OdinGuru

As most anyone who has been visiting this section of the forum knows, there is ample dyno and accelerometer based evidence that indicates that the Rensis is making significantly less than even the revised 238 HP Mazda crank spec. No one believes the six-speed manual transmission combined with the straight zero angle carbon fiber drive shaft should have losses of any more than the typical 15% (and probably should be significantly less). Thus we’d expect to see around 200 RWHP, but instead we’re seeing around 170 RWHP at most (down an additional 15% from what we’d expect). So where has our HP gone? Most will point the finger at Mazda and say even the revised spec is a lie. I’m a member of this camp, but it has never quite sat well with me that the engineers capable of putting together the INCREDIBLE feat of engineering (the RX8) I’ve come to LOVE could have been off by SO much on the capabilities of the Rensis. Granted, the spec is up to the marketing department, not the engineers, but they must have based it on something from the engineers (I don’t believe they just made it up out of “thin air”).

There has been one possibility that I’ve been kicking around for a while trying to figure out if it actually makes a difference. I’ve decided to write it up and post it here so that the rotary experts on the forum can either agree, or set me straight. Basically, the question comes down to this: what if there is a significant difference between STATIC HP and DYNAMIC HP? If so, perhaps there is some “truth” in Mazda’s specs, although they may still be misleading. This could be similar to the 1.3L displacement spec which has some “truth” but is also misleading in that there are good arguments for making the spec 2.6L or even 3.9L.

So Mazda, and any original manufacture, would be using engine dynos rather than chassis dynos to do their development and testing. As I understand it, on an engine dyno, an adjustable brake is used to measure the output of the engine. The pressure on the brake is increased (making more resistive force) and balanced against the engine output at some specific rpm. In this way, controlled repeatable measurements of output torque and power can be made at any part of the engine operational range. So as I see it, this would give them a measurement of static HP: one where the engine is operating at constant speed. The vast majority of techniques available to those of us who only have engines pre-installed in our cars, most of the techniques that are available are dynamic (engine is changing speeds during measurement). This includes the vast majority of chassis dynos as well as accelerometer based measurements (e.g. G-tech). Although the two techniques take different approaches to measuring power, obviously historically there haven’t been massive differences in the results, otherwise it would be common knowledge, and or different measurement techniques would be used. So for the difference to be significant, it would have to be due to something that was SPECIFIC to rotary engines and not found on cylinder based engines.

A chassis dyno will read a power less than an engine dyno because it takes into account several additional factors. The most well known factor is the frictional losses in the drive train, but also there are inertial losses due to the rotational inertia of the wheels, drive train, flywheel, and engine. Most of these factors are basically equivalent between normal front engine RWD cylinder based cars and the RX8. If anything we’d expect the RX8 to have a lower than normal loss due to the light weight carbon fiber drive shaft and the fact that there are only two zero angle universal joints. The only thing MAJORLY different is the engine. So the question becomes: is there a significant difference between the rotational inertia of a rotary and a cylinder engine? I believe there might be.

One of the main advantages of the rotary engine design is that there is SIGNIFICANTLY less reciprocating mass. In fact the only remaining reciprocation is a due to the eccentricity of the rotor movement. As a result, over 80% of the rotor momentum is conserved. In contrast, in a cylinder engine, most parts (cylinder heads, valves, rots, etc.) reciprocate and thus carry no momentum into the next cycle. Only a few parts like the crankshaft, camshafts, and pulleys rotate and conserve momentum. So depending on the weight of the rotor, could the rotational inertia of a rotary engine be greater than that of a typical cylinder engine (excluding flywheels)? If this is true, and the difference is significant, then we’d see a corresponding decrease in HP in dynamic vs. static measurements (because rotational inertia is not taken into account in static measurements). Thus it is POSSIBLE that this effect could be responsible for some of the gap between observed and published “specs.”

Am I barking up the wrong tree here? It seems to me that even if the engine itself had significant rotational inertia, then wouldn’t Mazda be able to get away with a significantly lighter flywheel then normal (perhaps this costs more). I know that there is a thread on installing a lighter flywheel, perhaps someone knows how the stock flywheel compares with typical ones used in this type of car. I don’t have an estimate of the extra rotational inertia that may be present, and thus don’t know how much of an impact it may have. I though it might be major considering the rotors are somewhat large, cast iron, and rotating at relatively high speed (in comparison with the wheels). For example at 90mph in 3rd gear the engine is at around 9000 rpm, but the wheels are only going about 1080 rpm. Nine times the rotational velocity results in 81 times the rotational kinetic energy for a given rotational inertia. Putting it another way, the wheels get accelerated better by the torque multiplication from the transmission, but the rotors don't so they are effectivly "heaver" by the overal gearing ratio (which is typicaly quite high in the RX8).

