Tuning for STi intake manifold swappers

[fahr_side]

Okay, a bit of a niche group since not too many of us do this, but for the DIYers out there...

 

Issue

 

You've swapped the STi or oldskool WRX metal intake manifold onto your 4th gen, along with some other bits, like maybe the STi TMIC. You may have done a TGV delete at the same time, or converted to top-feed injectors. Response is funny, or you have fueling errors your perfect MAF scaling doesn't seem to allow.

Now, I've tuned dozens of cars with larger than stock injectors, gutted TGVs etc. in the past but never before had I run into so much trouble with tip-in fueling as I did on my own car having swapped in the metal manifold. There are also a few other things you might want to pay attention to if you're trying to get the best driveability or fuel economy out of this setup.

 

Things to note

 

The plastic intake manifold is very nice and smooth and all, doesn't conduct heat well blah blah, but air distribution through the four runners is not as even as on some of the metal manifolds. Nabisco member ASF Machine was kind enough to test various intake manifolds on a flow bench and came up with the numbers that appear below. Why do we care about this? Well, unless you have a WBO2 sensor installed on each exhaust runner to check they are in balance there just isn't any way to know if each cylinder is running at the same AFR at all times. When you only read AFR downstream from the turbo, what you're looking at is just the average of all four cylinders.

 

Plastic 4th gen LGT / 3rd gen WRX manifold:

 

Cyl #1 293cfm

Cyl #2 282cfm

Cyl #3 283cfm

Cyl #4 291cfm

Max variation 10cfm (3.5%)

 

2007 STi manifold:

 

Cyl #1 303cfm

Cyl #2 304cfm

Cyl #3 306cfm

Cyl #4 299cfm

Max variation 10cfm (2.3%)

 

Note that while the difference between the highest and lowest flowing runners is about the same, the cylinders that are getting the most or least air are not. Note that Subaru is well aware of this and that is compensated for in the per-injector compensation tables A through D. The name is a little misleading in this case perhaps, but I believe they are tweaking IPW slightly to compensate for the different airflow through each runner to hit the same AFR on all four cylinders.

 

It goes without saying that if you're installing aftermarket injectors that come with a datasheet you should install the highest flowing injector into the highest flowing manifold runner and the lowest into the lowest.

 

The other difference you will find is in transient, or Tau, fueling. If you switch to a metal manifold it does somehow seem to affect how fuel evaporates off the port walls when fuel is being sprayed at a closed intake valve. I have no idea why this would be but experimenting seems to show a difference.

 

Note also that different fuel rails have different resonance profiles and the chances are you are using the fuel rails that match the intake manifold you're installing. I believe Subaru compensate for this mostly via the Engine Load compensation tables.

 

Last off note that the tip-in fueling strategies on the STi models are quite different from other models. Perhaps they care less about economy on that platform.

 

Solutions

 

Basically what you want to do is to tweak the tables in your ROM as close as you can to values in the ROM from the car the manifold came off of, to make the best starting point for further fine-tuning. I recommend the following:

 

Copy and rescale the per-injector pulse width comp tables

 

Open the ROM for the model year car your manifold came from and check out the Per-Injector Pulse Width Compensation table A and compare that to the table in your own original ROM.

 

Here's a comparo between my own GH8 tables and those from an '11 GRB from whence came my manifold:

 

http://i464.photobucket.com/albums/rr5/fahr_side/per_inj_comp_zpsgohice4s.png

 

So the first thing you may notice is that the STi table is much flatter than the GH8 one. This is I believe due to the more even air distribution through the four runners at various engine speeds and air flow conditions.

The second thing you might notice is that the tables aren't the same size. If you didn't notice then take a look at the tables below, one from an '05 LGT and the other again from the '11 GRB.

 

http://i464.photobucket.com/albums/rr5/fahr_side/per_inj_comp2_zps2bnqscy9.png

 

What you will notice if you add all four tables together and average all the values is that they sum to very nearly zero. This shows that the tables compensate for the variation in air flowing through each runner at each point in the RPM and IPW range covered but do not alter the total.

