brillo

I did a search on this but haven't found any good single topics on the volumetric efficiency of the rotary engine. I know that the design of the combustion chamber affects power mileage etc, so I was wondering if anyone had looked at way to improve the design or augment what we already have. How efficient is the rotary compared to a piston in numbers? Are there other rotary designs that are more efficient? Is it possible to reshape the indentions in the rotors to improve things? What about using direction injection engines, as VW and Honda are going to whereby they inject fuel right into the piston chamber? Could that help us? It seems to add both power and mileage for the new VW 3.2L V6. I just thought it would be fun to speculate on what we could do to improve this aspect of our engine design.

globi

Regarding the volumetric efficiency: I would bet that the rotary engine has a better volumetric efficiency than the 4 stroke piston engine since it doesn't have any valves that obstruct the incoming airflow.
If the rotary engine was run as a simple air compressor it should beat any piston engine with valves. Actually the rotary engine was used to supercharge a 2 stroke engine before a functional rotary engine was built (see picture on the bottom). http://www.nsumotor.onlinehome.de/baumm.htm

But there are other things than the volumetric efficiency that make the engine less efficient, which have been discussed before like the surface to volume ratio of the combustion chamber or the fact that combustion chamber is moving.

Anything must have been tried with rotor indentions, so I don't think that there is much potential.

Direct injection could be helpful but is somewhat tricky, since the combustion chamber of the rotary engine is longer and moving. You might need several injection nozzles.

In general any supercharger should make an engine somewhat more efficient, since you can build a smaller engine with the same power (less friction).

Anyway, the rotary engine might never challenge the piston engine regarding its efficiency. So as long as its being used where its power to weight ratio, its volume (low cg) bring a significant advantage, the lower efficiency can be neglected.

Just my $0.02

CERAMICSEAL

The wankel rotary engine has the worst combustion chamber shape of any modern engine (to the best of my knowledge), the area of flame propagation is terrible: long, thin and moving.
Direct injection does hold serious potential.

jsh1120

This is an old thread, but it's close enough to my question to add it here. As I understand it, the long, narrow shape of the rotary's combustion chamber is inefficient compared to a piston engine. I've assumed that Mazda has designed the engine with two sparkplugs to improve the efficiency of flame propagation. (True?) If so, would the addition of a third spark plug add to that efficiency? (I realize this is not a possible modification of an existing engine; I'm just wondering if such an approach has been used or found not to improve power/efficiency significantly.)

Aoshi Shinomori

I'm pretty sure Mazda's race engines have a 3 spark plug setup. I'm assuming that there has to be some benefits if it's used by the race cars.

globi

The 26B had 3 sparkplugs:
http://www.mymazdarotary.com/mazda_r...paper_html.htm
My assumption is that the added costs of a third spark plug didn't justify its benefits on the Renesis.

rotarygod

CERAMICSEAL summed it up pretty good in very few words.

The thing you need to understand about piston engines is that the fuel doesn't travel very far before it is ignited. The fuel enters the engine and right behind the valve is the spark plug. It gets ignited right there. Modern valve design swirls the intake as in enters the cylinder and this keeps the fuel fairly evenly distributed. The spark plug is usually pretty close to the center of the combustion chamber. The chamber being round also means that flame front movement doesn't have to go very far and it gets more of it. All of this adds up to everything that a rotary is not.

The rotary takes in air and then moves it over to the other side of the engine before firing it. This is a long way to go and in the process some will fall out of suspension and stick to its surroundings. Also remember that the rotor itself is turning at only 1/3 the eccentric shaft and we therefore have a 720 degree power stroke. This means that the fuel moves through this distance slower per rpm than fuel does in a piston engine. We arent swirling it either due to the chamber design. Moving fuel slowly over a long distance isn't very efficient. Then we have the leading spark plug located very far off of the center of the chamber when the rotor is facing that side. The flame front favors the direction that the rotor is moving which means that much behind the leading plug doesn't get burned as good or as fast. The solution to this is to add a trailing spark plug to help alleviate this problem. The trailing plugs timing vs the front one is changed though depending on rpm and load for best effect. Now we have 2 different flame fronts. It isn't necessarily better to fire them both at the same time though. Even with 2 plugs firing, there is still fuel that has fallen out of suspension which doesn't burn as completely. There is also a little fuel that still hasn't burned near the trailing apex seal area. This was addressed with the late trailing plug on the R26B engine.

