My new way of measuring on-road drag

[aerohead]

Originally Posted by MeteorGray

As noted before, I also have observed variations in the TPS (throttle position sensor) screen of my Scan Gauge when at WOT using a throttle-stop device. These variations seem to depend on on road grade and wind speed and direction, ie, the engine load.

For instance, at a consistent WOT pedal position, I found that my Scangauge reported TPS values from 26% at 70mph to 28% at 75mph. This was during an uninterrupted 10 minute WOT run over a stretch of approximately level and approximately north-south oriented road with a wind that the weather bureau reported was 10mph from the northwest. I was going approximately south at the time, so it was essentially a tailwind. I attribute the range in speed to imperceptible road grade and wind direction-and-speed variations. The only surprising thing to me was the 2% TPS variation. This is obviously something other than a mechanical throttle position being reported by the gauge on its TPS screen.

As part of the same test, I turned the car around and drove in the opposite direction, ie approximately north, which then put the wind as a headwind. On that run, which was only a minute or two so did not see as many road or wind variations, I recorded a speed of 65mph at a TPS of 27%.

This kind of puzzled me, since I didn't see why the throttle position value would change with these variations in load when I know that the throttle is being held in exactly the same position, mechanically speaking.

So, I must conclude that the throttle position sensor screen is not reporting the actual throttle position per se, i.e. mechanically speaking. I think in my drive-by-wire Mazda3, the TPS value is actually a calculated derivative, probably resulting from a combination of such computer-read sensors as the throttle sensor on the throttle pedal, the mass air flow sensor that measures the volume of air going into the engine which is used to determine the load for the proper engine settings, and whatever else the engineers felt useful. Just a WAG.

As noted before, I quit looking at the TPS value on the gauge because it is not useful in trying to measure the precise mechanical throttle position, which does not change with load. I just take it on faith that when the throttle is being held at WOT by the throttle stop, the actual throttle position does not change with load. The speed does change with load, of course, but not the mechanical position of my throttle pedal. What the computer is doing with that mechanical pedal position obviously is up to the engineers and what their design dictates.

Incidentally, I also looked at the LOD (engine load?) screen during my tests, but it gave values that ranged widely that seemed even more useless to me.

The world of computers has passed me by without even a courteous honk.

That's a very thoughtful 'WAG' !

[aerohead]

Originally Posted by j-c-c

My suspicion on the vette on the flats, there is a lot of effort there to gain rear DF with the least amount of drag within the rules, being getting HP to the salt is difficult at high speeds with OEM body shape. My guess, drag reduction was the lesser goal.

1) it's not unusual for many of the cars to experience wheelspin, even at the 8-Mile, for the power-to-weight ratios they have.
2) if the car gets sideways, it's aspect ratio, as a crude wing, can increase by an order of magnitude.
3) in 2014, a 1991 Corvette, which had been two ahead of me at DARKO, got sideways at the finish at World of Speed,at 231-mph, going airborne about ten feet in the air, doing a reverse-half-somersault, landing on its roof. The car was trashed, and driver shook for over a half hour afterwards, brought to tears by the event.
4) all those capping plates on the Vinson car give it a bit of weather-vaning capability, and if the driver is quick enough they can get the parachute(s) out in time to save the car.
5) the 'Bonneville' spoiler also helps the pilot chutes to find clean air for rapid chute deployment.

[aerohead]

Looking at the closeup photos, something came to mind.
It appears that the clay introduces an artificial roughness which by itself would alter Reynolds number, directly effecting the point of separation, as with VGs.
Not only the performance of the wing ( turning vane), but also any of the Insight's body covered or behind it.
Technically, this would have to be sorted out.

[aerohead]

Originally Posted by JulianEdgar

A good example of aerodynamic modification testing.

Modification:

Diffuser wing



Intent:

Working with full undertray, change wake pattern to reduce drag

Mock-up:

GOE222 wing extrusion with plywood endplates, attached to rear diffuser with aluminum angle.

Test:

Throttle stop testing, 30 per cent throttle, top speed in km/h, Breadalbane straight, New South Wales, Australia. Weather: 13 degrees C, light head wind, 26/5/20, 8.30am

Results:

No wing - 100
1 finger upper gap - 100
2 finger upper gap - 100.5
3 finger upper gap - 99.5
4 finger upper gap - 100.5

Conclusion:

Makes no difference to drag

Outcome:

Remove mock-up.

Please forgive me if this has been covered elsewhere.
It takes 30-miles (48-Km ) of continuous driving at 50-mph ( 80-Km/h ) before all lubricants and dynamic rubber structures reach equilibrium, steady-state conditions.
If all testing follows this pre-test condition, then we're a-okay, as long as all other possible variables are accounted for.
If not, we're introducing number of variables which are not accounted for during measurements, which remain unknown quantities, and are certain to influence measured data and conclusions.

[freebeard]

Please forgive me if this has been covered elsewhere.

Spoken like a true gentleperson.