 

In order to emulate the donor tables we'll have to rescale them and for that I recommend Airboy's Log File Interpolation spreadsheet that you should all be familiar with. Just remember that the axis values on the rescaled tables must increase monotonically, or IOW have equal gaps in values between them. In the case of the A2WC521C ROM I'd probably only rescale the values up to 6k4rpm since very few of you will be revving the LGT motor beyond there. You could also in this case only grab and rescale the STi values from 2.00ms IPW and up.

 

See the example here:

 

http://i464.photobucket.com/albums/rr5/fahr_side/per_inj_comp3_zpsgvgnjtoo.png

 

Now would be a really good time to compensate for a change in injector scalar you might need because (say) you installed larger injectors, or went from side-feeds to top-feeds. Select all the IPW values on the table and multiply by the difference in scalar like so: new scalar / old scalar = change and in this example that's 565.25 / 749 = 0.70, as below:

 

http://i464.photobucket.com/albums/rr5/fahr_side/per_inj_comp4_zpsaijvmhkj.png

 

Now go ahead and do this same process on tables B, C and D.

 

Next..

 

Copy and rescale the Engine Load Compensation tables

 

Below are the '06 WRX and '05 LGT tables. Why not the '11 STi you ask? Well, no way in hell I'd put the '08~ STi fuel rails on a poor unsuspecting LGT, given all the horror stories about resonance in these rails and the resulting lean spots and stumbling!

 

http://i464.photobucket.com/albums/rr5/fahr_side/elcomp1_zpsvujlpque.png

 

Now let's move the column headers over and selected rows of values, like so:

 

http://i464.photobucket.com/albums/rr5/fahr_side/elcomp2_zpsw2i2lbjb.png

 

One thing to note here is that later cars will have two EL Comp tables labeled Cruise and Non-Cruise which related to TGVs being closed and open. While the '05~'07 models also have TGVs fitted, they often only have one table, because basically the TGVs are open as soon you hit the gas pedal the first time and don't close again until you restart the car from cold. In that case I might copy just the Non-Cruise table from the later ROM or possibly, if the same manifold were fitted to an earlier car, use that single table version as the model. It should be fairly obvious that if you've gutted the TGVs, you should use the Non-Cruise table values and if you have a Cruise table, use the same values in that one, as there's now no difference in VE whether the ECU thinks the butterflies are open or closed.

 

Either way, to obtain the very best driveability and economy you will want to fine-tune the EL Comp tables yourself. I've made another tutorial on that very subject here.

 

Tau fueling

 

This is something that is not defined for every ROM out there. Again the differences between models can quite obvious. Here again is the '11 GRB vs. the '08 GH8:

 

http://i464.photobucket.com/albums/rr5/fahr_side/tau_zps9hmhgiyn.png

 

In this case we can just copy the axis and table values straight over.

 

Tip-in fueling

 

Again, this is where STi and WRX strategies were quite different so you might want to experiment with both approaches. In general the STi approach is a bit lighter on additional fuel at smaller tip-in throttle values and the same or very close at higher tip-in throttle values. On the '08~ models however this seems to be different and the STi gets less enrichment at low throttle delta and much more at high delta.

 

What you want to watch out for is whether or not your ROM has a table for Tip-in Enrichment Compensation (RPM) as this skews all the tables to some extent as the logic is different. If you have one, you probably want to base your new tables on a ROM that also has such a table, like the '08~ STi models. So far I have not seen a ROM earlier than 2008 that had this compensation. Here are three stock tip-in tables with their corresponding boost error comp tables below. From left to right they are '07 STi, '06 WRX and '05 LGT:

 

http://i464.photobucket.com/albums/rr5/fahr_side/tip-in_zpsse5vnxjw.png

 

My approach would be the try the leanest strategy first and go to a richer strategy if you experience bogging. Of course if you are installing different injectors you also want to apply the same scalar correction factor mentioned earlier to Throttle Tip-in Enrichment tables A, B and C as you do to all other ms-based tables. Tuning tip-in is one of the trickiest aspects to getting to a good tune and I'm not about to get too deep into that here, partly because my own understanding of it isn't that great.

