use my best judgement. If it is critical, I'm going to be using simulation software that does everything for me anyway.

 

[TheAssassin]

lol I may just be over-complicating it but it's refreshing for me to see a use for calculus in my life besides for business. Don't get me wrong I understand how powerful calculus is, I just only use it in finance and economic modeling. My dad is a nano-chemist, so I appreciate its use in the engineering/scientific world. It was simple enough doing it every 200 rpm, the only hard part was making sure I was getting a somewhat accurate value.

And I agree on the fact that different dynos read differently, and in response to christbone20, that factor would apply in the comparison of two cars' peaks as well. Therefore, it's best use it as you (dan) said, when comparing back to back runs, but the same applies to comparing peaks.

I enjoy conversations like this, I learn a lot and I hope to share what I've picked up over the years.
What simulation software do you use? Or might that be proprietary?

 

[ddd4114]

I write my own

 

[TheAssassin]

lol care to share?
If the dyno plots can be exported to excel then it would be easy to find averages for any rpm range. If this is possible then next time I get on the dyno I'm going to request that he does that for me lol. In my opinion, I think people should say their averages along with their peaks, if they can get them.

 

[ddd4114]

The stuff I write is crude and application-specific, so I'd rather not share. Plus, I wouldn't call it proprietary, but I'm not the only person using it. Also, like any simulation software, if you don't know how it works, you'll either get garbage results or you won't know how to interpret the results.

Nobody really uses average horsepower as a metric, so it would be hard to for most people to grasp unless you're doing a specific comparison. Even then, the powerband has a big effect on the average horsepower. 99% of the time, citing a peak power number and providing a torque curve is plenty.

 

[TheAssassin]

Nonetheless I am interested, however I accept your secrecy lol.
That is true, but a lot of times only peak hp is provided especially from the manufacturers. I can't remember the last time I saw a manufacturer produced dyno plot.

Regardless, I found your thread(s) to be very informative and the basis for my wanting to calculate averages. If nothing else, my thread may have channeled a few to your much more informative one.