ITB mode am I doing it right?

[lukep6470]

Hi,

I have converted my setup from alpha-N to ITB mode and I am VERY happy with the improvement in driveability. I highly recommend doing this to anyone running ITBs and haven't bit the bullet yet.

The only problem I'm having is getting my head around the effect of TP load at switchpoint. My tune runs a little lean at around 2000rpm and 17% TP. I've let autotune completely map my VE table up to 4500rpm so far.

I have attached my tune, a short datalog, and a spreadsheet of my switchpoints which I renamed to .txt as I couldn't upload a .xlsx

The car is a pretty stock Ford 289 automatic with a Microsquirt running msextra 3.2.5. The microsquirt goes into a eeciv adapter cable and into a Ford hot rod harness running TFI ignition. I've built a set of 46mm ITBs using Mikuni throttlebodies.

[muythaibxr]

The load at switchpoint curve just sets the amount of the table that is SD vs AN. Setting it lower gives more bins to AN and setting it higher gives more to SD. I recommend starting at 50% then tuning it for tunability.

[kaeman]

is that 50 % throttle position or 50kpa, my engine idles at 60 to 67kpa, and my cruising kpa is in the 70's, but my throttle position is only open 2 to 8% for cruising depending on the speed.

[muythaibxr]

If you are taliking about the curve I was talking about, it is neither. It sets the load that will be used in table lookups at the point where you switch from AN to SD.

[wmax351]

There are two curves: one is the switchpoint curve, and the other is the load at switchpoint.

The switchpoint curve is the more important. This is the TPS position for 90% baro, which is found out by pulling the MSL into some form of spreadsheet or statistical software. You plot TPS (y axis) vs RPM (x axis) for all datapoints with a MAP from 89-91 % baro, or just 89.75-90.25 if you have a big datalog. Then you run a linear or other regression to determine the formula for the switchpoints. I think mine is a quadratic or cubic regression that fit best. You then use that formula to calculate the switchpoints in %TPS for a range of rpms.

I looked at your datalog, it's too short, and doesn't cover the full rpm. You need a longer log. Ideally from your alpha-n setup, so it is running well.

[lukep6470]

Hi,

Sorry if I wasn't clear. That attached datalog is just a short example of how my current tune is behaving around town. I have 100s of MB of logs that I pulled my reference points out of for the TPS switchpoint graph. I went for around 87 KPA as my particular engine seemed to loose resolution over 90 and I decided to be conservative.

Looking at how consistently low in the TPS range my setup goes atmospheric I'm guessing my ITBs are oversized for this engine.

[wmax351]

They might be a little oversized, but not outrageous. The MAP value also depends on how you have the MAP sensor hooked up. If it is connected to all of the ITB's, it will have a much higher MAP for a given TP, though it is probably a more accurate representation. If it is connected to only one ITB, you will have more resolution (In the case of my bike, enough to run SD only if needed), but a less accurate picture of the whole engine.

Here's the MSQ for my bike, which is stock ITB's, though it would originally use a Vane type Airflow meter for fueling. At low RPMs, there is a very small throttle opening for the switchpoint.

Looking at your datalog more, how well synchronized are your throttle bodies? There seems to be AFR oscillation: the average AFR tracks the target value quite well, but bounces to the sides of it far more than would be explained by EGO correction. It jumps from 13-16 rapidly at lower rpm's, possibly indicating one part of the engine is getting the wrong mix. The time between spikes is correct for RPM.

[lukep6470]

The map is reading from a common plenum under the manifold so it is reading the entire motor. The O2 sensor is only reading one bank and I have used a carby balancer to get the idle flow as consistent as possible but it doesn't seem to make that much difference to the occilation. I do have an exhaust leak on 1 cylinder on that bank that I guess is upsetting readings at low rpm.

Another problem I'm having is spark plugs are getting a LOT of carbon build up since converting from the 4bbl holley to ITB's. You can see from my datalog that the engine isn't running that rich at all and I am getting better mileage than the carby that used to run clean. Do ITB's normally require a hotter spark plug?

[wmax351]

The common plenum will affect the balance. The throttle linkage needs to be balanced too, so all plates open identically. With the carb balancer, press down on the pedal and see if one moves first.

What's your air filter situation? If it isn't working as well, that could possibly foul your plugs.

[lukep6470]

I have a foam rally style filter on each ITB.

I have never been completely happy with the geometry of my linkages. I will have to organise a couple of days down time to sort them out properly.

[wmax351]

I have a foam rally style filter on each ITB.

I have never been completely happy with the geometry of my linkages. I will have to organise a couple of days down time to sort them out properly.

The foam filter, if the oil type, could be putting some oil into the cylinder and fouling the plugs.

The linkages seem like a real pain with ITBs, especially as they get bigger in number or have multiple banks.

[lukep6470]

The filters are dry. The carbon is dry and the engine isn't using any oil. There is just a hell of a lot of black soot. For some reason I think it really needs a hotter plug. I've currently got NGK AP5FS plugs in it and I have a set of Autolite 45s I'll give a go.

I've got the front four and the rear four linked properly using weber linkage wheels and tie rods. It is the link from the rear bank to the front bank that the geometry is off. I don't think this will be causing my carboning problem as all 8 are black.

My VE table is a bit weird as it has very high numbers in the top left. I tried a few hard take offs this morning with autotune on and it made that corner even richer!!

I think I have a runner length resonance thing going on as the car is a real stump puller from a standing start now. This is a stock 289 and it launches harder than my previous 5.0HO with gt40p heads, explorer intake and E cam!!

[lukep6470]

I just thought of something else. I'm running a 80 degree C thermostat. Could this be causing the engine to run too cool and not burn off carbon?

[dontz125]

Are you getting misfires? A too-rich condition causing misfires will give a lean EGO reading, as unconsumed oxygen goes past the sensor head.
From my 2-stroke days, carbon fouling is too-cold plugs, carbon dusting is too rich.

[lukep6470]

Nope no misfires. Currently the car runs like OEM.

[lukep6470]

Hi,

So yesterday I went through and sorted out some linkage issues and balanced all the throttles so at idle they are all flowing between 1.8-2.0 Kg-H on the flow meter. Idle and off idle is markedly smoother and more responsive. Idle and part throttle AFR is still jumping around by the same amount so I'm starting to think this is a 289 firing order quirk.

I will drive for a couple of weeks and check my spark plugs again.

[lukep6470]

Over the weekend I did a free air calibration of the LC-1. This is the 2nd time I've done it. The first was when the sensor was brand new. I then ran auto tune and went for a 20min drive.

It didn't make too many changes over the table except for steady cruise which it leaned out by 20% or more. This is with no other change except for the calibration.

We'll see if that has sorted my carbon issue out.

[phongshader]

I've just gotten my car on the road and I'm setting it up to run in ITB mode. This thread has been great at explaining what the different load/switch point curves do. I do have a question about setting the load switch point and the load at switch point. I've combined several logs to create an excel sheet and filtered out all entries that aren't between 89.5-90.5 kpa. I started to plot the point out in the itb load switch points graph and got a very non linear curve...am I going about this correctly, should I smooth the curve, should I not combine different logs? The log only goes up to 4500rpm...breaking in a new engine.

itbLoadSwitchpoint.JPG

The engine is a ford ranger duratec 2.3 with stock cams, headers, suzuki gsxr itbs, ms3/3x 1.3.0, synchromap.
Thanks,
Justin