PLEASE chime in if you think this makes any sense/no-sense.
- David

legokcen

One comment of correction is that you stated that :

------------

For example at 90mph in 3rd gear the engine is at around 9000 rpm, but the wheels are only going about 1080 rpm. Nine times the rotational velocity results in 81 times the rotational kinetic energy for a given rotational inertia. Putting it another way, the wheels get accelerated better by the torque multiplication from the transmission, but the rotors don't so they are effectivly "heaver" by the overal gearing ratio (which is typicaly quite high in the RX8).
------------


Actually the rotors are rotating at 1/3 the speed of the E shaft. At 9000 rpm Engine speed, the shaft would be turning at 9000 rpm, the rotors turning at 3000 rpm and the wheels at 1080 rpm.

Just so we don't get confused.

GeorgeH

iTrader: (2)

I'm not a rotary expert, but I'll make a few comments.

Interesting idea, and there may be some merit to it, but I am a bit skeptical that it can account for all or even most of the "missing" power.

One question worth asking is if RX-7s have had this same problem - I suspect not, or we would have heard about it.

Also, it's worth pointing out that the dyno pulls happen in higher gears - 3rd or 4th, I believe. Because of that, and the fact that the rollers are quite heavy, the actual angular acceleration of the crankshaft is fairly low during a dyno pull - much lower than a first or second gear pull on the street, I think. Anybody know how long it takes to make a pull on a Dynojet using an RX-8? From that we could compute the average angular acceleration of the crankshaft.

The above is a long-winded way of saying that the rotational inertia of the drums is much greater than that of the rotors, so even if there is more rotating mass in a rotaty than a piston-slapper it wouldn't register on a dyno pull.

Let's see what the rotary experts say.

Rotarian_SC

Hmm, this is an interesting proposal. Maybe it didn't affect the rx7 because it had turbos somehow, but I can't really think how that would come into play. Also as far as I know rotaries don't contribute to more drivetrain loss than piston engines, but is that true? Would have to look at the rx7 dyno and engine output for that. I know that most people have done the 3rd gear dyno pulls cause well it is the easiest, but there is another method I have heard about, remember reading it in a post somewhere from an expert. It can use a first gear pull and you have to manipulate the sensors in some way. So i guess that leaves it up to the experts. Maybe rotarygod could shed some light on this for us.

Spin9k

Any 7 owners who've dynoed their cars chime in and share some numbers and to the wheel losses as a validity/reality check here?

OdinGuru

You are quite right on this, so any effect will be "not as bad" as I was originaly thinking by this factor.



I haven't been able to convince myself that this is significant or not. Mostly looking for more info to throw in the discussion.


True, the drums will have MUCH greater inertia. However, I was under the impression that this inertia is calibrated, known, and in fact used to calculate the HP output. Putting it another way, the dyno computes power by seeing how fast the car can accelerate a fixed known rotational inertial load (i.e. the drums). Thus wouldn't this factor already be "taken into account".


If this theory does have an effect, then it would definatly be true for the RX7 as well (even more so due to the heavier rotors). The only diffence could be that Mazda's "official" specs may have been more appropriate. Basicly, if this is TRUE I belive that the engineers at Mazda are well aware of these properties, but that the marketing department gets to "choose" what it states is the official spec, regardless of which figure makes the mose "sense".

I've been trying to dream up an experimental procedure that would be able to "measure" the rotational inertia of the engine+flywheel, but I haven't fully figured it out yet. I've been thinking along the lines of recording the time it takes to decellerate a fixed amount from a fixed speed, but I can't figure out how to seperate the effects of friction from inertia, and I'm worried that things like pumping losses may cause too much of a problem for this to work. The other option is to look at differences in HP between different gear runs, but this probably requires figuring out the frictional and intertial masses of the rest of the drivetrain in order to cancel them off properly.


David

epitrochoid

Thats a very interesting idea...

We need to get someone to put a renesis on a engine dyno and see what kind of numbers they get. If they get ~238 then you might just be right. If they get ~220, then it's time to petition Mazda for some more free service

GeorgeH

iTrader: (2)

I wonder if Racing Beat will ever divulge the hp readings they get off their engine dyno? It would be interesting indeed.

Here's another point about rotary interial losses as compared to piston. You point out yourself that rotaries maintain much more of their momentum than a piston engine does.

After all, pistons completely reverse their momentum four times for each power stroke. Based on this, I tend to believe that piston engines have greater intertial losses than rotaries, even if the net rotating mass is higher in a rotary. Rotors don't change direction as violently as piston/pin/connecting rod assemblies do, and therefore have less inertial losses. Seems logical but I am guessing here.

I recall reading about the relationship between lighter pistons and recovered horespower - it's a well understood formula, IIRC. I imagine there is a similar formula for rotor weight, although it will be trickier given that the effect of the added weight will change based on how far it is from the center of rotaton. It would be interesting to compare the formulas. Another question for the rotary gurus.

OdinGuru

Yes, I definatly belive the rotatry has overall "less" inertial losses, but remember we are talking about the difference in measurement between static and dynamic power measurements. In other words, any loses from reciprocating mass ALSO occur even if the engine is opperating at constant speed, so this is effectivly "take in to account" properly by static dynos. So the root of the proposed theory is not that the rotary engine has "more" inertial losses, but rather the losses are somehow "different" in such a way that they could cause variations with well established measurement doctrines. I.E. only the part that is "swept under the floor" in a normal static test is greater.