Due to the engine shape, movement and design, there isn't a very easy way to improve this other than maybe using a smaller engine with more rotors. One thing that should hold promise would be direct injection with the fuel injector mounted on the same side of the engine as the spark plugs. This has been tried by Mazda but I have never seen any results as to how well it worked and obviously we don't have a direct injection rotary engine in the RX-8. It seems like a good idea though.

RX8-TX

Disclaimer: I don't know nuttin'.

Where would the fuel noozles be placed relative to the leading & trailing spark plugs? I am having a hard time picturing the injectors filling in the chamber @ a decent rate (especially @ high rpms)

jsh1120

Thanks, RotaryGod. As usual, excellent information. (Now if only your political views were as enlightened as your technical perspective. )

brillo

After I went back and reviewed my papers, the 26B had its fuel injection right near the port opening into the housing, so it was injected directly into the housing near the spark plugs.

I wonder if rather than add a third spark plug, if you injected fuel right in between the leading and trailing plugs you'd get the increase in economy and power you would expect.

The only thing about that setup would be, As the rotor moves around, its suppose to compress the air fuel mix, if you inject fuel at the end of the compression cycle as your about to fire the spark plugs, I'm not sure that would work. Any body have an idea on this one?

If not, I like RG's idea, go with a 1.5L 3 rotor, smaller chambers which would improve the a/f mix, which would mean you could go with only two spark plugs due to the reduced size and direct port injection.

The only real issue I see with a 3 rotor is its simply more complicated, not by much, but when were talking about mass production, it could be an issue.

globi

There's a catch though: If you reduce the diplacement of one rotor you automatically worsen the volume to surface ratio and therefore increase the heatloss. Also an additonal rotor doesn't not only increase complexity but also increase friction and therefore reduce mechanical efficiency.
Maybe you could have several injection nozzles to improve a/f mix or you place that injection nozzle rather at the beginning and inject several times while the rotor is moving.
Instead of adding a rotor they could make the rotors smaller anyway (assuming the a/f mixture gets indeed better) and add a turbo-charger instead to make up for the power loss. Mazda's electric turbo is a really good idea.
It also appears that a well engineered turbocharged engine allows one to generate an extremely flat torque curve as the new Volvo S40 T5 shows:
from: http://www.autoweb.com.au/cms/A_2259/article.html

RX8-TX

I'd suspect that you could place one injector before the trailing sparkplug and one in between the the leading & trailing. You could start the first injector right after the trailing seal of the rotor closes the intake port (you've pumped air so far)
and start compressing the mix, and possibly use the second injector to adjust for engine speed, load, etc....



The red lines are injector locations.

Would something like that float?

brillo

I guess you would have to model it out and see which was a better trade off, but these are good points. If you go with an all aluminum construction, that weight savings might cancel out the weight of a turbo.

I'm more inclined to see what you can get NA, and with a slightly larger displacement I think you could get the TQ up into the 175-185ftlbs range NA.

Keep it coming, this is good stuff.

globi

Instead of more torque Mazda you could also increase the speed range of the engine and reduce the axle gear ratio and add another gear or simply make the 6th gear longer in order to save gas on a highway (reaching maximum speed in 5th gear). This way you'd also have more torque just at a different rpm.
I think you should take advantage of its silent revving capabilities. A rotary engine will never be a torque monster.
Wouldn't it be fun to rev it up to 11'000 rpm?

rotarygod

The engine doesn't compress fuel. It only compresses the air around the fuel. If you add the same amount of fuel regardless of location, it is still the proper air/fuel ratio.

The advantage comes from being able to effectively burn more of the fuel entering the engine. By injecting it in this manner, fuel has less chance of falling out of suspension or puddling up at the trailing side apex seal. You can actually inject less fuel into the engine but get the same amount of burned, usable fuel into the chamber. The key is in how much was actually used and not how much you put into it. Basically there is less waste which means increased fuel economy.

rotarygod

A additional rotor my be added complexity but it isn't an issue. Yes it has increased friction. You'd never know it though. If we were to compare the additional area in terms of percentage between say a 1.5 liter 2 rotor engine and a 1.5 liter 3 rotor engine, it wouldn't be much more with the 3 rotor. Yes there is an increase. The percentage in frictional losses would not be the same in terms of percentage than the increase in frictional area though. Look at ANY piston engine and compare it's frictional losses to a rotary. It would take several rotors to equal the losses in even a small 4 cylinder engine. I don't really see an issue with this.

rotarygod

It's kind of hard to say that facts aren't enlightended. Politics is probably the single fastest way to make me mad. When I argue about it, I never intend to make enemies even though it comes out that way. I understand that there are different viewpoints and opinions. There's only 1 right way to do it though!