[California98Civic]

There is a small problem with this great and clever method in that the throttle position readings vary irregularly sometimes in some cars. Here is a solution.

Problem: I often got significantly different throttle position readings in successive applications of the throttle as I shifted to my test speed and then between different test runs: 16.1, then 15.6, then 17.0 and so on. Sometimes there was wide variation, such as 16.1 to 18.8. Once 17.1 became 20.1. Were the readings a real indication of movement of the throttle plate or was it meaningless sensor or gauge error? Was there "stretching" of the braided cable, shifting of the stopper, or compression of the carpet and sound deadening under the gas pedal? It was hard to tell on the road, so I spent time tinkering with the build and testing.

Testing results: The different readings are caused mostly by compression of materials under the gas pedal and by inevitable, slight shifting of the stopper's position due to engine and road vibration. However, the TPS reading variability is also partly caused by imprecision in the sensor itself. When the spring throttle on the intake is limited by securely-mounted metal pieces, the TPS's own imprecision can still yield variations in its readings but they become slight.

Conclusions: to me, this means that the throttle stop under the pedal can be a source of error in road test results. It would be better to control (1) carpet and sound deadening compression and (2) movement of the stopper relative to the pedal a little more rigorously. These sources of error may not be fully apparent to the driver because the throttle stop movement might allow the throttle to open a little more while the throttle position sensor's imprecision might mask the extra opening. Maybe this only applies to D16y7 Civics, but it likely applies to almost any gasoline car built in the 90s & 00s with cable throttle, electronic fuel injection, and carpets and sound deadening on the fire wall.

Safety: I'm a little concerned about the safety of the under pedal stopper. In an emergency, it could end up under the brake or clutch pedal. Many years ago that happened to me and it was terrifying.

Solution: I made a simple throttle limiter and mounted it on the engine itself. I'll write that up in a new thread, with credit to Julian and to Vekke, hopefully soon.

Here is the under pedal stopper I built and will now scrap. The black end fit into a slot on the floor for precise placement and the wood and rolled tape "legs" rested on the carpet for stability. It works quite well, but it could be better and safer at speed.



Hoping this is all helpful to someone else...

[M_a_t_t]

For those who have done this and watched the reading for TPS what kind of repeatability did you get? I was testing my hand throttle while reading the resistance of the TPS on the car and got a ≤50 ohm variance between tests. This equates to ~2.3% (3.44k ohm with no input and 1.26k ohms at full throttle). When I use it though I do get pretty consistent results as far as speed to the tenth of a mph.

Thoughts?

[JulianEdgar]

Originally Posted by M_a_t_t

For those who have done this and watched the reading for TPS what kind of repeatability did you get? I was testing my hand throttle while reading the resistance of the TPS on the car and got a ≤50 ohm variance between tests. This equates to ~2.3% (3.44k ohm with no input and 1.26k ohms at full throttle). When I use it though I do get pretty consistent results as far as speed to the tenth of a mph.

Thoughts?

2.3 percent throttle variation? Seems a little high.

I didn't bother using the throttle stop after the first few tests because I could manually hold the throttle to 1 per cent (watching the live TPS percentage readout on my Motec dash) eg 20-21, 20, 20, 19, 20, 20, 20, 21 (etc).

One thing that I have learned though is that you must enter the test straight at dead-on the same speed and then move to your throttle stop position. In my case I enter at exactly 100 km/h (there's a tree I sight against) and then go to throttle-stop position, which might then raise or lower that speed by the end of the straight, depending on the aero changes.

But best test is windows open / windows closed - then at least you know you're measuring real stuff.

[M_a_t_t]

Originally Posted by JulianEdgar

2.3 percent throttle variation? Seems a little high.

I didn't bother using the throttle stop after the first few tests because I could manually hold the throttle to 1 per cent (watching the live TPS percentage readout on my Motec dash) eg 20-21, 20, 20, 19, 20, 20, 20, 21 (etc).

One thing that I have learned though is that you must enter the test straight at dead-on the same speed and then move to your throttle stop position. In my case I enter at exactly 100 km/h (there's a tree I sight against) and then go to throttle-stop position, which might then raise or lower that speed by the end of the straight, depending on the aero changes.

But best test is windows open / windows closed - then at least you know you're measuring real stuff.

I thought the 2.3% seemed high as well, but thats what the numbers say .

[JulianEdgar]

Originally Posted by M_a_t_t

I thought the 2.3% seemed high as well, but thats what the numbers say .

Have you made a mistake with the maths?

3,440 ohms closed
1,260 ohms full throttle

= 2,180 ohm range

50 ohm variation

= 50/2180 * 100 = 2.3 per cent.

(Sorry, I am bad at maths so always work out everything again.)

Interesting! I wonder if most of that 50 ohm variation is noise? You can be sure the ECU filters the TPS input.

At this stage, I'd just go ahead and do some aero testing like windows up/down and see what happens. If you are consistent in the change in speed, and the change in speed makes sense, then do some more subtle aero mod testing.