Tuning Motronic-based ITB racecar motor

[preston_brown]

Everyone,

My latest acquisition several months ago is a independent throttle body-based 3.9L motor wrapped up in a 964/993 hybrid ex-European GT body. The car was imported in the early 2000s and run in PCA racing for awhile. Then it was parked and had not run in 6+ years when I bought it.

After changing plugs, rebuilding injectors, changing oil, fixing a high pressure fuel leak(!), and a few other things, we got the car started. The car is still running a Motronic 993 DME. It (obviously) has a custom chip in it from RS Tuning in Germany from wayyyy back in 1997. I've tried to get in touch with them to see if they can provide me with any information, but have heard nothing. I'm not sure what kind of throttle bodies I have, but would love suggestions/ideas? They appear to share a common shaft. The intake trumpets are carbon fiber and look totally custom. The air cleaners are the standard ITG ones originally designed for carbs. I think.

The problem is, the car is running EXTREMELY lean at idle, and really anything less than about 1.8k RPMs. Also, it has a seriously tough time transitioning out of low range RPMs with lots of stumbling and backfiring, I assume because it is too lean. And then it is a mess when you let off the accelerator as well as the revs drop.

I have a wideband O2 sensor hooked up to monitor what's going on. From what I can tell, the car is set up to use the throttle position sensor (TPS) only. No MAF, no AFM, no IAT. Knock sensors and O2 sensor are in place. So there can't be any speed-density type mapping going on here. Must be Alpha-N alone, from what I understand.

I'm trying to determine the right way to chase down the problem. I don't see any vacuum leaks. There are two rubber hoses leading from small ports on the number 2 and 3 throttle bodies that are tee'd together and then are feeding into the fuel pressure regulator. The throttle bodies all have air bypass screws, I assume for balancing between them just like on carbs. Most of them are screwed out a fair way, and when we checked with a synchrometer, balance was quite good (when pedal is held at 2k RPMs so we don't stall).

I have to assume I have either a problem with fuel being under-delivered, air being over-delivered for what the DME is believing should be correct, or the map in the DME chip is seriously bad. I think it must be one of the first two. Can I turn the air bypass screws all the way in and see what happens? Counting the number of turns each? It *must* have a richening effect, right? I also plan to check fuel rail pressure -- I have a fuel pressure gauge on the dash though and it is registering 4 bar. Which I think is OK.

My injectors are Bosch part 0280 150 786.

I have ordered a Ostrich 2.0, PROM burner, etc. in case I have to work on the chip fuel and timing maps or more. I wonder how the chip is set up. I know that some people think Motronic tuning is some super secret black art but I'm a software engineer by trade and hexadecimal and assembly don't scare me. Learning curve will be there but if we have to make adjustments, it shouldn't be the end of the world.

Any ideas on where to start? Here are some pictures to give you a better idea.

Hmm, what do we have here?



The engine cover open:


Left side throttle bodies:


Right side throttle bodies


993 DME chip:

[Chuck.H]

Sounds like a neat project! If you go down the assumption that 'it ran well once' then I wouldn't dive into changing the chip right away, but if you can grab the binary image off that EPROM, some of here can help I'm sure.

Since you see it trying to run the fuel/spark off the TPS only, what do the voltages from the TPS look like? Is it adjustable?

I'm guessing if this company modified the chip from a standard 993, they may have bypass resistors in place to hold those inputs at a constant state - do you see any connections like that to the standard 993 DME harness?

Have you checked the distributor drive belt as I understand these are a common 993 failure point.

Good luck, keep us posted!
Chuck.H
'89 TurboLookTarga, 336k miles

[preston_brown]

The TPS has been monitored with a Durametric Porsche OBD1 interface and it sees between 0 and about 70% as I depress the throttle(ignition on but motor not running). Does the TPS also have micro switches for idle and WOT? I don't see anywhere in the Durametric software to see if those are present and/or getting tripped.

The DMEharness appears stock and unmodified. Likewise, the board itself didn't look like it had been modified when I opened it to look at the chip. Could be wrong of course.

I looked at the distributor belt and it appears fine. The distributors are properly synced and both work.

[Quicksilver]

Wow!
Not an easy road sitting in front of you...

Originally Posted by preston_brown

. . .
I have to assume I have either a problem with fuel being under-delivered, air being over-delivered for what the DME is believing should be correct, or the map in the DME chip is seriously bad.
. . .

They are all the same thing. Mixture is the only thing. Once you have that the rest of the stuff is just throttle pedal.


It occurs to me that it might be insanely easier to switch to a Megasquirt or MOTEC. That gives you a whole development environment where you can quickly plug in adjustments and see results so you could conceivably get close to a good tune in a day's worth of work.

To do the same with the DME you will need some sort of mod where you can jack into the chip socket with an interface cable to a laptop that allows you to change your maps on the fly.

