45 replies to this topic

[xFactor]

So, I started it up last night, and it ran well, but sounds like my subie. Idled 14's afr after warm-up. Short story long, the #3 and 4 holes run about 80 degrees cooler than the 1 and 2 using a temp gun on the header. We swapped the plugs, wires, and injectors around to see if the problem moved, but it didn't. Then we did a compression test...



Yeah, the coldest cylinder just made 255psi on ten cranks. Everyone else was around 240 psi. Pretty sure that isn't the problem...

We've ordered a 4 vacuum gauge manifold from teh ebay. My plan is to run each gauge to it's own tb, use a hand-held vacuum pump on the map sensor to bring it down to 18" or so to fake a good idle signal, and not worry about the fuel pressure regulator for syncing them together. I also flipped some bolts around on my fuel rail adaptor plates so I can get to my adjusters at idle.

I figure that you can't tune each tb when they are all tied together normally through the mini manifold (aka balance tube) that feeds the map and pressure regulator, right?

later, matt

Edited by xFactor, 12 December 2012 - 08:07 PM.

[xFactor]

The gauges arrived, and we hooked them up according to the plan. 18" on the map sensor was too lean to start the engine. 13" was the magic vacuum on the map sensor to let us start, idle and crack the throttle during adjustments.

First off, the gauges are in centimeters. Second, I've never done this before...
Cylinders 3 and 4 were pulling 60cm, and 1 and 2 were pulling 20cm, meaning the 3 and 4 throttle plates were completely closed, and the vacuum on them was closing them to the point of completely sealed, and 1 and 2 were doing all the work. It took a few minutes to figure out which way to turn the screws, and I started to balance everyone out. But every time I got them close, the idle would be way too high, like 2000+ rpm's with the main idle adjuster backed all the way out.

After an hour of fail, I pulled the tb's off the manifold and looked at them with a light behind the plates. It was just wrong. The light showed #4 was the tightest seal, gradualy opening up more each plate up to #1. The throttle cable is coupled the the number 4. I figured start there, and it would be a cascading effect down the line. What I didn't figure on was when I cranked down an adjuster to close it, it would exert pressure on the other tb's messing up any previous adjustment, making the others open a bit.

I dialed everyone down to an almost tight seal visually, with a flashlight behind the throttle plates. I found that the slightest change, like 1/8 of a turn, would have a dramatic effect on the adjacent plates. Hopefully this is my new zero point.

So thats where we are at. More to come,

Matt

[Mark L]

[cbstdscott]

Love this thread and follow every step.

Help me understand the build process: Did you start with four individual tbs or did you start with a set of four from a motorcycle? It seems to me if you keep four motorcycle tbs together you avoid these issues you are having. The motorcycle factory has built the ITBs as a unit and they should all draw the same.

I see from your manifold design that the motorcycle spacing may not align with the engines cylinder spacing. Straight runners are always best. But some compromise might be in order.

Scott

[cbstdscott]

Love this thread and follow every step.

Help me understand the build process: Did you start with four individual tbs or did you start with a set of four from a motorcycle? It seems to me if you keep four motorcycle tbs together you avoid these issues you are having. The motorcycle factory has built the ITBs as a unit and they should all draw the same.

I see from your manifold design that the motorcycle spacing may not align with the engines cylinder spacing. Straight runners are always best. But some compromise might be in order.

Scott

[xFactor]

They all came off as one unit, and at one point in time they were probably all synced up. But even if they were still on the motorcycle, eventually they would need to be re-tuned.

Little inconsistencies between cylinders, like valve lash, will effect how they pull vacuum, so I'm pretty certain however I cobbled this together it would still need tuning.

If I would have kept the motorcycle spacing, The outer runners would be a little longer, and in my mind that seemed to be a bad thing. I also would have needed to use the cbr fuel rail, and it had a different pressure regulator assembly that I wasn't too fond of.

I jumped into the deep end, and now I need to learn to swim .

I appreciate the input,

later, matt

Edited by xFactor, 13 January 2013 - 05:37 PM.

[xFactor]

Finally back after the holidaze. It seems my elves put everything back together for me, and they did it right. Bonus.



It also seems my dial in strategy worked out. Extra bonus. It now sounds like an angry Honda, not a pissed off Subaru, just the way it should.

http://s75.beta.phot...4e5690.mp4.html

But now with everything plumbed back correctly, it runs way too rich, like 9's. We added an adjustable fuel pressure regulator, backed it down from 48psi to 37psi, and it now idles high 10's low 11's.

