Can't get a smooth idle with homemade ITBs

[Metro]

Since my car was running so well last fall I decided it would be a fun to ruin it by building a turbo ITB setup. It's very basic using GSXR ITBs clamped on tubing with DCOE flanges on a TWM manifold. I went with the TWM as it was the shortest easily attainable manifold I could find. After many months of small successes and large failures, I discovered my "new" TWM manifold was machined unevenly, not to mention the bolt holes for the #1 carb put it 2mm out of center. That was very annoying to say the least. With a belt sander and some flat aluminum stock, I was able to get it leveled and as best I can tell, it's sealing.

MS2-extra v3.3.2
L28et ("A" cam, but otherwise stock)
450cc DSM injectors (flow tested, 2.3% variance)
46mm GSXR ITBs
Aeromotive RRFRP set at 37psi or so
GM DIS coil packs with BIP373 ignitors inside.
No idle control (yet)

 

So my problem is that I cannot get it to idle smoothly. It will pick up, then miss and drop 150rpm and repeat. It's consistently inconsistent. My AFRs are bouncing around from like 12.5 to 14.2 as it stumbles and catches. At this point, I don't believe it has anything to do with my Megasquirt configuration, I can't help but feel it's something endemic to my setup. I haven't really attempted to tune at higher rpms, but when I do rev it up and hold it, the AFRs bounce around just like at idle.

The injectors are mounted in the stock location on the GSXR ITBs, putting them about 5" from the head. I've noticed when I take the manifold off, the gasket is permeated with fuel around the intake ports. Like there's a half inch ring where the gasket is wet. I've never really worked with carbs or a wet manifold, so maybe this is normal. In general it seems like there is a substantial amount of fuel in the intake tract and I've wondering if it's just pooling and not being atomized. But it seems like everybody with DCOE style ITBs would have the same issue.

For the time being I'm just running Speed-Density while trying to fix the idle. Switching over to MAF or Alpha-N doesn't help.

Thus far I've:

• Balanced the ITBs with a carb syncrometer to 4 Kg/hour
• Used a caliper to set the plates to open within 0.25mm.
• Fixed a few small vacuum leaks I found using ether or by listening with a hose.
• Replaced plugs several times
• Readjusted the valves
• Fuel pressure looks good.
• Several compression tests, open and closed throttles, all good.
• Swapped in a set of GSXR 240cc (high impedance) injectors.

The only thing I haven't swapped out or replaced would be the coil packs. As best as I can tell they're working fine though.

 

I'm about ready to throw in the towel and pick up some Jenvy DCOE throttle bodies. The only thing stopping me is the fear that they will do the same thing.

 

I'm just wondering if anyone has any ideas.

[Tony D]

I can't tell too much from your photo as the fuel bar covers everything I need to see. A shot from straight down overhead would be helpful. You don't REALLY mean "RRFPR" do you? You just have a standard 1:1 Manifold Vacuum-Referenced unit correct? With MS you DO NOT use a rising-rate FPR, it totally defeats the purpose of having programmable injection! Anyway, here goes:

 

Balanced the ITBs with a carb syncrometer to 4 Kg/hour

Used a caliper to set the plates to open within 0.25mm.

 

The above two are basically mutually exclusive. You set your balance how? Bypass screws per barrel? CLOSE the throttles it idle and adjust your balance with the individual bypasses. Check off-idle balance separate from idle.

 

 

Fixed a few small vacuum leaks I found using ether or by listening with a hose.

 

Use water or some sort of spray. No leaks are permissible. Any present will screw with you forever.

 

Replaced plugs several times

 

Why? Are they gapped properly?

 

Readjusted the valves

 

What is idle vacuum? Alpha-N should give you a ROCK STEADY idle as it discounts MAP Spikes in the ITB setup. This may be your issue, look to see if your idle kPa values are sufficient to keep you in ONE LOAD CEL. If it's jumping between four cels at idle and they all call for different fuel you will constantly be chasing it. Make all cels at idle that it jumps through identical and see if your AFR settles out.

 

Fuel pressure looks good.

 

Nix

 

Several compression tests, open and closed throttles, all good.

 

I doubt that from the above synch and setup concerns. But likely they are not an issue if you standardise their setup.i. Terms of throttle angle and relative position / reaching WOT.

 

Swapped in a set of GSXR 240cc

 

The 450's may have had latency issues, they were also high impedance?

