Defined Autoworks Header Install!

[j9fd3s]

i haven't checked backpressure on the P port yet, but it is on the agenda shortly. i did check it on my stock FC though, and it made tuning so simple its funny, i don't know why we don't use it more.

[Libor]

Originally Posted by j9fd3s

i haven't checked backpressure on the P port yet, but it is on the agenda shortly. i did check it on my stock FC though, and it made tuning so simple its funny, i don't know why we don't use it more.

Could you contribute more? This is something very interesting, unfortunately not much examined on public rotary forums

[j9fd3s]

sure, um the gsl-se and the 86-88 NA cars make doing this very easy, so its a good place to start. my car happens to be 100% stock, mileage is 76K now, 74k at the time.

so the 6 port cars use exhaust backpressure to move the 5th and 6th ports, so all i did was tee a boost gauge in the 6 port hose, and then go for a drive.

the results were interesting, the gauge and the butt dyno agreed 100%, or more accurately the boost gauge and the tq curve agreed.

the stock exhaust did ok, up to about 4500, then backpresssure starts to climb. so its under 1psi, until 4500 then it comes up pretty quick and peaked around 4.5psi.

i pulled the cats off, and behold the center cat was all melty, so i hollowed it out, and was rewarded with 3.5psi of backpressure. seat of the pants improvement was bigger than the SAFC.

the next step would be to put the gauge actually in the exhaust manifold, i'm not sure but i suspect i'm measuring the rear half of the exhaust. as the pickup is in the main cat.

the concept works though, and its easy to do. higher exhaust pressure = lower power.

[NoDOHC]

Update:

My previous test setup was not as accurate as I had hoped, so I decided to build a setup that could graphically display what the exhaust was doing.

To that end, I used two Ashcroft K1 150 psi pressure transducers in the exhaust to measure backpressure (watching the pulses on an oscilloscope). These transducers are the same ones I use for compression tests.



Unfortunately, I didn't think that the neighbors would appreciate me starting the car and reving it up right now, so I will have to wait for tomorrow to collect data.

Anyhow, what waveforms do you guys want to see?

I can tell that the headers scavenge well at higher speeds, but I can't get a good scope trace there as the engine revs too fast.

I watched the exhaust pressure at idle and was appalled by how frequently the engine misfires. About every 15th power stroke is a full and complete combustion, the rest are partial or complete misfires. I think that 14.2:1 is a bit lean after all.

I will post some traces tomorrow - what would interest you?

[NoDOHC]

Here are a few waveforms with open headers.

The time scale is listed on the bottom, the pressure scale is 3 psi/division.

This is what the exhaust pressure looks like at Idle (1000 rpm due to e-fan):


Revving from 3000 - 5000 rpm, WOT:


Revving from 5000 - 6000 rpm, 40% Throttle


Revving from 7000 - 9000 rpm, 50% Throttle:


Bouncing off the rev limiter (10000 rpm):


No wonder rotaries are loud! A piston engine exhaust output looks much more like a sine wave, this looks like a triangle wave (more high frequency components). The blowdown (pressure remaining in the chamber when the port opens) is ridiculous. This is probably due to the early exhaust port opening (75 degrees BBDC).

It is easy to see the dip in the pressure that is caused by the header scavenging. This dip moves relative to the wave due to the wave frequency changing and the header length not changing.

By looking at this data, it appears that the headers would produce optimal tuning about 7,000 rpm (we want a minimum pressure on the exhaust port as it closes - negative pressure is not indicated my my transducers).

[RICE RACING]

Header wrap on a Peripheral Port will destroy your headers! Take it off and make a heat shield mate

[NoDOHC]

Thankfully, this isn't a P-port, it is a stock-timing mild street port. The EGT is seldom over 1500F even at WOT (although I haven't had it on the dyno since I installed EGT thermocouples).

You may be speaking to j9fd3s (who has a P-port) as you can probably tell from the pictures that mine has the stock intake manifold and is therefore not a P-port.

Anyway, I did some testing after adding this:


The waveforms changed considerably with the addition of a collector.

Before I show them I should explain that 100 mV is 0 pressure and these transducers read as low at 1.5 psi negative. Some of the pressure seen below is going to be significantly negative. This means that the headers are working.

Most headers require a collector for proper scavenging.

Anyway, here is idle. Please don't let the weird waveforms confuse you, the engine is only firing once every 60 ms (acoustic and sonic properties of the headers/chamber are causing the rest):


Here is revving from 2500 - 4500 rpm:


Here is revving from 3000 - 5000 rpm:


Here is revving from 4000 - 6000 rpm:


Here is revving from 7500 - 9000 rpm:


Bouncing off the rev limiter:


I will do a writeup of each change when I am done. Right now, I am building the exhaust system piece-by-piece and re-collecting data each time.

[j9fd3s]

i'm not sure i know what i'm looking at, but that's very interesting!

[j9fd3s]

Originally Posted by RICE RACING

Header wrap on a Peripheral Port will destroy your headers! Take it off and make a heat shield mate

even not wrapped the headers are probably going to live a short life...

