rxphink

The engines in the Auto and Manual are not identical. The maunual trans engine has a different intake configuration with two extra intake ports available for high RPM running

TybeeRX-8

suggest you find a bunch of fans and put them aimed at the car when dynoing. This will at least try and simulate the volume of air this car is going to take in at driving speeds.

I asked about this on another thread and was told they did have a fan or fans to supply air while on the Dynojet. So...
Guess we'll just have to wait and see. Sports Compact mag had the most logical explanation I've seen yet. Some cars just don't work properly on a Dyno for any number of different reasons particular to the car being tested.

Xavier

Wow! The civic boys will be all over this one

an intake that gives you 40 more hp!

Quick_lude

That was probably me.. While I can't speak specificallly for that place, any dyno place worth its salt has at least one fan under the air intake to simulate road conditions. If the lack of air was the issue with low dyno numbers, how do you explain the Japanese mag 1/4mile numbers which were virtually identical for the low power 5mt and high power 6mt versions?
I would like to see some real 1/4 mile numbers from the drag strip, none of these "I raced my buddy" on the street and got a 14.8 stuff.

rxphink

Not a different intake, that would imply that the differences are before the throttle body.

The MT has a different intake system, this includes intake runners and intake ports in the engine.
Like so:




This is how it works in the high power engine:

ChurchAutoTest

Red 6-spd, 2650 miles, completely stock.

Dyno: Dynapack 2000 chassis dyno - hydrostatic loading, hub dyno (wheels removed, dyno bolted directly to drive hubs).

Power (SAE corrected) output: 189 hp @ 8179 rpm
Torque output: 138.5 lbs-ft @ 5918 rpm

If someone can host pictures, I will email the scans to you, as well as comparisons with a typical S2000 and a typical 350Z, two cars which share similar drivetrain layouts and power levels.

Just for reference, the RX-8 curve looked just like I would expect it too, right up until 6000 rpm where things got wierd. Compared to a typical S2000, it was down 21.5 whp at peak. Compared to a 350Z, it was 63.8 whp at peak. In either case, this would suggest an actual power output of around 220 crank hp. Mixture was rich, but only about a half point richer than the average S2000 - so while rotaries do tend to be a little more sensitive to mixture than piston engines, they aren't _that_ much more sensitive.

Another interesting point - My scan tool could not read anything from the ECU. I've tested several OBDII Mazdas, including an MP5 a few months ago and had no problem pulling OBDII data, so this was interesting. The scantool was receiving power from the port, but it kept telling me there was no signal and to check if ignition was on.

More info coming.

SC

Xavier

Take off those big 18s and it makes a difference. It might explain something... similar cars with smaller stock tires dyno better ex: RSX

P00Man

was the AFR around 12?

cause people are saying that it should be A LOT closer to 14, like 13.7 or even 14.1 or something like that

anyway, thats very strange with the ECU thing... is it possible its using a new version or something along those lines?

Xavier

More fuel to the ECU fire, considering they were programmed at the docks. Im thinking Mazda didnt meet emissions or something along those lines... So they got reprogrammed and had to be locked? so the dealers couldnt revert back to the old ones.

RomanoM

Thank you very much!

Basically, and correct me if I'm wrong, you got about what we've seen so far?

How big are the pictures you have. I have a bit of room left on my site, I maybe able to post them for you.

P00Man

"So they got reprogrammed and had to be locked? so the dealers couldnt revert back to the old ones." - Xavier

then wouldnt it probably say something like "no signal" or "locked" instead of "check ignition"?

my guesse by the "check ignition" thing is that the scantool "simply" couldnt recognize the ECU, and therefore "concluded" that the ignition was off

P00Man

"Im thinking Mazda didnt meet emissions or something along those lines... " - xavier

and if they reprogged to meet emissions, then why the rich mixture, that doesnt really make any sense, the engine is running so rich its practically blowing gass droplets out of the muffler

that said i still think the best explaination is the the break in fuel map theory that we have, either that or that it cant be dynoed well, which also can be thrown away given the results Mr. Church (is it Mr. Church?) posted due to the type of dyno he used which would have negated those previous possible effects by bolting straight to the drive hubs, which i imagine would eliminate slip and thus eliminate the ECU trying to negate that slip in order to save the driver going off his desired course, in other words, i feel that this particular type of dyno would "bypass" the TCS and DSC systems

loco4rx8

OK, I've been reading about this horsepower issue with great interest, since I'm about to plunk down $30,000 on this car.

I know nothing about this stuff, didn't even know what a dyno was before RX-8 Forum. But here's something that seems to make sense to a "layperson."

ChurchAutoTest dyno'd the RX-8 at 189 HP and noted "things got weird" at around 6000 rpm. Now, doesn't Mazda ask that we break the cars in by keeping them under 7000 rpm for the first 1000 miles (or something to that effect)?

I think this might lend creedence to the break-in ECU mapping theory. Maybe Mazda doesn't trust us to break the car in the way it needs to be done.

Keeper

That car had over 2500 miles on it. Well beyond the manufacturer's prescribed break-in period.

loco4rx8

Yeah, I know. But some folks here have speculated that it might take 5,000 miles for the ECU to "remap." I dunno.

