You guys are thinking of this drivetrain loss in completely the wrong way. Drivetrain loss is not a fixed percentage. Thinking of it in percentage is not correct. A drivetrain will have a certain amount of loss associated with it. At stock power levels you can say if it is rated at 100hp and made 75hp on the dyno it must have a 25% drivetrain loss. If you then double the HP of the engine and dyno it again- with the same exact drivetrain- it is not going to dyno at 150hp, probably in the 160-170 range. The reason is the drivetrain losses were ~25hp, not 25%. As the loading of the gears went up, because the higher power, you will create more frictional losses and the HP loss will be slightly more. It's not linear though.

This still doesn't explain why the RX-8 dynos at 180hp.

As far as rotary specific dyno procedures there are none. I've owned, modified and dynoed RX-7's for the last 7 years. They put down power just like any other car. MY last RX-7 which was an 84 put down 245hp with a 1990 turbo motor in it. Maybe it was really 300hp?? I'm going to go brag to all my friends now that my car really did 300hp to the wheels but the dyno doesnt read it right so I can't prove it. :D

lol

that actually is what i would think it is, but i think that everyone is trying to say that "most" engines lose "this percentage" at the wheels when dyno'd
________
Dodge Caliber Specifications

EDIT: Actually that IS completely accurate.

As you suggest losses are not linear, but they are not fixed either.

Drivetrain losses change with load (torque) and RPM.

As the load increases the amount of friction force in the system increases. This part is basically linear, but not completely. Like all friction the greater the load (force) between moving parts the greater the friction force. The more power is wasted as heat.

Also, with increased RPM the amount of power lost in overcoming rotational inertia increases, actually by the square. So you are using more engine energy (power) to spin a part at 9000 RPM than you are at 4500 RPM. This is probably why the RX-8 has a carbon fiber driveshaft to mitigate this effect by reducing mass. Energy used spinning gears and shafts doesn't go into spinning the wheels.


The 15%-20% loss comes from emperical data. Most cars of the RX-8 layout lose between that much. It isn't by any means though a "law."

Synopsis of Dyno/Horsepower Issue

BTW- Just because a car has more power or higher redline than another does not mean it will have higher losses.

There are many factors like coefficient of friction between parts, mass of parts, effective lubrication.

isnt that what he said, to an extent?

"As the loading of the gears went up, because the higher power, you will create more frictional losses and the HP loss will be slightly more" - baz
________
Toys Cam

I need either more sleep or more coffee! Oppps!:( :p

Another thought. The OEM tires are, unfortunately, rated A for traction. Is it possible that at higher speeds, the tires simply lose traction with the rollers (or whatever they sit on) on the dyno?:cool:

i think we should all stop thinking cause we aer just putting in more possible variables to look for

just kidding lol, though it is true... the more thoughts and minds thinking about it, the sooner it will be solved

happens to the best of us Rom
________
Vaporite""

I only use premium fuel; but those that use regular do not seem to mind the loss in HP.

Eccles,

I use that type of dyno (a Dynapack from New Zealand actually). We actually remove the wheels and bolt the dyno directly to the hub which prevents issues with wheelspin, tire deformation/inflation, etc. This results in higher readings than a Dynojet (no wheel mass to spin up), but is very repeatable.

For baseline runs we still use a ramp method, but we set the ramp time. Perhaps on the RX8 we're testing Saturday we'll do a step test too, although we don't baseline many cars that way, so there will be little information for comparison. We used the fixed speed mode to tune of course, especially at part throttle.

SC

Cool! I saw those portable (well, transportable) units in a recent issue of Racecar Engineering. I'll be interested to see the results.

I love my RX-8 just the way it is but one of the things that got me so excited about the car back in January was the 247-250 HP rating. Would I have been the first person to pre-order in January (at my local dealer) if was rated by Mazda at 227-230 HP I don't know but it (the higher HP) was a definite factor.
I love my car but at the same time I feel a little miss-led.
Vince

I just remembered something that could be relevant. I remember a dyno for the Integra GSR where someone dynoed it with stock 15" wheels and then again with 18" wheels/tires, and the dyno showed a 10whp loss. The RX8 comes with 18" wheels, right? Unless those are very light wheels/tires, it will be detrimental for a low torque car. The S2000 has similar power ratings and dynos ~200-205whp, but comes with 16" wheels/tires. I always kinda thought that 18" was too big for the power specs and that they are there partly because bigger wheels are in style right now.

The only thing is that is still wouldn't explain why the Japanese dyno showed correct power numbers.

Did I miss something about a Japanese dyno?

Vince

Finally somebody other than me beating the parasitic loss theory! Be prepaired to duck often lad!

Heavier wheels do make a difference and I have no idea what the weight of those big tires are, even if the wheels are light-ish.

