ChurchAutoTest

Remember, what you feel when you are accelerating is proportional to the _torque_ curve. HP is a better measure of overall accelerative ability because it allows us to account for gearing, but the acceleration curve, in g's, will follow the torque curve precisely for any given gear.

In fact, allow me to make a prediction for anyone choosing to use an accelerometer on a stock car. This prediction assumes that the dyno curves represent how the car performs in the real world. It also assumes a 150 lbs driver, full tank of gas and average conditions near sea level. Variances can change the absolute levels, but not the _relative_ acceleration.

1. You will see peak acceleration between 5500 and 6000 rpm. In 2nd gear it will be approximately 0.45 G.

2. By 8000 rpm the acceleration curve will drop off to approximately 0.37 G and will keep dropping after that point at about the same rate (0.04 G per thousand rpm). In other words, the acceleration will drop off about 10% per thousand rpm above 6000 rpm.

3. In 3rd gear, the peak acceleration rate will be approximately 0.32 G. Above that point the acceleration rate will drop a little faster than in 2nd gear thanks to aero effects.

Now, in 2nd gear, the rate of change of rpms will be about 1000 rpm/second. Trying to resolve a change in acceleration of 0.04 G/sec is pretty darn tough. So let's think about 3rd gear. The rate of change will be closer to 600 rpm/second. But now the change in acceleration will only be 0.015 G/second. Again, very tough for the human butt to resolve

If the results are not similar to what I've predicted, then maybe something is different between dyno runs and the real world. We'll see.

SC

Genom

Well, I doubt my *** is calibrated that fine, but I'll give it a shot tomorrow :D

Elak

CAT,

I assume your calculations are correct, however I'm not sure you can ignore the air friction at high rpm in even 2'nd gear. I would suggest doing all tests in first gear to minimize that effect.

A change from .45g to .37g over a few seconds would be quite noticable. It is equivalent to your 150 pound driver losing 12 pounds of weight over the same time in the earths gravity field (Atkin's eat my shorts), however absolute values will be hard to "feel".

Now what would we expect if Mazda is right, and the dynos "wrong"? Mazda's published torque curve is flat as a pancake between 6000 and 8000 rpm. If we assume we can ignore the friction buildup, the acceleration should be pretty much constant over that rpm range.

I'll try to find an accelerometer at work...

/Elak

ChurchAutoTest

Air drag at 60 mph is equivalent to about 0.1 psi of positive pressure acting on the surface of the car. That pressure is acting on a CdA on the RX8 of about 720 inches. That equates to about 70 lbs of drag. Compared to the 1400 lbs of thrust or so the RX8 produces in 2nd gear, you're talking about a couple of percent. And the pressure at 40 mph will be about half that, so the relative difference would be about 30 lbs of drag or 2% between those two speeds. If you do the test in first gear, you shorten the sample time so much that you risk additional introduction of errors/variables that could have significant effects (whereas a longer sample time provides more time for errors to accumulate, but have a smaller impact).

Furthermore, I would assert that most people will not be able to discern the a change in acceleration of 0.04G/second, especially when you are already accelerating, which is an unnatural condition for the average person. Furthermore, humans are less sensitive to acceleration in the horizontal plane than they are to vertical (which is perhaps one reason why cars which squat heavily on acceleration "feel" faster). There is a great deal of research on human perception of acceleration, which is often visual, aural and also physically perceived by the inner ear and other organs (something called an otolith I think?). One of the perception tolerances that are often used for various engineering projects is 0.025 G for RMS acceleration. Furthemore, studies on horizontal elevators have shown that jerk must be below 0.2 g/sec to remain below the average detection level of humans.

Not to say that some people (perhaps yourself) aren't above average in the perceptive abilities - by the definition of average, some people must be!

Anyways, enough of the funky theory, let's just get some accelerometers in cars and check out the actual levels!!

SC

MP3Guy

Guys, I've been following this thread for awhile, and all I can say is, get back to ENJOYING your cars without agonizing over this! All of you new owners are indeed very fortunate to own an RX-8, and I'm sure whatever is going on will be either found to be of little consequence, or easily rectified.

It's Sunday morning. Crank it up, pop in a CD, turn it up and go for a ride.