Keeper

Yes, higher revs CAN yield that much of an HP increase. In fact, you can get that much of an HP increase without increasing the torque output of the engine at all.

HP is a function of torque --> HP = TQ*rpm/5252 rotations. Most people think of HP as "power" or "force", but it isn't -- it's a measure of work. "Work", in physics terms, is a force applied over time. Torque is the force exerted by an engine. The more often you apply the force, the more work you do. The faster the engine revs, the more often the force is applied, and the higher your work/hp number is. I could get into where the 5252 rotations comes from, but then this post would get even more boring than it already is.

So let's translate this to a real world example. In this case, let's relate it to the dyno that was just posted.

If you were putting out 140ft/lbs of torque at 6000rpm, you'd be making 159hp. If you were putting out 140ft/lbs of torque at 8000rpm you'd be making 213hp.

That's actually a 50hp gain that you get just from making the engine spin faster.

Obviously, the trick is to keep the torque up at higher revs, which is what the problem here is...when the intake ports switch, torque output drops drastically instead of staying flat.

eccles

Not just an intake. Because of the way that a rotary works, altering the placement of the intake ports alters the timing and duration of their opening. So when the second and third ports open as the revs rise, the effect is not dissimilar to Honda's VTEC, as the extra ports incrementally alter the "intake valve" timing and duration. (But without the obnoxious VTEC din.)

ChurchAutoTest

Yes, the results were in line with what others have been reporting.

I have 3 GIF files each about 150k in size. I can email them to you if you like.

One displays the RX8 compared to an S2000 (hp and tq), another shows the same comparison between an RX8 and 350Z. Finally, there is an A/F ratio comparison between an RX8 and S2000. Please note that on the charts it says "flywheel" hp. That is not correct. The Dynapack sw allows you to enter a correction factor for transmission losses (you will see it on the charts as "TCF"). We don't attempt to correct for this, so the factor is left at 1.0 which means the numbers you see are what was measured at the hub.

Someone else mentioned comparing an RSX. I didn't think that was totally valid for a comparison since it is a transverse FWD drivetrain, something more efficient than a longitudinal front/rear setup. However, the average stock RSX-S produces about 172 whp on my dyno, which is 17 less than the RX8 - again supporting the 220 crank hp number.

I think the RX8 compared to the S2000 is the most valid comparison. The RX8 is rated at 6 lbs-ft of torque more and 7 hp more. Both cars have high revving powertrains, similar rear end ratios, similar layouts, etc. Up until 6000 rpm the RX8 is handily ahead (as expected, it displaces more and has the variable intake system). The curve is smooth and steady. And then at 6000 it gets rough and takes a dive while the S2000 curve takes off. If the RX8 was performing as advertised, I fully expected the two curves to be virtually parallel at high rpms, with the RX8 being slightly higher of course.

Hopefully Romano will be able to post the charts up for everyone to see.

SC

RomanoM

450k..no problem I have 650K left of space

I can make some room if need be.

vrooom@verizon.net

send them all at once, I have broadband.

I will put them up on the server and then e-mail you back the URL address so you can post them the way you want in this thread or anywhere else.

ChurchAutoTest

Regarding some of the other questions raised, many will be answered when the charts are posted, but...

The average AFR was right around 12:1. This is not optimal for power, but not unusual in high revving, high performance engines. By comparison, over the same rpm range a Honda S2000 averages about 12.5:1.

The scantool returned a "no signal, check ignition" response after about 1-2 minutes of trying to make a connection. As I said, it was able to power up from the port, but was not able to interface with the ECU. The only other time I've seen anything similar was when I tried to link up with a European market Honda development ECU from a pre-production car.

The TCS was fully disabled (hold button for 5-10 seconds). However, the Dynapack would not preclude the same dyno issues that could happen on a roller dyno as one set of wheels is spinning while the others are not.

In light of the performance under 6000 rpm, which looks just like one would expect based upon rated power, published numbers, etc., I would have to say that something is holding back power above 6000 rpm. Whether its a wheelspeed based issue on the dyno, or an ECU issue, intake issue, etc. I don't know.

One thing I did notice were two vacuum ports on the lower part of the intake manifold, capped with black rubber vacuum caps. Any idea what those are for?

SC

rpm_pwr

on every injected rotary I've seen, Mazda leave a few spare nipples to connect gauges / etc to.

BTW Nice work ChurchAutoTest.

-pete

RomanoM

The Graphs from ChurchAutoTest:

RX-8 v. 350Z Power and Torque

Link: http://mysite.verizon.net/romano.michael/RX8-350Z.gif

RX-8 v. S2000 Power and Torque

Link: http://mysite.verizon.net/romano.mic...S2000pwrtq.gif

RX-8 v. S2000 Air/Fuel Ratio

Link: http://mysite.verizon.net/romano.mic...8-S2000AFR.gif

ZoomZoomH

iTrader: (1)

dude i can barely read those graphs, can you enlarge them?

but I do see from what I can see, that at 6000rpm the power 'leveled off' instead of keep climbing, in the RX-8/S2000 HP chart.

that's definitely NOT what the engine should do

RomanoM

By your command:p

ProtoConVert

What would happen if instead of an ECU mapping issue there were some kind of US-production derived obstruction in the tertiary intake path, or a lack of resonance? does this make any sense?