Damn reading this thead makes me wish I never sold the 9.0 rotors out of my 20b to buy the 9.7's.

 

Well, there is definitely a good reason Mazda progressively upped the compression ratio of rotaries through their production!

The number one complaint consumers have with the Mazda rotary is lack of torque, by which they mean low rpm power.

The higher compression rotors definitely provide better throttle response and tip in/low rpm torque.

I think the rotary itself proves the best race engine isn't necessarily anything like the best production engine.

 

The higher compression also likely improves emissions.

Heck, one thing that stuck in my mind in a discussion about compression ratio is that most (piston) automotive engines have compression higher than what is required for best WOT performance. The compromise is made for better part throttle economy since cars spend most or even all of their lives at low load/low RPM.

I know that is certainly true for turbo engines but hadn't thought it to be the case for nonturbo engines! Given the amounts of timing pulled out by the OEM maps I've seen, it's probably more a case of tuning for the Lowest Common Denominator - the proverbial guy towing an overloaded trailer up a mountain in August and ignoring the temp gauge or the marbles rattling around under the hood.

 

If you are racing and the rules allow you should run race gas.

On a piston engine with race gas you will be able to add all the timing back in on a high compression motor- like wow.

Friends tuned a basic Honda B16 (type R EVERYTHING! *cough* which is stock in Japan ofc). Gained nearly 30hp on the race gas map with very aggressive timing. Nearly 30hp added to a 135hp motor is a lot!

On my 8.5:1 turbo rotary on race gas- we never noticed power gain adding in timing, but did see power gains taking more timing out of the base map.

Adding timing in on low rpms sure was better for throttle response, so we stopped pulling timing in the low rpm for more power LOL.

On a loading dyno you have to remember to tune for transient response (opposite of tuning on a dynojet where it is all transient) because if you lower timing + add fuel in low rpm on a turbo rotary you gain power spooling the turbo.

That might help you pull away from 1,500rpm in 5th faster, but it doesn't apply to performance driving. Gotta do the sweeps.

 

Dynoed sat night in a hurry before race on sunday. First run laid down a solid 290 hp. added some more timing and a bit more fuel pressure and smaller air correctors as top end was too lean. put down 306hp with still too lean top end, also torque was really good, around 260-265 Nm at the flywheel. left it there are there was no more time. Car ran really well during the race. 500hp camaro which weights only 100 kg more could not overtake on the straights

 

Right on!

So, it was just the exhaust port/sleeve was too small?

 

yep. Also it seems the myth everybody is telling to folow about the late exh port opening as described in old mazda comp manual of around 73-75° is not really valid. I am opening exh now at turbo specs (82 deg) and get strong midrange (6-7k rpm), with peak troque of close to 270 Nm

 

After reading your other comments, I'm confused on what you actually did to the exhaust to find the improvements?

I haven't gotten to the point of dynoing my 20b just yet but I built my exhaust sleeves into the header flange (so I have no sleeves at all). I did this based on a thread RG had a few years back when he was making custom exhaust sleeves and the discussion was about port flow. I'm also running 1.75" headers and have not ported the exhaust as of yet as I'm trying to get a base and then port as needed later.

 

I first had stock opening at 74deg (same as mfr spec) and rx4 exh sleeve (almost constant cross section like mfr housings I had) both mfr and rx4 housings I made produced same number (270hp and 230 Nm at the fleewheel), which is low considering they both had 52mm ID intake port with 80 deg closing.

Now I removed the rx4 sleeves from the housings and machined hole to accept TII sleeves and changed port to open at TII specs (82 deg). First run without changing anything was 292 hp and 260 Nm, with smaller air correctors and more timing advance it makes now 306 hp and 270Nm, picks up better at 5k rpm then before too on track.

 

interesting. the factory literature, i think uses BSFC, so i guess we just substitute power? it may also have something to do with the rest of the combination, runner size, intake, exhaust, carb, or maybe just the thing i'm sitting on

 

Actually the engine with the small sleeves felt the same as a renesis, didnt react much on mixture or timing changes. With the rx4 sleeves we got near same peak hp regardless of intake length, jetting or timing. Now with the bigger exhaust we picked up 20hp with 3-4 deg advance and a bit more fuel.

 

Ok! I'm just not that familiar with the early rotaries porting configurations.

 

Yeah, my though is that people that try to run a small exhaust sleeve to match the exhaust port opening have to think more about the dynamics at work not a static picture.

What works on the flow bench (laminar flow) may not work on the engine where the exhaust volume is rapidly expanding as it leaves the port.

Our little internal + external combustion engines.

Then there is the fact the TII/REW exhaust sleeves have that big flat angled top side that can reflect the rotaries strong reversion pulse back down/out instead of right back into the exhaust port.

Anyways- more of my theories (like the CR thing).

What we do know is that on a turbo set-up the small exhaust port sleeves/exhaust runners really spool up a turbo, but EGTs go through the roof melting exhaust wheels.

I theorize this is because the exhaust gas expansion is delayed until it hits the turbo exhaust wheel expansion area.

Maybe on an NA we could use this effect as well?
Keep the small exhaust sleeves along with small header primaries and then after the collector go into an (pulsejet) expansion chamber with a large 3" exhaust at the end.

Could put a low pressure zone at the header collector.

So, if you have the small exhaust port sleeves- take heart and experiment with the rest of the exhaust before you tear down the engine to replace the sleeves.



This design uses the Coanda effect. Pretty cool!

