Exhaust Port relocation ideas.

[NoDOHC]

Ok, so I have hinted around for a really long time, now I finally got a moment to post this for discussion.

Simple explanation:
Exhaust port opens too early, hurting thermodynamic efficiency, increasing EGT, and exhaust sound from blowdown.
Intended Fix - Machine exhaust port in different location on RX8 housing (weld coolant passages and fill coolant seal grooves).

Complicated Explanation:
Although a full stroke on a rotary engine requires 270 degrees of eccentric shaft rotation to complete, the time that a port is open is the same as a piston engine given the same speed and port timing.

Unlike a piston engine, A rotary engine exhaust port completely opens in about 25 degrees of eccentric shaft rotation and remains completely open until about 45 eccectric shaft degrees before completely closing (due to a different position of the rotor).

A normal cam for a piston engine opens the exhaust valve at about 45 degrees BBDC. The opening really means leaving the seat. The valve is barely open by BDC. On a rotary engine, the exhaust port is completely open by 50 degrees BBDC. This means a lot of blowdown and wasted energy (any loss of gas pressure in the chamber BBDC is a loss of useful work that the gas could have done). A wild (barely idles) piston engine camshaft will open the exhaust valve at 75 degrees BBDC, but once again, the valve doesn't completely open until 40-50 degrees ABDC. These engines are not typically noted for their stellar fuel economy either.

Anyway, the point that I am trying to make is that the rotary engine efficiency problem may very well be linked to the early exhaust port opening.

Description of proposed solution:

This is the position of the rotor in the chamber when the factory exhaust port opens. As can easily be seen, there is a significant amount of chamber volume remaining when this port opens (about 18%).


This is the position of the rotor at maximum chamber volume. (What I have been calling BDC).


This is the position of the rotor at the proposed exhaust port opening.


This is a closeup of the rotor when the eccentric shaft is 230 degrees ATDC.


A mockup of how the exhaust port sleeve (which would have to be modified) could go in.



This is a comparison between old exhaust port and new proposed exhaust port.



These pictures illustrate that a set of RX7 headers could be easily bolted to the RX8 housings.

The exhaust port would also close later, but it is almost completely shrouded by the rotor at that point, so the little bit of port overlap caused by this shouldn't be an issue.

What do you guys think?

[RETed]

You basically did the same thing a bridge port does but by moving the exhaust port?
How do you address the increase in overlap?
Or are you moving the intake port(s) too?


-Ted

[sen2two]

I started working on this last year with a set of rx8 housings i have. A different purpose than yours though.

Its pretty easy to talk about welding the coolant passages. But a different story to actually do the welding. (I am a welder) You would have to do TIG to get a clean weld that would be strong enough. And its almost impossible to get the TIG torch down in there close enough for good penetration. You can get a MIG gun in there and try to spray weld it, but you will not get good gas coverage. But this is cast aluminum, and even with billet aluminum with good gas coverage, spray/MIG welding will not be strong enough for this situation. And the heat needed to weld such a piece, will warp the housings.

So I'm looking for other ways now.

Another way that i wanted to try is similar to a PP on the intake. Press fit (extreme press fit) a sleeve for the exhaust. I have heard of a few people machining exhaust sleeves from aluminum and them working fine. I don't see how, as i would think that the aluminum would melt.

Then after a lot of searching, i found this...

http://jhbperformance.com/downloads/...racingmods.pdf

[NoDOHC]

Originally Posted by RETed

You basically did the same thing a bridge port does but by moving the exhaust port?
How do you address the increase in overlap?
Or are you moving the intake port(s) too?


-Ted

Actually, I am not trying to achieve the same effect as a bridge port. I am not desiring overlap, but rather better cycle efficiency. I will get overlap, but not a lot, as the rotor will be mostly shielding the cusp on the housing during the overlap time.

Believe it or not, the overlap generated by this modification would be very minimal.

I will attempt to illustrate:


This is a stock 4-port rear iron with the RX8 housing on it and a rotor installed upside-down (just to make it easier on the housing, as there are no planetaries in the housing right now). You can see that the exhaust port is already nearly shielded (the dish is almost past the cusp). The engine has already passed BDC, but the volume is barely changing at this point. This is the big reversion enemy, but only occurs for about 10 degrees of eccentric shaft rotation.


This image shows the intake port open. The dish has completely passed the cusp on the housing, allowing the rotor to housing clearance to effectively block the exhaust port from communicating with the intake port.


At this point in the process, the exhaust port is closed.

I agree that I may have reversion issues, but I am hopeful that they will be minimal.

I have no intention of modifying the intake port timing at all. (I will use my other housings with the very mild street port).

[NoDOHC]

Originally Posted by sen2two

I started working on this last year with a set of rx8 housings i have. A different purpose than yours though.

