I see a use for a temp controlled valve... seems like it would work... Or RPM controlled, however that would be harder.

or just buy a stage 2 and do it the right way without messing around!!!
phil

the whole point behind this is for those of us who can't afford Stage 2, or, in the case of the OBD-II guys it isn't available for :)

and i thought you used to be a dsmer :( :( :(

Just kidding haha. Seriously though, just for having your MAF not max out, running a say, 100*F temp switch (20$) that opens and closes a valve (20$) in a tube that bypasses the MAF (10$) seems cheap.

Although I'd kinda like to know why subie would use the MAF to control idle...

Anywho. its all good.

I don't believe the MAF is used during idle... Mychailo you have the ability to monitor MAF voltage...what is the reading at cold idle and warm idle?? also some in between. The Idle Air Controllers should handle that as the engine runs in open loop when warming up and closed loop when fully warm and shutting off the SIAC and using only the IAC to function for breathing at idle. This is only an assumption so don't chew me out for it...

I agree that the Stage 2 is the proper thing to do for anyone with OBD I but the OBD II guys might be able to utylize this.

Tom

i like this idea. so i was reading the charts and is it better with the hose spliced or just curving around everything. i still need to figure out where to drill into my cold air intake that runs in the fender. how big of a hose is to big to use?

I'll get some data logs tomorrow. I'll do one with the engine stone cold, then one when its lukewarm, and then when its full warm.

My general observations are that the engine switches to closed-loop fairly quickly after the car is started. It generally takes only two minutes before its running in closed-loop. When the car is lukewarm, there is a tendency when the car is idling for the afr to swing from ~13:1 to as high as 17.5:1, and the stalls occur when it swings super lean.

KC, I'd recommend first trying 3/4" ID. That's the largest you'll want to use with an Impreza RS fpr. If you install an adjustable fpr and go with a higher fuel pressure, then you can try larger diameter hoses as you increase the fuel pressure, but you'll have to be very careful about building in too much ignition advance.

i see i see ty

One thing that has not been discussed is that the smaller bypass tube has fluid-dynamic (aerodynamic) differences other than just being smaller than the main intake tube. If you were to plot the flow resistance over airspeed, you'd notice that the resistance will increase on a curve. If you compared the curve of the main intake tube to the bypass tube, the curves will be different. (The small tube will have a much steeper curve.)

Here's another way of saying it: Let's presume the total area of both tubes is 10 square inches. Let's say the small tube is one square inch and the large tube is 9 square inches - 10% and 90% respectively. At idle, each tube will handle almost exactly their respective percentage of air flow. At higher air speeds, the small tube will handle a lower percentage, say 8% while the large tube will handle more, 92%. At the extremes, the difference could be so large that the small tube might as well not even be there. Honestly, I don't have any precise values. It might be negligible, but then again, it might not. Those joints and right angles in the small tube won't have much effect at low flow rates, but at high rates it'll really add up. The smaller diameter alone will make a difference simply because the aerodynamics are different on a smaller scale. It sounds like the primary goal is to fix a problem that occurs at higher speeds. Unfortunately the bypass tube will function best at lower speeds. Perhaps you could install a vacuum-actuated valve in the bypass tube that opens when the engine is under load, or maybe a vacuum-actuated electrical switch that changes the output value of the MAF.

Just something to think about.

That's absolutely true, UberRoo. I figure that at 6000 rpm the engine is drawing about 350 cfm of air (the volume of air at standard temperature and pressure would be a little less because of reduced pressure in the intake tube). The inside diameter of the intake tube is about 3", giving it a cross sectional area of about 0.05 sqft. That means the average velocity in the intake tube is about 7000 fpm. Pressure drop across the intake tube at that flow is about 0.03 psi, including the bend, but not including the corrugations (I can't figure that until I do the CFD). An equal pressure drop through a 1 inch tube would yield between 16 and 18 cfm (only about 3200 fpm average velocity). So, even though the cross sectional area of the 1" tube is 11% of the 3" stock intake tube, the flow rate through it would be only 5% of the amount flowing through the main tube at 6000 rpm. That's because the wall/flow area is so much greater in the smaller tube, so that frictional losses rise much more quickly in the smaller tube as flow and velocity increase.

Made a big conceptual error in my last post. Forgot about velocity pressure. It takes a lot of pressure difference to get the air in the intake tube moving at 7000 fpm; in fact, the velocity pressure of 7000 fpm is over 1.3 psi. So, if you look at the total pressure required to push flow through the two tubes and balance them, you wind up with 332 through the 3" duct and 28 through the 1" ID bypass duct. If you use a 3/4" ID line to bypass the MAF, you'll get about 15 cfm going through the bypass and 335 cfm going through the main line at 6000 rpm.

At 1000 rpm, you would be pulling a total of 58 cfm through the intake system, and about 53 would be going through the MAF and 5 cfm would be going through the bypass. With a 3/4" ID bypass, you would wind up with about 2.5 cfm going through the bypass and 55.5 cfm going through the MAF.

Dan,

Your numbers say that at both 1000 rpm and 6000 rpm, about 4.3% of the air will go through the MAF bypass. Is this correct?

Finally had time this weekend to get the MAF and AFR data for the smooth tube configuration. Results are excellent. From 3000-5000 rpm, the afr is right around 13.2:1 which is right about where it should be for maximum power, and then from 5000-6500 rpm, it settles at about 12.4:1 which is also pretty good. Nothing that can be done about the richening of the afr past 5000 rpm. The ECU is doing this for some reason. It shows up in the non-bypass afr curve too. If I compare these afr results with the non-bypass AFR data, it looks like about 5% of the air is going through the bypass tube regardless of air speed which is right in line with Dan's predictions (if I interpreted them correctly).

http://www.subaru-svx.net/photos/fil...czko/31494.gif

Here are the MAF, AFR, and rpm data vs time for cold start and engine full warm. For the cold start condition, the afr starts swinging up and down when the ECU goes into closed loop mode. For temperatures below probably 40F, the swings can become great enough that the engine will stall. If I compare the position of the spikes in MAF voltage, rpm, and AFR, it appears that the spike in the MAF voltage just slightly precedes the spike in the AFR data. That would suggest to me that the AFR is affected by the MAF signal. It would seem more logical to me though that the afr fluctuations are preceding the MAF and rpm spikes. I would have guessed that the ECU is trying to lean out the afr, and the lean condition is killing power which causes the rpm to drop and the MAF voltage to drop. Anyhow, as best as I can tell, the car doesn't stall during warmup if the air temp is above about 40F.

http://www.subaru-svx.net/photos/fil...czko/31496.gif

Just an observation. You could potentially shorten the bypass dramatically by making it only long enough to go from the airbox to just after the MAF. By shortening the length, I believe you would reduce the low-speed vs. high-speed airflow difference, and perhaps help keep the top end AFR closer to 13.0.

Instead of tapping the plenum, just tap the elbow right after the MAF (see picture). Just make sure things are as smooth as possible on the inside of that elbow, or you could be doing more harm than good!

BTW: This is some nice creative, old school, engineering. :D

Mychailo:

Yes. That's what my numbers say. One thing I'm not taking into account is that you are bypassing the corrugations on the main intake tube. I really want to know what is the effect of these on flow, but our CFD shop is booked up with one of my projects for the next week or so. Have prepared measured drawings of the intake snorkus to input to the computer, just don't have a worker and workstation to spare right now.