 

we know of a few different ones. the 77 Rx3 competition manual lists a round port 12A housing, 1058-78-100A. i think its 86/75 intake timing, and 73/65 exhaust.

then there is the Rx7 4352-10-100 12A and 4801-10-100 13B housing. its a D port, same timing as above.

then there is the one in the 900032 SAE paper, its a D port, ports are wider but the intake timing changes to 80/80 and exhaust is 73/58. no known part number.

the R26B housing is again different, if the drawing is to scale it has bigger ports and runners, no known part number, although i would try K0BY-15-G0D

 

Hmm I still think that past a certain point the small area exhaust port is just a restrictor. there will be no way to get more gas through it unless changing the density which is in a NA not going to happen. by just opening the ports sooner I gained 20hp and 30Nm, without even touching the carb or iginition.

I massaged the TII sleeve roof to match the port and thus it is higher and expands slightly less, I made the exh port as high as I could in the housing to maximize overlap before hitting water (one housing I had to reweld)

Actually I run an expansion chamber, which now in all in a sudden seems to work well as Im seeing close to 270 Nm at the flywheel.

 

the factory setup was sort of like that, header collector into a megaphone, i guess using the entire world as an expansion chamber... if you've ever poked your head under the green #7 IMSA car, those header pipes are TEENY

 

by just opening the ports sooner I gained 20hp and 30Nm, without even touching the carb or iginition.

And changed the exhaust port sleeve right?

Hmm I still think that past a certain point the small area exhaust port is just a restrictor. there will be no way to get more gas through it unless changing the density which is in a NA not going to happen.

Yes.
But I am providing hope for t-von/John Huijben and others that have built an NA motor with laminar flow exhaust sleeves/small primary diameter that they can work with the system they have and see gains over their baseline with a proper exhaust including expansion chamber. Maybe they will find some low rpm power before the primary diameter becomes a choke point in high rpm power.

I agree with your race set-up that allowing exhaust port expansion for high exhaust gas velocity at the port where the rotor shape is creating an intake low pressure area during overlap coupled with larger primary diameter is best.

In fact, after I saw your port pictures I thought you could improve your power as your D shape exhaust port looks upside down to me.

If you flip the D you get very high velocity gas at port opening to punch through the exhaust pulsation gas slowdown and the flat top would maximize port area available for the beautiful rotary overlap intake scavenging.

Obviously, just more theories as your motor is working beautifully now!


Actually I run an expansion chamber, which now in all in a sudden seems to work well as Im seeing close to 270 Nm at the flywheel.

Clearly working! It is nice to have such a graphic illustration of the rotary external combustion in effect.

 

i guess using the entire world as an expansion chamber

It is no wonder racing organizations clamped down on decibel limits

I did some off the cuff calculations using Yamaguchi's published #s for race rotary exhaust decibels before and after the 90lb Mazda competition lava rock muffler.

Before they added the 90lb muffler the race engines were slightly quieter than a top fuel dragster. That is a noise that rearranges your internal organs from the track to the stands for those that haven't been exposed to that sonic assault.

 

i wonder if exhaust pipes similar to 2 strokes would be better suited.. like this guy..

Ever Wonder What a 2 Stroke V-8 Engine Sounds Like? | Muscle Horsepower

it's my understanding the rotary exhaust tuning on PPs is very similar to 2 strokes in the fact intake and exhaust overlap heavily.

 

I opted for the inverted D port because a wide flat closing edge I was thinking to be a little risky, probaly I am just being to much on the safe side here. The other reason was that a straight opening edge would create a really strong pulse as soon as the port opens. this pusle would serve to help to scavenge the other rotor due to the 'tuned'primary length from port to collector. also I was thinking that the pressure is the greatest when the port opens, so the more area there would mean a higher massflow through the port. and everything that is out, is out and could not dilute the fresh charge during the overlap phase.

just wild idea in my head. one really needs a realtime pressure sensor in the port to see what is going on.

Oh and I measured the pressure at the EGR port in the side housing (is open to the volume between the sleeve and the housing) and at high rpm maximum pressure is around 50-60 mbar. Less is better but I think it means that the collector area is still sufficient and rest of the exhaust doenst from much of an obstruction.

 

^ Since we are on the subject of measuring back pressure, what is the best location to do so? Wouldn't it be at the collector? I'm thinking of rigging up something while my engine is out.

 

Depends, for the engine what counts would be at the port. but we can only measure average pressure, what really counts is realtime pressure depending on E shaft angle but we need a quartz sensor for that :/

But Still average pressure at the port will be a good indication to see if the primaries/collector/rest of system forms much of a restriction. if you measure 'high' average pressure at the port you can measure further downstream to find which part(s) are the bottleneck

 

i put a gauge in the 6pi hose on my last S4 NA, and its quick and dirty, but i found it was very eye opening. basically the gauge = seat of the pants power. you could really feel the engine loose hp as the pressure in the exhaust stacked up.

i was gong to say that because the exhaust is open at one end, you only get a reading after that, but mine was initially high, i dropped the exhaust and found a clogged cat. once that was fixed, pressure dropped and car was faster.

car was stock, but i had it down to about 3psi @6500rpm, which is as fast as i could find room to go.

 

How fast is a MAP sensor. Could we use it for an exhaust pressure sensor and record the 0-5v signal?

Who makes a cheap quartz sensor? How fast are they?

I need something like 22k cycles/min if possible. (360 degrees X 60 sec)

Then we can compare pressures at overlap.

Barry

 

Being an analog sensor, there is not a refresh rate. The limiting factor is going to be the sampling rate of your logger. MAP sensors are commonly used on high end applications for just this purpose.