Edited June 22, 2014 by Tony D

[Tony D]

It appears the plates have small holes in them. THAT is your idle air. Close the throttles and affect balance/synch at idle throughout what I (think) is your bypass line (black tubes?)

 

Off idle synch is a mechanical issue and needs relatively rigid mounting. Probably not an issue in this problem though.

 

That it does it at idle AND off idle...but who knows where in the rpm range. Give specifics... Hold it at 1500, then 2500, then 3500 and look on your tuning map....is the dot bouncing around the cels? If so equalise their values and see if AFR Stabilises and then determine why the MAP is jumping around so much. It should be relatively stead off-idle, and NOT A FACTOR AT ALL in Alpha-N.

 

Indeed, something is being left out here as switching ITB's to Alpha N should stop all MAP-Based AFR jumping about. And off-idle for sure!

[Sam280Z]

Post your msq and a data log of the behavior.

[Metro]

Top View

 

You don't REALLY mean "RRFPR" do you?

My mistake, yes, I meant a normal 1:1 FPR.

 

The above two are basically mutually exclusive. You set your balance how? Bypass screws per barrel? CLOSE the throttles it idle and adjust your balance with the individual bypasses. Check off-idle balance separate from idle.

The plates are completely closed and the idle is being set by bypass screws on the bottom of each TB. The bypass screw is just a hole in front of the TB -> screw -> exits behind the plate. The black tubing is for my vacuum log.

 

Use water or some sort of spray. No leaks are permissible. Any present will screw with you forever.

I discovered a while ago that spraying any liquid on a vacuum leak with individual runners will make a loud puffing noise. I've also been using some tubing as a stethoscope to probe around. I haven't been able to find any other vacuum leaks since my other problems. Unfortunately, the way they're attached allows for lots of potential problems.

 

Replaced plugs several times.

Long time ago I used to foul plugs like crazy, so it's a cheap test. I have probably 30 lightly used plugs in a box, waiting to be cleaned.

 

What is idle vacuum? .. If it's jumping between four cels at idle and they all call for different fuel you will constantly be chasing it. Make all cels at idle that it jumps through identical and see if your AFR settles out.

The idle is stable enough and I've setup the load cells so that's it only ever sits in one cell. So the fuel being delivered is fairly consistent (around 1.18ms), but I still get a wildly fluctuating AFR. I mentioned alpha-n because a long time ago I tried using that to keep it in one bin. In the log, it says my range was 41kpa to 48kpa and the VE was mostly limited to moving from 23 to 24. As you can see in the logs, my spark advance does move around a bit, but even with a fixed advance it still runs rough.

 

The 450's may have had latency issues, they were also high impedance?

The 450s are low-impedance and the 240cc are high impedance, they're about as different as you get, which leads me to believe I can rule out injectors. Granted, I don't know the exact opening and closing time of the GSXR injectors. The 450s are configured with settings I had been using previously with MS1.

 

I've attached my log and current settings. I should point out the MAT is from the Ford MAF and I had it sitting on top of the radiator, then I moved it off to the side, that's why it drops like 20 degrees in the log.

2014-06-21_18.35.28.zip

Edited June 23, 2014 by Metro

[madkaw]

Cool set up! I could see where you could be chasing vacuum leaks. I am currently chasing an inconsistant idle with my triples and MS3. I too have the dreaded vacuum log. Have you capped it off to eliminate it. Might want to try the propane torch with a hose to pinpoint possible leaks. 

How is your MAP signal with your vacuum log? I've read that some have a hard time getting a steady signal from ITB's and that MS has a program for that.

[Sam280Z]

Just had time for a quick look.

Your AFR target table needs some work.

 

Try turning off "include AFR" and setting your timing to "fixed". See what that does.

[Tony D]

Yeah, I'd say your idle speed will move when your timing changes. In fact, that is how the old TEC2 would keep idle speed steady: varying advance.

 

If you are moving around, and your pulsewidth is consistent, there is a variable. As stated fix the timing and see if it settles  down. This may or may not affect the AFR's but it will eliminate a variable. 

 

Basically lock everything down and re-enable ONE variable at a time to see what effect they have. That is kind of the way you're supposed to tune it... almost nobody does with so many pretuned maps 'close' these days. But when something doesn't work, go back to the basics, lock it all down and if it's still moving it's a mechanical leak someplace, possibly variable. If you are sucking PCV into the manifold, make sure the hose doesn't have a drip leg in it where condensed oil could puddle and make for an intermittent plug.