Its pretty easy to talk about welding the coolant passages. But a different story to actually do the welding. (I am a welder) You would have to do TIG to get a clean weld that would be strong enough. And its almost impossible to get the TIG torch down in there close enough for good penetration. You can get a MIG gun in there and try to spray weld it, but you will not get good gas coverage. But this is cast aluminum, and even with billet aluminum with good gas coverage, spray/MIG welding will not be strong enough for this situation. And the heat needed to weld such a piece, will warp the housings.

So I'm looking for other ways now.

Another way that i wanted to try is similar to a PP on the intake. Press fit (extreme press fit) a sleeve for the exhaust. I have heard of a few people machining exhaust sleeves from aluminum and them working fine. I don't see how, as i would think that the aluminum would melt.

Then after a lot of searching, i found this...

http://jhbperformance.com/downloads/...racingmods.pdf

I know that you are right, but this will be easier than the intake I think. I intend to lay a small piece of aluminum down in the end of the single passage that I will use for the port and TIG it fast. I think I can do it. I have Inconell rods and sheet, so I figured I would lengthen the exhaust sleeve and weld the holes on the side shut. This will give me isolation between the exhaust gasses and the aluminum. Really the aluminum plug should hold up fine to the 13 psi of coolant pressure that it must withstand. If the housing is too porous and I have leaks, I will use high temp RTV (its just coolant).

It is really neat that the article you found mentions that you should not open the port earlier for the same reason that I want to open it later.

I really want a set of junk RX8 housing to try this on, the good ones I have are too valuable to experiment with. After I perfect the procedure, then I use the good housings.

[scotty305]

Very cool idea, was this influenced by Barry Bordes' in-chamber pressure testing?

Will your new port clear the tension bolt near the top of the current exhaust sleeve?

[GorillaRE]

This will have a very negative effect on engine performance and consistency. The extreme amounts of exhaust reversion this port timing change will cause is something that we are trying to ovoid.

I like your attitude and thinking, but its not applied to the right issue.

-J

[NoDOHC]

Originally Posted by Scotty305

Very cool idea, was this influenced by Barry Bordes' in-chamber pressure testing?

Will your new port clear the tension bolt near the top of the current exhaust sleeve?

Yes, When I saw that there was 91 psi remaining in the chamber at 195 ATDC, I decided that some of that blowdown should be used.

Yes, the port will clear by about 1/16" It is almost like Mazda intended this to be the exhaust port location.

Like Sen2Two said, the machining is not going to be as hard as welding up the cooling jacket and modifying the sleeve.

[NoDOHC]

Originally Posted by GorillaRE

This will have a very negative effect on engine performance and consistency. The extreme amounts of exhaust reversion this port timing change will cause is something that we are trying to ovoid.

I like your attitude and thinking, but its not applied to the right issue.

-J

I agree that exhaust reversion is not our friend, but I don't see how this modification would be anywhere near as bad as a bridgeport. The significant overlap is minimal (about 10 eccentric shaft degrees). Most modern cars have significantly more overlap than that.

If you can explain why this will give terrible inconsistency, please do, as this will save me the trouble of building it.

I am curious myself as to what this mod will behave like. That is why I want to try it.

It really should act a lot more like a factory-cammed piston engine after moving the port than it does now. This should provide for better fuel economy too.

Please read my above post about the cusp on the housing shielding the exhaust port. Even at 1,000 rpm, it will be hard for air to move through a 0.050" slit. The overlap on a bridgeport occurs with the dish clearing the cusp, allowing free air flow.

[GorillaRE]

Originally Posted by NoDOHC

I agree that exhaust reversion is not our friend, but I don't see how this modification would be anywhere near as bad as a bridgeport. The significant overlap is minimal (about 10 eccentric shaft degrees). Most modern cars have significantly more overlap than that.
What modern cars? PISTON POWERED ONES! In these engines, overlap works totally different than in a rotary. During the duration of overlap in a piston engine (a properly designed modern one) the exhaust valve is "sucking" the fresh intake air across the top of the cylinder and into the exhaust. And although in a "tuned" cylinder head/exhaust/intake system it also "pulls" some exhaust back in, it does this in a perfectly timed and proportioned manner. The reason it can do this so well is because the pistons and valves are only doing "two" things: Up and down, open and close. This allows eeeveryone out...aaand eeeveryone back in.
This works very well and is a basic law of physics.

In a rotary on the other hand, it cannot and dose not have the same effect. Once the apex seal has crossed over around 50% of the exhaust port, the exhaust, back pressure from the turbine, natural engine rotation and the intake port acting like a vacuum on the smaller "squish" area on said rotor chamber during that portion of rotation, all cause a negative overlap effect. This equates to a uneven and inconsistent amount of exhaust gases being re-circulated. Since the rotor rotation is moving the chamber through the same steps over and over again, the mixture becomes more and more diluted.