I've always found it funny that shows and mags spend thousands on all manner of hp increasing gee-gaws and then put the cars on the dyno, but they completely ignore the heavy chrome weights shackled to the hubs.

Yet another reason why a variable-load dyno is superior, because it measures how much torque is available at the wheel at a given speed rather than how fast the wheel can spin up a roller.

Wheels: 20 lbs. (9.1 kg) - very light

Tires: 26 lbs. (11.8 kg) - average

Overall diameter: 26 inches (660 mm) -
larger than average, but same diameter as BMW M3, Nissan 350Z, Infiniti G35. The M3 and G35 wheels are heavier, the 350Z Track wheels are lighter.

Note: it's not wheel size that matters, it's overall diameter, overall mass and where the mass is located on the wheel/tire.

Yes, a larger and heavier wheel has greater parasitic losses.
The average diameter for a sporty car is 25 inches and average weight for the wheel/tire combo is ~50 lbs.

BTW - I know this because the effect it has on acceleration is also the effect it has on braking.

I have a stupid question. With the rotary engine needing oil injected to build compression between the apex seals and the outer housing, what does running rich do to the compression at high rpm? Would adding MMO to the fuel actually help compression at high rpm, if the engine is running way rich?

The oil is injected to lubricate the apex seals, not to help them seal.

Your confusion may arise from a trick used by old rotorheads to start a stubborn motor: injecting a small quantity of oil into the intake can help the seals seal at low cranking speeds.

Under normal running conditions, however, they don't need any help.

Sometime in the next week or so on a nice cool day I'll dyno my car again now that it has 3000 miles on it.

Vince

Hi all, this is my first post here. I'm a Mazda6 owner very active at the Mazda6club, but I have a pretty good interest in the RX-8. Had I known it would be so inexpensive, I'd of definately waited to test drive or buy one. Back in November when I was shopping, the price was projected to start over $30k.

Anyway...

It seems to me people here are set on horsepower figures, and why they fall short.

They don't, at least not relative to its power curve. The horsepower doesn't drop off at the high end any faster than should be expected- it's just 7-8% at most. That's a really hard difference to feel, and I'd bet those thinking they "feel" the drop in torque after 6300rpm are just experiencing a placebo effect.

It looks to me that the ENTIRE dyno plot is shy of about 20 ft-lbs of torque, from idle on up. It rarely goes above 120ft-lbs, and is closer to 110ft-lbs where it's torque peak allegedly resides.

To make 207whp at the car's alleged horsepower peak, it needs to make 128ft-lbs of torque at 8500rpm. 128ft-lbs is reasonable to make at that speed, and peak torque should probably come in close to 135-140ft-lbs. For comparison, the Mazda6i is rated at 155ft-lbs, and makes 130ft-lbs to the wheels.

As I said, the entire graph looks to be shy by about 20ft-lbs. Don't forget, torque is what moves the car- the high RPMs (and hence high horsepower) just allows for the really fast gearing of the car, which multiplies the torque to the wheels. It appears that the torque curve is the correct shape- it's just low.

At least that's my analysis.

So- what do I think?

I think that the car might be electronically limited to 90% throttle for the break-in process. Do we have any air-fuel ratios to any of the dyno charts? Perhaps someone could remove their air filter and have someone floor the gas pedal. You could see if the throttle body is opening all the way or not. Correct me if I'm wrong, but if the engine is only getting 90% throttle, it'll only get about 90% of its fuel. To prevent the engine from running lean (or perhaps rich, since perhaps the air would come in with less velocity?), the throttle body would also only open 90% of the way. Does this sound like a reasonable experiment?

I don't know of any mechanical break-in that could account for 15-20ft-lbs of torque, especially since the dyno at the beginning of this thread happened at 2000+ miles.

Various people have already overlaid a power graph from Mazda from last year against actual dyno results (scaled to allow for transmission losses) and have more or less shown that the graphs are very similar from idle upwards except over 6250rpm (where tertiary port kicks in) where the actual dyno shows a lack of torque relative to Mazda's graph.

The car is fine from idle up to 6250rpm. From there to redline it may or may not be fine. Current data suggests there may be a problem with ECU or tertiary port behaviour in this area, but only Mazda knows for sure.

I used my calibrated internal accelerometer to measure the difference, at high rpms the lack of power is very noticeable.

It's a difference of 8%! How is that "very" noticable? Forgive me for I don't own the car, but I'd think 8% would be just shy of negligable. More importantly, the ENTIRE torque curve seems to be down by almost 20ft-lbs, more like 15%.

Can you point me to this graph? I've seen Mazda's factory dyno, but not one that was at all detailed. Certainly I wouldn't be able to overlay the graphs into anything meaningful.