 

The idle kPa is pretty good for an ITB.

[Sam280Z]

Your PW is sub 2ms. That is in the non-linear range of the injectors. You have no precise control over how much fuel is being injected. Try changing "number of squirts" to 1. (I'm running sequential and can't remember if this doubles the pulse width - I think it does.) You may want to turn on "sequenced batch fire" so you can have consistent injector timing every time you start.

 

Just to be clear, your target AFR table has a different load range than the VE table. You should make them the same and adjust the values accordingly.

 

I only suggest turning off "include AFR" for testing purposes. Your current settings bounce between 12.5 and 13.8 target AFR. This is included in the fueling calculations when "include AFR" is on. (Once you figure out the fluctuation problem, turn it back on.)

 

I think this combined with the small pulse widths (due to large injectors) causing you to lose control of fueling is causing the fluctuations. 

 

To get the PW down, you can try any combination of the following: bump the idle up, pull some timing at idle (try 10 deg), switch to 1 squirt per cycle, bump up your AFR target.

 

With large injectors far back in the manifold it will be more difficult to tune idle, cruise, and transient conditions. See my MS3 build thread: http://forums.hybridz.org/topic/102280-ms3-w-ms3x-install-l6-turbo-sequential-injection-boost-control/

[Tony D]

The 250's should have idled like a stock car... 450's not necessarily so.

 

I found mine "Alternating 2squirts" (with 240's) was the smoothest on my setup. I think I didn't have sufficient rail volume to keep steady pressure in the fuel system so pressure dropped enough that it affected injection events with single pulse.

 

JeffP put a large pulsation damper on his rail to keep the 450's, 550's and eventually 720's from drawing-down rail header pressure and making his AFR move all over on the Extreme280ZXT. It really helped in that regard--with and ITB, what kind of vacuum log signal do you get to the FPR? It could suck if you were in an opposite-phase dynamic where due to the strong pulse in the intake tract the FPR was at it's lowest pressure just when the injectors fired off....and then out of phase at highest pressure point in the rail on a subsequent injection event!

 

I mean, we're reaching here now... but all within the realm of possibility. A good gauge would show if this was the case. I don't think JeffP sees more and a 2-3psi fluctuation at idle now with the 720's in there with the damper (using a 6" Diameter Mirrored 0-60 psi Reference Gauge tapped mid-rail.) Before, it was swinging like 15+ psi!

 

BIG injectors really get easier with sequential as it really does limit the rail fluctuations.

Edited June 24, 2014 by Tony D

[Metro]

@tony

MS2 and MS3 have a cold idle advance table and I was thinking of using it in lieu of an IAC. The problem I have is, the vacuum port on the TBs are only 1/8th of an inch. Playing around with it, I  can only move the idle by about 150rpm at the most. Of course, it might all change once it's idling properly.

 

@sam

Turning off include afr targets was a good catch, that should not have been enabled right now. Turning that off, I had to up my VE values slightly, but ultimately it's still exhibiting the same problem.  I honestly haven't paid much attention to the AFR table as I thought it should have been disabled along with EGO. The load range is from my old MS1 configuration and using a 60mm TB on an N47 manifold. I was hoping to use MAFload for the axis at some point.

 

I think my injectors run fine down around 1ms, if I wasn't clear before I've been running MS1 since 2006 with the same injectors. My old config used 2 squirts and the VE value was down somewhere near 25 in the table. Unfortunately I'm having difficulties with finding an old log, but I'm confident I didn't have problems running these injectors around 1ms. The only thing I was unclear on when going from MS1 to MS2 was the Injector PWM period(us), so I left it at the default "50".

 

I should also point out the GSXR injectors are half the size, so the idle PW was around 3ms, same problems. I also just went out and tried running it 1 squirt and that didn't help any.

I have been following your thread too Sam, that's why I was a little concerned with the placement on mine.

 

@madkaw

Early on when I was having problems keeping it in one bin, I tried pretty much everything MS recommended. First I tried using a fuel filter to dampen the pulse, no change. Then I inserted a restrictor from the GSXR map line, nada. Then I tried all together and still didn't see any change. So in conclusion with my weak cam, I don't see any need for MAP tricks to help smooth it out. If not for the RPM going all over the place, it would only fluctuate 1-3kpa or about the same as it did when running a single TB. I haven't tried removing it from the equation yet, it hadn't really crossed my mind.