With a vacuum gauge tied in showing MAP vacuum, we saw about -15" hg, and the reading was very smooth, but I'm not sure if it's low enough with the cam regrind and the itb's.

Can anyone chime in with vacuum readings on their ew-4 intake manifold at idle?

Later, Matt

Edited by xFactor, 13 January 2013 - 04:28 PM.

[xFactor]

Getting ready for the dyno involved fixing a few things.  The throttle plates were binding, better brackets, throttle stop, blah blah blah...

 

Ok, screw the bs, time for the numbers.

 

On a Mustang Dyno!!!  Dynojets usually always read higher due to less friction...

Old combo from a year ago 92hp 92 tq vs. new combo 137hp 118tq

 

The Mustang dyno has two rollers

 

A little bit of cage...

 

 

Good dyno videos...Click to watch.

 

 

 

It rolled onto the dyno very happy.  Nothing we did during the session made more power.  It didn't like any more advance than 23dg static timing (32 adv).  It didn't like 3dg cam advance either (keeping ignition timing the same).  Afr's were something we were also happy with and didn't adjust.  It still idles rich, but under acceleration load started out high 12's, into an leaner mid rpm spot of 13.7's and fattened out to low 13's up top.  If I really want to worry about the idle, I'll put a micro switch on the throttle and trim back some MAP signal.

 

Track day is next....hellz yeah

 

later, matt

Edited by xFactor, 28 January 2013 - 06:31 PM.

[cbstdscott]

What are you doing for a MAP sensor? Are you using an O2 sensor?

The stock ECU and sensors are not nearly as accurate as modern systems... but they do work pretty well. Stock fuel pressure is adequate for all but turbo applications. The stock system does a reasonably good job of lengthening the injector cycles to deliver enough fuel.

I am still working on my intake manifold design. I think I am going to use a single tb but I am still working out the runners and plenum. The magic ingedient: velocity stacks.

[Buford]

Hey cbstdscott, I still have those ITBs FS, I need to start selling some good stuff that I'll never need now.  PM me an offer.

[xFactor]

Stock MAP sensor, Innovate LC-1 using the 1volt simulated output to the ecu.  Integra ecu and injectors at 48 psi idle.  The sensors are fine, and the ecu isn't that bad.  I've had good luck manipulating it on past turbo builds via MAP signal modification and different injectors.

 

I'd still like to find some sort of dirt cheap inlet trumpet...

 

We've done a few nitrous ew-4's before, and this combo makes as much power as they did on the bottle.  45 more hp and 26 more lb/ft, I'm excited to see how it runs on the track.

 

Later, Matt

[cbstdscott]

"Innovate LC-1 using the 1volt simulated output to the ecu"

I am sorry, I do not know what this means. Is this what you are using for the O2 sensor? If so, why not use a regular O2 sensor?

Have you tried stock injectors and ECU? I think your rich condition is caused by the Integra injectors and ECU. Contrary to popular belief they do not play well with the EW.

Scott

[dwend]

The Innovate LC1 is a wideband O2 sensor. It has a 0-5v output for wideband and 0-1v for a simulated narrowband. The wide band output usually goes to a gauge or if the ECU can read wideband both. If the ECU can't understand wideband, you can wire up the simulated narrowband to the ECU and it will behave like stock.

If the AFR are good when running I would not worry much about them. How often are you stopped in chumpcar? If you pull in the pits you turn it off to do the driver change and fuel.

Edited by dwend, 28 January 2013 - 10:05 PM.

[xFactor]

The LC-1 is a wideband controller...the 1v output from it "simulates" a narrow band o2 sensor output, but with better resolution.  A 1-wire o2 sensor is only good at reading .5volts.  I can explain more if you want...

 

There is also a high speed gauge output too.

 

 

For us, the Integra ecu/injector combo has been fine for several builds. It idled stoich on all the old ew-4 manifold builds.  I fired up a stock 3g civic si we have, and it pulled 20" of vacuum, vs. our 15" on the itb engine, and that's why we idle rich- less map signal..  I tried to trim out fuel pressure to clean up the idle, but it was way too lean under load.  In reality, the engine is now on the verge of out-flowing the injectors.  There were a few spots that could have used more fuel, even at 53 psid rail pressure.  We are way beyond oem crx injectors...

 

later, matt

[cbstdscott]

Thanks for answering my noob questions.

Scott