Lensman

Sorry if this has been covered elsewhere but did SCC ever publish the dyno graphs for the RX-8 that they discovered had the stuck third ports? If so how does it compare to the other's that we've seen recently?

Hercules

Good question... I dunno. We will have to see I suppose.

nk_Rx8

I don't believe they did. I have that issue and they just said there was no point in showing the dynos since the port wasn't opening. But I bet their plots would look just like the ones we have seen here.

Elak

I think you are forgetting that the drivers whole back will act like a preassure sensor and even a 0.08g difference would in my mind be noticable. If you try an lie down on the floor of an elevator, you'll sense the acceleration much clearer than if you stand on your feet. Your co-riders may think you're crazy though.

/Elak

ChurchAutoTest

Actually Elak, all those hi falutin PhD and MD types and their controlled experiments say its the opposite - you sense acceleration better in a vertical plane (i.e. standing up).

Your skin/back isn't the best means for detecting acceleration either. Even if we limit it just to your back (say, 500 sq inches?), that 0.04 G/sec jerk rate (an additional 6 lbs of perceived weight) would equate to an effective pressure change of 0.01 psi. And that's not even taking into account the fact that the acceleration in the horizontal plane combines with the 1G acceleration in the vertical to create a vector with a magnitude smaller than 1 + 0.04G.

But again, just get an accelerometer in the car. The human butt lies.

SC

5Gen_Prelude

Are you saying people are pulling these numbers out of their ***?

Lensman

:D

ChurchAutoTest

Pulling something out of their asses yes, but if someone has an *** that's outputting some actual numbers, I think we could get them a lucrative contract somewhere!!

SC

XUrotaryrocket

I second this notion. I too am interested in this thread, but I'm not going to worry about it until someone from "corporate" actually gives us an answer. All I know is that the car is more powerful than my 1991 NA RX-7, but not as powerful as my 94 turbo - which I expected from day one.

Elak

I believe vertical plane in this case means falling (i.e. losing altitude) not whether you stand or lie down

If we simply look at the torque curves from the dynos before 6000 and after, the positive gradient before the peak is actually less steep that the decline after. Does my back register the acceleration before 6000? You bet! Thus I would expect to sense the decline (change - not actual value) after 6000rpm.

Back to a previous discussion about the impact of air drag. Where did you get the 720 inch CdA (you mean inch^2)? The Cd is quoted as 0.31, but I can't find a cross area value. If a Porsche 944 has a cross area of 2790inch^2 then the RX-8 can't be 2322inch^2?

If I put in the Porsche value (CdA=910inch^2), I get about 8hp loss to air drag at 45mph, and 20hp at 60mph. This seems to me to be too high to ignore if we are doing accelerometer tests in 2'nd gear. We are after all looking for a torque deficiency manifesting itself as 20-25hp in that specific rpm range.

/Elak

rpm_pwr

Funny you say that! Just yesterday I was looking at my accelerometer and wondering if that's what's needed to put to rest any suggestions that they need to be on the road to make the full power. Pity I don't have an '8

-pete

ezradg

Greetings everyone,

I am a newcomer to the RX community and have been watching this thread with great interest over the last couple of days. I am also a novice about rotaries, dynos, and the like, so I was hoping that you folks could help me out.

I saw the RX-8 for the first time last Tuesday, parked outside the Ritz in downtown Boston. I literally stopped and asked "what is that?" Since then I have been doing a lot of research, and have taken two test drives. (I have also driven the G35 and 330CI.) I have to say that I am a bit hooked on this car. One of my local dealers has the combo I want coming in within the next month (6MT, GT package, nordic green w/chaparral) and I am trying to decide whether to plunk down a deposit.

Even for a novice, it seems reasonably clear from the evidence presented here that there are "power issues" above 6,000 RPM. So, the questions I have are:

-- Do people believe that this is an ECU/tuning glitch, or is there something more fundemental that needs to be resolved?

-- If Mazada were to say "oops," what we really meant was 210 hp like the automatic, do those who have actually been driving the car think it is still worth $32k?

-- Is the hp problem a short-term show stopper? Should I wait until this is resolved before diving in?

I am not in any particular rush, but of course would rather have the car sooner than later.