 

 

I really want to think there's vacuum leaks or the manifold is leaking exhaust gas into the intake ports, but I just don't see any of the signs of it doing any of these things as I had in the past. As far as I can tell, it's fine and yet it still idles terribly.

 

I went out a attempted to datalog some higher rpms and I definitely have a noise issue that I'll need to sort out. I used to ground MS to the intake manifold, but I can't with this manifold. Instead, I ran a heavy lead out from the cylinder head to a spot on the fender and bolted all my grounds together on there. Obviously that isn't working very well or MS2 is a lot pickier than MS1.

2014-06-23_18.28.24.zip

Edited June 24, 2014 by Metro

[Metro]

I found mine "Alternating 2squirts" (with 240's) was the smoothest on my setup. I think I didn't have sufficient rail volume to keep steady pressure in the fuel system so pressure dropped enough that it affected injection events with single pulse.

 

JeffP put a large pulsation damper on his rail to keep the 450's, 550's and eventually 720's from drawing-down rail header pressure and making his AFR move all over on the Extreme280ZXT. It really helped in that regard--with and ITB, what kind of vacuum log signal do you get to the FPR? It could suck if you were in an opposite-phase dynamic where due to the strong pulse in the intake tract the FPR was at it's lowest pressure just when the injectors fired off....and then out of phase at highest pressure point in the rail on a subsequent injection event!

 

I mean, we're reaching here now... but all within the realm of possibility. A good gauge would show if this was the case. I don't think JeffP sees more and a 2-3psi fluctuation at idle now with the 720's in there with the damper (using a 6" Diameter Mirrored 0-60 psi Reference Gauge tapped mid-rail.) Before, it was swinging like 15+ psi!

 

BIG injectors really get easier with sequential as it really does limit the rail fluctuations.

 

I used to experience what you mentioned when I had the FPR mounted by the fuel filter on the passenger fender. This put it about 4-5 feet away from the rail. I had this same VDO gauge mounted on the fuel rail and would bounce around 10-15psi as the injectors fired. I think I was able to tune around it, it wasn't much of an issue. Now I've moved the FPR to the end of the rail and the gauge barely moves at all. The FPR is getting the same vacuum signal as MS - it's attached to the end of vacuum log. I've also removed the vacuum line from the FPR numerous times and it hasn't had any noticeable effect.

 

I might be able to get my hands on some quality fluid filled gauges to test with, we'll see.

 

For whatever reason, I've never been able to get it to idle smoothly with alternating, I've always gone with 2 squirts and batch fired. For many other reasons, I'm wishing I splurged and got an MS3 instead of the MS2 chip. It would be great to go full sequential without having to add a bunch of drivers. Maybe next fall when I have time to rewire everything.

 

Also I just wanted to toss out, thanks for all the ideas guys. This is pretty much my hail mary attempt at finding a solution. I've been screwing around with this for about two months now.

[Tony D]

My single vacuum bypass at 1/4" would give me from zero to almost 2200 rpms in terms of idle bypass. I should think 1/8" per barrel would be more than enough to balance and give adequate basic idle speed.

 

That you had it running on MS1 for so long and did not have the issues I would tend to agree there is something in there altering it with everything locked down like you say.

 

From that first log photo, it looks like everything is steady but your 200 rpms swing which inversely correlates to the MAP signal... Your TPS is obviously closed.

 

I don't know how receptive you are to some bushleague troubleshooting, but get some ignition sealant...or clear krylon. While the engine is up to temperature, start at the head and spray the hell out of the connection hose joints, the manifold flange all your vacuum hoses paying attention to the joints. Lay it on THICK. 

 

We would do this  to Lycoming Engines when the inlet seals would start going out, drip in high temp paint that was semi-thickened from being open, and let it suck into any crevice and crack on the engine's intake system (they had similar hose-clamped hoses on the "spider's legs" from the intake) and as it was sucked in it dried and plugged the hole.

 

It's clear, so it shouldn't have a big impact on visuals... It's a quick and dirty way of confirming and temporarily plugging the vacuum leaks. I would have a logger running and when you spray log.... if you see the MAP and RPM start to even out or stabilise you're on the right track.

 

ALL THAT BEING SAID....

 

Traditionally noise issues have plagued these installs. I would not be surprised to find what appears to be minimal noise or voltage fluctuations to be at the  root of all of this! I'd filter and ground the hell out of everything really well just to be sure!