 

Hopefully this makes the tuning part of the conversion a bit less painful than swapping the parts over was.

[Th3Franz]

Good writeup. I do not believe the X-tau tables are defined for any of the ROMs. These come into play more for tip-in variations for IAT, etc. The way the fuel film evaporates between the plastic and metal manifolds seems to be fairly different. Toss a big FMIC into the mix and it get tricky for sure. Also take a look at some of the JDM cars if you're able to get the definitions.

For anyone who wants to read more about the theory, here's a good reference:

https://books.google.com/books?id=vmq7NiuaA78C&pg=PA87#v=onepage&q&f=false

[LatentWagen]

Franz, that is some good reading. Makes sense, thinking about it. Just hypothesizing here:

 

The bumpier surface of the metal intake manifold would hold a more consistent film versus the smooth plastic surface, which might be "splotchy" and more defined by surface tension of the film, as opposed to adhesion to a rougher (metal) surface.

 

With the metal intake, the enleanment on tip-in ought to be more predictable, and so the film evaporation more easily accounted/compensated for.

 

Of course it's not as simple in real life, and very interesting to see the tuning in practice, thanks for that Fahr!

[utc_pyro]

Franz, that is some good reading. Makes sense, thinking about it. Just hypothesizing here:

 

The bumpier surface of the metal intake manifold would hold a more consistent film versus the smooth plastic surface, which might be "splotchy" and more defined by surface tension of the film, as opposed to adhesion to a rougher (metal) surface.

 

With the metal intake, the enleanment on tip-in ought to be more predictable, and so the film evaporation more easily accounted/compensated for.

 

Shouldn't the film in this case be in the head and maybe the very bottom of the TGV assembly? Injectors are pointed at the back of the valves, so unless there is a pretty serious flow inversion there shouldn't be any notable amount of fuel in the plastic part of the manifold.

 

That whole book that Franz linked to is extremely informative.

[fahr_side]

Shouldn't the film in this case be in the head and maybe the very bottom of the TGV assembly? Injectors are pointed at the back of the valves, so unless there is a pretty serious flow inversion there shouldn't be any notable amount of fuel in the plastic part of the manifold.

 

 

Exactly what I was thinking.

[LatentWagen]

So even at part/low throttle, under vacuum? Wonder how far that mist & film actually makes it's way up? A clear plexi IM might be a fun project.

[fahr_side]

So even at part/low throttle, under vacuum? Wonder how far that mist & film actually makes it's way up? A clear plexi IM might be a fun project.

 

 

 

That's a good question and one I can't answer. I can tell you that injectors are sprayed at closed valves in certain situations, otherwise known a batch-fire. We know base pressure is 43psi above MAP. How far up the runner gets wet in the split second between intake cycles?

[rhino6303]

Why change the entire injector compensation table when using larger injectors?

 

From my research it is sufficient to multiply the header column values by the old/new scalar. In other words, multiply row 0 by the scalar and nothing inside the table itself.

 

I'm not disagreeing by any means. I'm just curious. Thanks.

[fahr_side]

Why change the entire injector compensation table when using larger injectors?

 

From my research it is sufficient to multiply the header column values by the old/new scalar. In other words, multiply row 0 by the scalar and nothing inside the table itself.

 

I'm not disagreeing by any means. I'm just curious. Thanks.

 

 

 

Answer is in the thread title. You're correct in that if you're just changing injectors it's quite enough to rescale the IPW axis. I'm talking about what to do if you swap the intake manifold. The text mentions rescaling the axis because a lot of people will be installing larger squirters at the same time as a manifold swap.

[rhino6303]

Answer is in the thread title. You're correct in that if you're just changing injectors it's quite enough to rescale the IPW axis. I'm talking about what to do if you swap the intake manifold. The text mentions rescaling the axis because a lot of people will be installing larger squirters at the same time as a manifold swap.

Gotcha. Thanks.