Mazda 6 Forums banner

1 - 14 of 14 Posts

·
Registered
Joined
·
123 Posts
Discussion Starter #1
I know the AP power measurements are based off interpolations from acceleration data, but does anyone on here have any experience with the accuracy of the AP "dyno"? How does it's numbers compare to physical dyno numbers? I tried 4 pulls today just for fun and they all seemed fairly close together.
 

·
Premium Member
Joined
·
2,168 Posts
I tried a few pulls and it was very inconsistent for me. Some looked similar, while others showed some oscillation that made no sense. Maybe this is COBB's algorithm for calculating data points:

Code:
while (1)
{
    sleep(2500); // Sleep 2.5 seconds
    dataPoint = rand(); // Draw a random number
    ...
}
 

·
Registered
Joined
·
123 Posts
Discussion Starter #4
Lol. Nice. :D

I hope that's not the algorithm for all their data logging...
 

·
Registered
Joined
·
2,746 Posts
The AP uses MAF, elapsed time and rpm to estimate hp. It is quite a bit optimistic, but should be consistent if you drive consistently to gather the data. It would be useful as a tool to measure improvement after mods. You shouldn't take a single data point as thruth though. As with any estimating algorithm, a series of 5 runs, toss out the high and low, then avg the 3 remaining will get you the most realistic number(s)
 

·
Registered
Joined
·
3,414 Posts
The AP uses MAF, elapsed time and rpm to estimate hp. It is quite a bit optimistic, but should be consistent if you drive consistently to gather the data. It would be useful as a tool to measure improvement after mods. You shouldn't take a single data point as thruth though. As with any estimating algorithm, a series of 5 runs, toss out the high and low, then avg the 3 remaining will get you the most realistic number(s)
The other thing is you have to be on a completely flat surface each time. Although the AP asks you to dial in the approximate weight of your car (fuel level matters), it makes no compensation for incline/decline, which will impact your time, and skew your results.

EDIT: I tried it one day many months ago, and saw +/- 30hp differences with back to back runs, and wrote the AP dyno off as useless.
 

·
Premium Member
Joined
·
2,168 Posts
Yeah I was trying to measure improvement after my DP install, but couldn't get consistent results to average. By consistent I mean one curve would look like an HP curve and another would just be a series of ups and downs. Same "flat" road, same gear, etc. Unfortunately I no longer have the logs, because I got angry and deleted them lol.
 

·
Registered
Joined
·
186 Posts
The AP uses MAF, elapsed time and rpm to estimate hp. It is quite a bit optimistic, but should be consistent if you drive consistently to gather the data.
FORZDA 1, do you know the exact algorithm that the AP uses? I'm very curious to learn how MAF, RPM, and time can generate HP numbers. I'm thinking if I knew how it worked I could decide for myself how valid/reproducible the values would be.

I'm only familiar with the vehicle speed, elapsed time, rpm and vehicle mass method of calculating hp (and torque). That method is very sensitive to the coarseness of the vehicle speed sensor and the slight variations in the speed sensor time stamps. With this method you can also compensate for drag and tire resistance to the get hp at the wheels. This method is definately still good enough to compare the impact of various performance parts if you do take multiple samples.
 

·
Registered
Joined
·
2,746 Posts
FORZDA 1, do you know the exact algorithm that the AP uses? ....
No, I don't. I have not seen the alg quoted, but Christian said something about it in a post or email a while back when he was doing the E-tunes and setting up the HP calc. It may be over in the Cobb forum.
 
  • Like
Reactions: PIXLGAZR

·
Registered
Joined
·
123 Posts
Discussion Starter #10
Here are three second gear pulls I did over lunch:



The spread doesn't look too bad. 2nd gear defiantly worked better than 3rd. Got some weird looking spikes in 3rd for some reason, although it is really gusty today, so that may have been affecting the calcs more in a higher gear. Here is a composite of those three pulls:



Peak numbers do seem a bit 'optimistic', although it seems fairly repeatable.
 
  • Like
Reactions: MWSpeed6

·
Premium Member
Joined
·
2,168 Posts
... The spread doesn't look too bad. 2nd gear defiantly worked better than 3rd. Got some weird looking spikes in 3rd for some reason ...
Hmm... my pulls were all 3rd gear... maybe I should try 2nd or 4th. Your results + averages look good - thanks for posting them.
 

·
Registered
Joined
·
186 Posts
I didn't find Cobb's thread talking about using MAF and RPM to calculate HP but I did find other information on the subject.

MAF and RPM can be used to calculate HP if you have a map or curve of Brake Specific Fuel Consumption (BSFC) as a function of RPM. If this is the method used by the AP then it has these BSFC numbers programmed in.

If this is the case then we can conclude that:
1) The HP and Torque values represent Brake or "at the flywheel" values.
2) The calculation doesn't need or use vehicle mass
3) The road incline is not relevant
4) The accuracy is primarily dependent on the quality of the BSFC numbers.
5) Since MAF and RPM are quick and fairly precise sensors, the resulting HP and Torque value will be fairly precise as well.

So... I'd say that the MAF & RPM calculation method is better than the Mass & Acceleration method for calculating HP and Torque because there's way less variables in play.

Is the AP Dyno accurate? Not a chance.
Is the AP Dyno a good way to validate performance mods with before and after dyno runs? You bet.

With the mass & acceleration calculation method the HP numbers have a +/- 10 hp precision. This is really not enough precision to measure, for example, the performance gain of an SRI upgrade. The MAF & RPM method used by the AP may offer better precision making it a good tool.
 

·
Registered
Joined
·
123 Posts
Discussion Starter #13
If this is the case then we can conclude that:
1) The HP and Torque values represent Brake or "at the flywheel" values.
2) The calculation doesn't need or use vehicle mass
3) The road incline is not relevant
4) The accuracy is primarily dependent on the quality of the BSFC numbers.
5) Since MAF and RPM are quick and fairly precise sensors, the resulting HP and Torque value will be fairly precise as well.
Well, the AP method does require vehicle mass, you enter it every time you run the "dyno" routine. And I'm pretty sure they're not flywheel values, as the AP specifically states "at the wheels".
 
  • Like
Reactions: PIXLGAZR

·
Registered
Joined
·
186 Posts
^So this means that the AP isn't using the MAF & RPM method afterall or it uses some sort of hybrid calculation that we are not privy to.

Either way then, the AP is no more precise than mass & acceleration method used in other software dynos.

Also, any software dyno that doesn't compensate for air and tire resistance when calculating WHP has a major flaw and it even less useful at evaluating performance mods.

How's that for a reversal of opinion. :) Data rules.
 
  • Like
Reactions: dougefresh_
1 - 14 of 14 Posts
Top