If the RX-8 is ONLY lacking power after 6250 RPM, that means it must have a 20ft-lbs jump after 6250RPM for that to be true. Highly unlikely. Think about that- that's larger than most any VTEC jump, and certainly wouldn't be "smooth" like the rotary is known for.

The increase in torque could be linear after 6250, not a jump as you imply. Also any jump in torque in a vtec system is due to improperly timed cam switch or bad tuning. The whole purpose of vtec is to have a flat torque curve with no jumps or spiking

...i guess you've never seen a Honda dyno sheet (most notably, one for the F20C) :p

...did you do the math?? and also, have you looked at the factory dynographs?? peak power comes right at redline, so the torque difference isn't 20ftlbs large, nor is there suppost to be a jump: the curve just isn't suppost to level off and turn down until 8500 rpm.

Probably to stay in the nice emissions friendly low lift cam for epa testing.

So you're saying the torque curve will peak around 120ft-lbs around 4-5000rpm just as it does now, start to dip around 6000 just as it does now, and then at 6250rpm start a linear rise from about 115ft-lbs to 135ft-lbs or so at 8000rpm before going back down to 130ft-lbs at 8500rpm?

Even if it is linear, the engine spools up so quickly that a 20ft-lbs increase over 1500rpms would feel like a spike. 1500rpms for a rotary is like 800rpms in a normal piston motor- you'd get a 20ft-lbs increase in force in about half a second's time in first gear, even though on the dyno the increase would look linear.

To make 207whp at 8500rpm, the engine would have to make nearly 20 more ft-lbs of torque than the dynos currently show at 8500rpm- that's about 128ft-lbs. That's where I got the number from. Since 128ft-lbs is more torque than the engine currently produces ANYWHERE and 8500rpm isn't even the torque peak, you can see why I think the entire graph is low and not just the top end.

I'm not suggesting there should be a peak- quite the contrary. That's why I think the entire graph is reading low, which lends itself to the theory of the ECU only allowing ~90% throttle or something.

oy, i'm not arguing this BS with you... and peak power comes at 9000 rpm from the original Mazda dyno sheet.

I thought you were some veteran here- crap, I've been reading your posts forever on this site. Peak power is claimed at 8500rpm- 9000rpm is redline.

In fact, this graph shows a pretty close torque curve match too discussed here.

Okay, my car now has 44xx miles on it and this week I will dyno it again and post the results.

Vince

Excellent

Do one run in 6th to settle the argument about the dyno not having enough load time for the flow to stabalize.

...funny how you say that, then cite the thread i went ahead and plotted all those curves together in :p

I edited my first post with new graph and data.

Completed new dyno runs today (9-3-2003) 10 degrees cooler and 10% more humidity.

Substantial increase in both horsepower & torque. I also did runs in fourth and fifth gear and they were within 1% of run 002.

Vince

Your updated numbers look better. If we assume a 22% loss, which seems high, I suppose this looks more like what Mazda has stated.

So, the parasitic loss is still about 22% from the stated 238HP figure. Can anyone comment and give a synopsis, as to good, bad or indifferent.

Hey Vince, sorry for sounding stupid but how do I see your updated numbers

huh, interesting that the entire curve showed the ~8-10 increase, not just at higher RPM's...

~10 deg cooler, 10% more humid, and an extra 2k miles.... :eek:

I updated my first post.

Vince

Next time try 87 octane and maybe youll top this guy's run, which i think is the highest dyno weve seen yet. Youve got 2000 more miles so you should top him with 87.

https://www.rx8club.com/showthread.p...5&pagenumber=1

Well the good news is that you are making all the torque you should. I would think that your car should never ( unless modded obviously ) pull HARDER than that. But on the down side, the torque drops off a couple hundred RPMs early, so it's just not pulling strong as long as it should. It is looking better though and i do think there is an easy fix.

Very encouraging that an additional 2000 miles freed up that much power (did you change the oil in that timeframe?).

If it frees up another 10-12 hp you'll be very close to Mazda's 238 hp rating.

SC

Changed it right after the first dyno test.

Vince

fuel?
 

if I missed it I apologize, but the type of fuel in both tests? assuming premium unless stated otherwise.

California 91!

Vince

Hate to be a pain in the ass, but with the original data gone, can we get a quick summary comparing the two?

Ie, peak hp, torque, temperature, miles on the car?

The original data isn't gone. Go look at the updated first post, which compares the old and new runs, including details re temp, humidity, and miles.

Doh. My bad :P The joys of skim-reading the forums with your boss in the room when you're supposed to be doing work ;)

Interesting. Might be worth having other people recheck after an oil change. I do a lot of oil testing for several companies and we've found power differentials of up to 3-4% by changing from standard factory fills to high quality forumlated oils.

Maybe Mazda is using an unusual oil formulation for break in. On piston engines it usually has a high Moly content, but for rotaries I don't know.

SC