David

GeorgeH

iTrader: (2)

Ah yes I see.

BaronVonBigmeat

Hmm, I was thinking something along these lines...wouldn't a lightweight flywheel car show up as the same HP as a heavy flywheel car on a dyno, only with a tad more acceleration, since it changes RPM faster? Ditto for a lighter driveshaft, wheel/tire combo, and internal engine mass--anything which lowers the moment of inertia.

Another question, is there any reason the renesis has cast iron rotors, other than cost savings? Wouldn't a good forged alloy be stronger/lighter, maybe 4340?

OdinGuru

Yes, any of those things would show up with more acceleration, but how much is the question? Also a dynamic dyno measures HP/Torque by looking at the acceleration of a known inertial mass, thus "a tad more acceleration" means by definition that the HP numbers would NOT be the same (measured difference would be proportional to the increased acceleration). This is basicly because RWHP is the measure of power transfered to the road (or drums a dyno), so any reduction in inertial mass means less energy is wasted accellerating "useless" parts and more goes to the road (more efficient power transfer).



I'd bank that it is purely a cost issue. Any material they make piston heads out of would seem suitable for the rotors (similar stress / heat environment). Although if this "inertial" effect turns out to be small then perhaps it's not worth the extra cost.

Anyone know how much a rotor weights? I've been trying to find out by searching the web, but no luck so far.

- David

BaronVonBigmeat

Well actually, I was thinking more along the lines of using lighter/stronger forged rotors to increase the maximum RPM of the motor for more power, although you'd have to come up with a redesigned intake manifold too. 270 HP @ 10,000 RPM in a streetable car sounds awfully sexy :D

Rotarian_SC

I am pretty sure you can get over 11k redline if you use ceramic or titanium coated rotors, right now though the problem is you would need ignition timing to be able to make more power at that increased rpm

ChrisW

Isn't the short answer to this that the rotational inertia of the accelerated load in the dyno is very much larger than the rotational inertia in the drivetrain of the vehicle being tested, so we can safely ignore the latter?

And this is why the technique works and produces reasonably reliable results on a variety of engines and vehicles of widely varying sizes and configurations and therefore widely different rotational inertias (e.g. 2 liter 4 cylinder vs 6 litre V8).

FWIW, I suspect the rotational inertia of the rotary engine is not that high. The rotors themselves are made of cast iron and are fairly heavy (8 kilos or 8 pounds, I can't remember which) but their rotation is only at one third eccentric shaft speed and there are only two of them. The 9000 rpm orbital motion only has a radius of 1.5 cm so the inertia due to this will be small. And then there's the light weight one-piece carbon fibre driveshaft...

MazdaManiac

iTrader: (3)

The lower rotational inertia shows up as a flattening of the torque curve over time.
If one were to read a dyno plot completely and correctly, the differences between motors with different inertial capabilities but identical horsepower would be obvious.
Time is usually ignored as an axis in favor of RPM which, to the wheels which have a transmission to talk to them, is irrelevant.

To summarize:
HP to RPM is relevant to tuning.
HP to time is relevant to comparing different engines.

Rotary787

I have suspected the Rotary might have a larger rotating Inertia than the piston engine. Also I would think any rear wheel drive engine would have larger rotating Inertia than a front wheel drive.

The more rotating inertia combined with frictional losses the greater the difference between wheel hp and engine horsepower. Any power made by the engine must accelerate the rotating masses then overcome frictional forces. The remaining power is used to ACCELERATE the car or drum on a wheel dyno. The wheel dyno does not take into account this rotating masses and friction therefore the horsepower ratings are lower. When horsepower is measure on an engine stand it does not accelerate anything. (Power = measured_torgue * rotating_velocity). The same hold true for horsepower and top speed (Nothing accelerated.)

ChrisW

I've just checked in the Yamaguchi book - the Renesis rotors are 8.6 pounds (3.9 kilos) each.

Rotary787

I am interested in comparing the the rotary to a simalar piston engine (~3.5 v-6). I may try to calculate rotational inertia for a rotary engine. I have about 5 car design/car dynamic books so I will check and see if their are a some sample numbers for the drivetrain. I think the piston engine will be tougher to calculate, but I may be able to find sample data.

Rotarian_SC

A similar size piston engine would be a 2.6L V6

Prod

You could make the argument to compare it to a 3.9L V6. The rotary fires 3 times for every one time a piston fires. 1.3x3=3.9 The output shaft rotates once for every time the spark plugs fire. In a piston engine, it twice for every time the spark plug fires. In this way it is generally more suitable to compare it to a 2.6L V6.

Rotarian_SC

I know, we are doing it for power through the output shaft so I think 2.6L would be more suitable, plus the Renesis was placed in the 2.6L category for the International Engine awards, so I would generally go with that.