[GtoRx7]

You can leave them wrapped NoDohc, 321 stainless is some very tough stuff. I personally ran my n/a 3 rotor with wrapped headers for over a year. Un- wrapped we have have ran our headers on my daily driver for over 4 years with thousands of dyno pulls, racing, snow, ice and rain. Never once failed or cracked.

[GoopyPerformance]

Originally Posted by GtoRx7

321 stainless is some very tough stuff. I

Yes 321 stainless is very good tubing...and Defineds welding is very nice..I have seen it first hand!

[NoDOHC]

Thanks Logan! I thought that the stainless would be ok. Header wrap kills headers because it holds moisture against them. I doubt that any 3 series stainless will corrode significantly from simple moisture.

As the pictures show, the only thing that separates these headers from a true art form is the usefulness of the headers.

I really need to take more backpressure readings now that I have my exhaust completed, I have been super busy and haven't had the time to update this thread, maybe later this week. I used dual Gibson Superflow mufflers with dual Flowmaster dBx mufflers behind them. Exhaust tone is pretty good (not raspy), but not as quiet as I was hoping.

For what the butt dyno is worth, this car pulls very hard. Unfortunately, I can't give tirespin comparisons for anyone, as I have a set of 215/65R15s off a minivan on the car right now and they spin effortlessly. I wore my 245/40R17s out and my Bridgestone Ecopias are on my '91 (I don't spin them anyway).

I picked up some low end torque, but the real gains were above 4,000 RPM (after I installed 4 real mufflers in place of the tractor muffler).

With my tune which previously ran 13.2 on the dyno I was running 17.2:1 AFR at 7200 RPM. The largest VE gain over the racing beat headers was about 5,500 rpm, where the engine actually lean misfired at WOT (lean of 18:1). I can't explain this, as the headers are not tuned for 5,500.

With the minimal overlap this engine has, I am surprised that there was that much VE on the table.

Strangely, my 6,500 map point was correct (no changes required).

My cruising MAP has decreased, as has my idle MAP (although it is still very low vacuum, I am running a MAP of 45 kPA at idle due to my 5 degree ATDC idle timing and 14:1 AFR).

I was able to go 0.5 AFR leaner at idle without misfiring. The cruising vacuum increase is a pain because I am running too much vacuum to run lean of stoic. (My cruising MAP at 60 mph is about 50 kPA, which is very close to my idle MAP).

Above 9,000 RPM I was also lean misfiring, but I never got this tuned on the dyno, as I always let up on the car about 8500 RPM as the coolant temp was above 110 C.

I really need to go to the dyno again, but I am finding it hard to part with another $80.00.

[Libor]

Originally Posted by NoDOHC

With my tune which previously ran 13.2 on the dyno I was running 17.2:1 AFR at 7200 RPM. The largest VE gain over the racing beat headers was about 5,500 rpm, where the engine actually lean misfired at WOT (lean of 18:1). I can't explain this, as the headers are not tuned for 5,500.

With the minimal overlap this engine has, I am surprised that there was that much VE on the table.

So if I understand correctly, you´re pulling 30% more airflow at 7200 RPMs?

[NoDOHC]

That is a good question, I compared fuel maps and I only added about 12% to the fuel to get it back, so I am guessing that the engine was misfiring. I have trouble keeping the AFR constant at 1800, 2400, 3600 and 7200 RPM. The AFR moves between 13.2:1 and 15.2:1 right now at 7200 rpm, I had to put a filter in the E8 for the MAP sensor, as it was seeing 5 kPA boost at 7200 (the MAP was jumping around about 10kPA).

This is the effect that I was not expecting. The headers dramatically change the VE, I had to move my RPM points around to fix a newly acquired '3600 rpm hesitation' which results from the MAP reading being very unstable at this speed. The VE peak (about 5%) is very narrow, as I need a 3550, 3600 and 3650 rpm point in my map. I was able to get around the 1800 and 2400 peak by just letting the engine go lean there, but the 3600 rpm VE peak was actually causing a misfire.

I may have to revisit the 2400 rpm MAP dance later, as it becomes a little problem when I am trying to cruise at 60 MPH (I usually don't drive 60 mph, but I was in traffic at that speed yesterday and there are still some driveability issues at that speed).

I haven't gotten to the dyno, so the 7200 rpm point is still running 'close' I can't tune that spot on the street, as it revs way too fast in second gear to do anything and I would be going > 100 MPH in 3rd.

The 5,500 rpm point is very weird, as I get misfires there even now, but my AFR is good above and below this point. I really need to get to the dyno, as even 5,500 rpm is a little fast for street tuning, although this could possibly be done in 3rd gear.

EDIT: With the minimal overlap this engine has, I thought that a MAP tune would be ok, now it is looking like a TPS tune may be the best solution again. (I hate TPS tunes because they are so temperamental.)

[j9fd3s]

hmm i don't really like MAP tunes on NA engines, it seems like cruise and idle are always the same spot! the last car i did though was a 56 chevy so it doesn't rev over 2000rpm until about 50mph!

i've actually thought about running some sort of air flow meter next time, a modern one, like an Rx8 or something, it SEEMS like it might be a little simpler to tune.

-mike