Thanks

ChurchAutoTest

Nope, the research says you feel acceleration more acutely in the vertical plane of your body.

As for acceleration, can you really tell that the acceleration increases from 4500 to 5500 rpm? I'll bet you couldn't. But you're watching the speedo, perceiving your surroundings, hearing the engine rev, etc. and those all give you cues. If we put you in an isolation chamber and began accelerating it at the same rate, you wouldn't be able to accurately perceive the change in acceleration. As I said, research shows that a jerk of less than 0.2g/sec is difficult to perceive.

On the frontal area, I used a Cd of 0.28, I didn't realize the RX8 Cd was so high. Take the height of the car minus the ground clearance, and take the width of the car minus the sideview mirror width x2 (remember that car widths are quoted at the widest points and the mirrors account for a very small frontal area). With the higher Cd the CdA will be closer to 800 sq inches. I calculate hp losses at about 8 hp at 45 mph and 18 hp at 60 mph. This 10 hp differential is relatively small in terms of the deficit shown by the dyno curves. On the RX8 I tested, it produced 189 whp when it should have produced 220 whp, so you're looking at a 30 hp deficit with only a 10 hp differential between 45 mph and 60 mph in terms of air drag.

SC

Sputnik

Any response you receive here will be pure speculation. It will be worth as much as a poll that asks "Which NFL team will go all the way this year?".

Quote:

-- Is the hp problem a short-term show stopper? Should I wait until this is resolved before diving in?

That is entirely up to you.

---jps

Elak

Ok, you obviously have looked into it in more depth than I, so I'll back off that one.

Rotarynews interview with the Senior VP of Marketing and Product Development, Robert Davis quotes 204-207hp to the rear wheels on the high power car. I.e. what we are trying to figure out if 15-18hp are missing or not. That means 10hp matters a lot.

/Elak

ChurchAutoTest

You're mixing your dyno's. The 204-207 quoted hp was on a Dynojet. To date, the cars tested on a Dynojet have averaged less than 180 whp. Or a deficit of at least 24-27 whp.

The Dynapack I use reads, on average, 10-12 hp higher than a Dynojet since the wheels are removed for testing. The average 2000-20001 S2000 (very similar drivetrain, power characteristics) puts down 212 whp on my dyno, the average 02-03 217 whp.

If we assume the RX8 should put down 7 hp more than an S2000, it should be producing at least 219 whp on my dyno, which is precisely 30 hp higher than what it did produce. Quite easy to resolve even with drag losses increasing 10 hp over the targeted speed range.

SC

Elak

1stRX8 got 184hp on his dyno. I believe that was a regular dyno. However the car had over 1000 miles, what was the milage of the one you tested?

So lets meet in the middle; we are looking for 20-23hp, overlayed by an 10hp drag effect.

Edit: Going back over the numbers I would compare the RX-8 with a 911 Carrera 2003 (RX-8 is slightly larger) - which would give a corrected CdA of ~940inch^2. That in turn means 20hp at 60mph, i.e. a 12hp loss over 45mph.

/Elak

rx8racer

I have had an accelerometer (a G-Tech pro) mounted in my car for a couple weeks now. I weighed the car on a certified scale so I can get some pretty good numbers. I have done numerous dyno pulls on many different roads and under many different conditions. I've tried DSC on and off. Every dyno curve I download to my computer is the same thing: the horsepower dips bad at 6300 rpm and the torque takes a dive. I have been watching the power as the car breaks in. From what I have seen from 900 miles to 1600 miles on the odometer, there may have been a 2 hp increase. I'm not holding my breath for another 30hp from break in.

Shamus

Blah Blah Blah Ginger... Anybody have any actual information from Mazda on this as opposed to inconclusive dyno run evidence that we've had since 3 weeks ago?

Look, the 204-207 number was something uttered by a marketing guy at Mazda at a Q & A session, and NO he didn't say that was a dynojet number. He simply uttered the words with no specificity. If you want to ASSUME it was a dynojet number, or that he actually knew what he was talking about keep whipping yourselves into a fine froth if you've got nothing better to do.

It's become quite the speculation parade in here.

Elak

rx8racer,

Care to share your data with us?

/Elak