Edited June 24, 2014 by Tony D

[Sam280Z]

In my experience, vacuum leaks just raise the engine speed when the ecu can compensate for the mixture. I had a pinhole leak in the weld when I modified my manifold. It just hissed. You could actually hear it over the engine running. Not saying you don't have one, but a can of carb cleaner should be able to tell you pretty quickly if you do. Tony's spray paint idea will work too.

 

Your deadtime is .8ms. That is over half of the PW. Sub 2ms PW is problematic. MS3 has a whole section on tuning this area.

I don't have any experience with low impedance injectors, so I can't offer anything on the PWM settings.

 

Can you get it to idle at say 1200 - 2000 RPM? Does it do the same thing?

 

Try dropping your spark advance to 10 deg or even less. You will have to open the idle up letting in more air. This will require more fuel and will drop manifold vacuum. Mine runs at more like 65kPa at idle.

 

The same behavior with the small injectors could have been from the AFR table and spark advance affecting your fueling. Need a log to tell.

[Tony D]

I have seen flexible connectors like that vibrate and let in air intermittently. That could account for the MAP changes and easily account for a 200rpm bump.

 

But again, so can electrical spikes and noise giving the ECU Fits. 

 

It could be either. 

 

I'd caution around tuning around something like this...if the MAP was constant and not swinging I'd say try tuning. But until that MAP stabilizes there is either some sort of intermittent leak or noise is causing an issue.

 

It shouldn't vary but a few kPa, and under 25rpm at idle with everything locked down.

[Sam280Z]

I have seen flexible connectors like that vibrate and let in air intermittently. That could account for the MAP changes and easily account for a 200rpm bump.

...

 

I'd caution around tuning around something like this...if the MAP was constant and not swinging I'd say try tuning. But until that MAP stabilizes there is either some sort of intermittent leak or noise is causing an issue.

 

It shouldn't vary but a few kPa, and under 25rpm at idle with everything locked down.

I agree. I'm not suggesting that he tune around it. My suggestions are to see if he can eliminate the behavior s so we can see what is causing it.

 

How flexible are the connectors? Have you tried grabbing the throttle bodies and giving them a good shake?

 

Could it be possible that the vibration is causing the throttles to open and close slightly?

[Metro]

@tony

While I like your idea, damn that seems like a mess to clean up after. But if I get desperate enough..

 

 

@sam

I've done the carb cleaner thing and that's how I ultimately discovered my manifold wasn't flat. The engine would rev up and even out when I sprayed around runner 1 & 2. I haven't been able to find any other leaks since I fixed the others.

 

In the second data log I tried revving it up and holding it around 1500rpm and it still seemed a bit problematic. I'm not sure I have enough idle adjustment to get them much higher than 900rpm, I would have to crack the throttle.

 

The ITBs are attached fairly securely. I can barely move when I've tried wiggling them, but I haven't tried it with the engine running. I suppose I could try it and do it again with the pushrods disconnected. It seems like there could be some problems there.

[Metro]

The jiggle test didn't have any noticeable effect.

 

 

Now here's something interesting. I found the data log of the very first time I fired it up using MS1 and my old config.

No wonder I was so optimistic back then. There was no special setup, no balancing, or vacuum log, I just bolted the manifold on and fired it up. The manifold definitely wasn't sealing though. It was night and I could see it spit fireballs where the manifold met the head. The only unusual thing in the log would be the alternator. I don't remember why, but it clearly wasn't charging.

 

After this test, I rewired the coil outputs to use the SPR instead of the IAC outputs and did the boost control mod.

 

So this gives me some new things to look at.

• try to raise the idle up to get the PW around 2ms
• disconnect the alternator - could be the source of my noise problem
• check the wiring
• check the board for bad solder joints

2013-12-21_20.48.29 fixed timing.zip

[Sam280Z]

To get your PW up, you can drop the fuel pressure a bit and recalculate your reqfuel as well. There are conversion calculators on the web. I did this on mine when I had only the large injectors. If that helps and you need to keep it that way, I recommend getting actual flow values and dead times via test mode testing.

[madkaw]

Still following this hoping to shed light on my issue since they seem closely related-though I have carbs. I think I have eliminated induction issues. Recently my plugs have shown something a miss(pun intended) , so now I am troubleshooting my MS.

I have been on the MS forums and there are other issue similiar. Broke, but not broke enough to pinpoint-hate it.

Good luck.