Have you ever really listened to a loud un-muffled rotary idle? Do you know how the idle seems to be fairly smooth/consistent and then all of a sudden not? e.g. brrrrrrrrrruh........brrrrrrrrrrrrruh........brrrr rrrrrrrruh.........breeehhhrrruuh etc.

That is the sound of to much exhaust polution. It builds up and builds up and builds up, then breeehhhrrruuh.

If you can explain why this will give terrible inconsistency, please do, as this will save me the trouble of building it.

During Me and Barry's testing with the ICPS we noticed an UNBELIEVABLY erratic pressur/burn rate during idle, low load and a few other key moments..... can you guess what the erratic behavior was caused by
?
I am curious myself as to what this mod will behave like. That is why I want to try it.

It really should act a lot more like a factory-cammed piston engine after moving the port than it does now. This should provide for better fuel economy too.

Please read my above post about the cusp on the housing shielding the exhaust port. Even at 1,000 rpm, it will be hard for air to move through a 0.050" slit. The overlap on a bridgeport occurs with the dish clearing the cusp, allowing free air flow.

-J

[scotty305]

Would it make sense to open the exhaust port later than OEM rotor housings but close it earlier (or at the same location)?

[NoDOHC]

Gorilla's post is too long to requote, so please bear with the new reply.

Let me make sure that I understand what is going on here:

The exhaust reversion is occurring under high vacuum conditions for the same reason that is occurs on a piston engine, exhaust pressure exceeds intake pressure and overlap exists. Tuned runners don't help this appreciably, as they are tuned for the speed of sound at WOT and the speed of sound is significantly slower at lower manifold pressures. The air also lacks enough velocity to actually dynamically charge the cylinder.

What I hear you saying is that the intake port is basically opening too late, causing the vacuum that is generated in the chamber to pull air back out of the exhaust. Closing the exhaust port later will exaggerate this. I see how this design could cause additional exhaust reversion at high vacuums (ie: cruising speed).

I still think that the reversion would be far less drastic than with a bridgeport. At 32 ATDC, A rotary engine has displaced about 8% of the total volume of the engine. This means that the effective intake displacement is actually only 92% of 0.65L or 0.6 L. This is the effect that the bridge port removes. I don't see how this is changed appreciably by relocating the exhaust port, as the engine already has this effect and already has a small amount of overlap (16 degrees) in the least shielded position. The only overlap added by this modification will be while the rotor face is very close to the housing (should be mostly ignorable).

I am curious how the small amount of exhaust reversion (should never be greater than 20%) can cause the engine to misfire. This may be obvious, but I am not getting it. All my knowledge of camshaft design is from the piston engine world and I admit that I may have drawn the parallel of the two types of engine too far.

The understanding that I have of the above explanation is that the exhaust gas continues to build up in the chamber and eventually the engine misfires. I see how this could occur at very high vacuums (> 60 kPA), but under normal operating conditions, the engine should force enough exhaust gas out each time to keep net air flow through the engine. The exhaust dilution should reach a still-functional maximum.

Kenichi Yamamoto mentions the reversion misfire effects in his paper on the rotary engine. He only speaks of it as a serious problem under deceleration conditions.

Here is some data from page 52 of his paper:


From what I see, it improves low and high end power, while make little effect on midrange. The misfiring data given is at 0 degrees BTDC, While a good idle timing will be ATDC.

I can say that on my '88 4 port engine (Naturally aspirated) I did have misfire issues if I leaned the fuel mix up too much or if I advanced the timing too far (running a lot of intake vacuum) however, the fueling (and hence the VE) matched what I expected from a piston engine. In fact, the rotary VE seemed to be higher at high vacuums than the piston engine. This would seem to contradict any theory that I can form about the exhaust reversion causing the misfiring, as any reverted exhaust is directly subtracted from the VE of the engine. If the timing was retarded enough or the fuel mix was rich, the car idles (I should say idled, as it barely runs right now with my fuel system problems) smooth as butter.

[warwickben]

Any thing cone from this?

[sngl13b]

i just saw this and was wondering the same thing..that jhbperformance has some nice stuff and good info on their site.

[NoDOHC]

I have not started this build yet, I have almost all the parts i need, I just need time and ambition.

My hope was to dyno an engine with normal housings and then re-dyno it using modified renesis housings to determine the performance difference.

I will update the thread when I get a chance to test this out. If someone else would like to, please feel free.

Disclaimer: I do not know if this will help power, hurt power, move power band, make completely undriveable, etc. This is a test I want to run. The reversion may very well make it completely impractical.