globi

So have better a/f mix on one side and more heatloss and more friction on the other side, if you'd choose a 3 rotor design. I'd think that the added heatloss would be more significant then the added friction.
We don't know to what level the a/f mix quality adds to the relatively poor fuel efficiency of the engine, but with fuel direct injection and/or an additional spark plug that problem could possibly be solved and then there wouldn't really be an argument left for the 3 rotor design other than marketing.
Marketing could be a factor though: I actually never understood why Mazda built a 1.8l V6 engine. It wasn't even revving like a Honda VTEC engine, so what was the point? I'd even say the 20B was just built for marketing reasons or simply to make a point, they could have reached the same power output with a turbocharged 13B.

Given the added complexity, weight and at least to my mind increased fuel consumption, I don't see a point in going over 2 rotors, as long as enough power can be produced with 2 rotors and a turbocharger or anykind of forced induction for that matter. The 2 rotors are already running very smoothly, so there wouldn't really be much to gain on this part.

rotarygod

By that comparison it would be better to run 1 single large rotor. The more rotors you have, the smoother it is. If you think a 2 rotor is smooth (it is), find a 3 rotor engine and take a ride.

globi

Maybe, but first of all a one rotor probably wouldn't run very smooth and second 1.3l displacement on a single rotor rotary might really be too big. I believe the largest commercially available gasoline engine has about 1l per piston, so 1.3l might really be trouble regarding the uniform distribution of the a/f mix especially considering the shape of the combustion chamber.

Regarding the 3 rotor: I've actually never seen a 3 rotor live, so I couldn't say anything about its smoothness, I just wanted to say that the 2 rotor already appears to be satisfactory (at least to my feel).

rotarygod

You should have gone to Sevenstock and heard the 4 rotor engines cranked up. They are pure smooth rotary heaven and the exhaust tone is awesome.

RX8-TX

Could someone verify wether these numbers below are somehow accurate? They are my attempt at getting to the theoretical & actual VE of the Renesis.

Edit: BTW, I picked 8561rpm to correlate to some measurements I took using a CANScan from the MAF.

Theoretical Engine Airflow (TAF)


(ED)(rpm)(VE)
-------------- = TAF
(ES)(C)


Variables:

ED = Engine Displacement (in^3)
1308cc = 79.8190569in^3
2616cc = 159.638113in^3
3924cc = 239.457170inc^3

ES = Engine Stroke (ES=2 for 4 stroke engines) ==> Use 2.6L engine discplacement.
rpm = Engine speed
VE = Volumetric Efficiency (100% Theoretical)
C = Conversion factor from in^3 --> ft^3

Solving TAF @ 2.6L:

159.63*8561*1
------------------- = TAF
2*(1728 in^3/ft^3)


==================
| 395.42ft^3 = TAF |
==================

Edit: I modified the calculations above to reflect what globi suggested. I am sticking with the 2.6L figure, since it seems to be the closest way to approximate a rotary engine's set of parameters to a piston engine. And please ignore the attachment. I will edit it and revise the calculations, and be more descriptive about it.

globi

RX8-TX, I didn't check your numbers but looked at your equations. You don't need that ES factor.
This rotary engine theoretically displaces 1308cc per revolution that's all you need to consider. ED is 1308cc that's it.
So it's TAF=ED*rpm*VE/C

(Since you're using that ES factor, you're calculation with the 2.6 l should be correct, assuming all your other numbers are correct.)

RX8-TX

Thanks globi! So for flow purposes (and to level the playground with a piston engine) its proper to use 2.6L of displacement with a stroke factor of 2.

rotarygod

Just to prove that it works this way I'll tell you how my standalone ecu is setup on my RX-7. I use a Megasquirt ecu that I built. You start from scratch with no base maps to get you going. Sounds like a nightmare! You need to enter some parameters so it knows how much fuel to inject. If you get it right, the engine starts. From there you tune it. I entered in a 2616 cc (2.6L) engine . The rotary is 1308 cc (1.3L) rated by Mazda. Since the typical engine fires all cylinders over 720 degrees of crankshaft rotation, I had to tell the ecu how many ignition pulses would happen to the engine over this same amount. It has no idea the rotary actually completes all pulses over 1080 degrees and it is irrelevant. There are 4 events in this 720 degree time period. I entered in 2616cc and 4 cylinder to the ecu. That is what it thinks it is controlling. The car started on the very first try! If I enter in ANYTHING different, it doesn't run. Tuning is a breeze also. This is definitely how it compares.