Either way, you need a recording wideband O2 setup with data collection that will let you collect the mixture and the RPM as you run up through its range. You also need a way to record knock sensor activation so you can adjust the timing curves. This will need to be done at multiple throttle positions from idle to full throttle.

Frankly this would probably be a lot easier with a MAF sensor in the system. The system was designed with that additional input because it makes it easier to approximate the fuel delivery and spark timing. Assuming the linkages make all of the butterflies move accurately in unison there is no reason you couldn't adapt a single hot wire sensor on one cylinder. All the cylinders are going to be doing the same thing so it would provide valid data.


Thinking about this I keep going back to the most basic concept: What can I get out of this that I can't get out of something that is more sorted out, easier, and most likely cheaper?

[preston_brown]

I have a ROM emulator so I can tune the motronic in real time.

I have a innovate LM-2 wideband O2 data logger with an inductive clamp to simultaneously log RPM.

Yes, we could go to MS or MoTeC (pricey!!!) or VEMS or Haltech or whatever. And we might. But this car raced with Motonic once so it must be able to work. This isn't an untested system. Therefore getting it running right again as it has in the past is my first goal.

Where do you put a MAF on an ITB setup?

[scarceller]

Originally Posted by preston_brown

Where do you put a MAF on an ITB setup?

You could add a MAF like the 89 BMW M3 does with ITBs. You'd need to add a air intake hood over all 6 ITBs and then have the MAF on the inlet pipe to this chamber.

But, if this is a race car I would just do a TPS AlphaN setup, which is most likely what was done. Does the current setup have any of the following components:
- Throttle Position Switch
- MAP sensor
- Intake Air Temp sensor
- Barometric sensor
- Cyl Head Temp sensor
- anything else?

Verifying what sensors you currently have will possibly help identify what fueling strategy was employed.

I would not change the system if you have the ability to tune it.

[preston_brown]

We've got a TPS. I don't see any MAP sensor. I don't see an IAT sensor either. There is a CHT sensor. I don't see any pressure sensors for ambient air/Altitude.

I'm not putting a goofy plenum over these carbon fiber works of art!

[scarceller]

If you are running lean it could be the TPS sensor not working correctly or out of adjustment?
Assuming the chip must be correct because it raced at some point.
And no air leaks
And fuel pressure is correct
Then in a AlphaN setup the heart of the air metering is based on a properly adjusted and functioning TPS sensor.

But what has me more worried is you mention NOT having a IAT sensor? You can't do AlphaN properly without it, you have to adjust fuel based on Air Temp.

You should also have a Baro sensor to adjust for altitude changes, although you can live without this one assuming the car does not see huge changes in altitude with the same tune.

[preston_brown]

Agree an IAT makes sense for alpha-n. I will look harder but there isn't one mounted on or near the stacks that I see, and I don't see one randomly placed elsewhere in the engine bay. The TPS could be out of whack I haven't disconnected it and bench tested it yet. Maybe Friday, I'm out of town at the moment. Just gathering ideas.

Agree a pressure sensor could help but I stay near sea level.

Does motronic have the ability to monitor both an IAT and a CHT? I didn't think our 3.2 or even 3.6 cars had an IAT sensor.

[scarceller]

Originally Posted by preston_brown

We've got a TPS. I don't see any MAP sensor. I don't see an IAT sensor either. There is a CHT sensor. I don't see any pressure sensors for ambient air/Altitude.

I'm not putting a goofy plenum over these carbon fiber works of art!

Given that you have only a TPS sensor it must be a AlphaN system, can't really be anything else. But with no IAT sensor I'm not sure how they pulled it off? Here's an example of IAT fuel correction:

Here's a typical IAT correction table for fuel:

TempC %fuelCorrection
-32c = +9.38%extra fuel
0c = +6.25%
20c = 0.00%
60c = -5.47%

The above table assumes a base fueling at 20c and as IATs increase we start removing fuel. If we assume the car will run in temps from 0c through 60c we can have fueling changes (AFRs) of as much as 12%.

[scarceller]

Originally Posted by preston_brown

Does motronic have the ability to monitor both an IAT and a CHT? I didn't think our 3.2 or even 3.6 cars had an IAT sensor.

Both the 3.2 and 3.6 motronic system have IAT sensors, they are built into the AirFlowMeter. The example IAT compensation table I gave above is from the stock 89 motronic DME. Only a MAF setup can live without a IAT sensor and even in my custom 3.2L MAF setup I still employ a IAT sensor to compensate ignition and fuel if IATs get out of hand, it helps control engine thermal runaway.

[scarceller]

The baro sensor is not required. The 3.2L does not exactly have a baro sensor, it has a Baro switch that closes at 1000meters, when closed the DME removes 6% fuel. You can easily live without the Baro

[preston_brown]

I will see (and hope) that I have missed the IAT sensor somewhere.

[scarceller]

As for your lean mixture I suspect the TPS switch may have been moved. Keep in mind that at low throttle angles you get the most air flow changes. For example: if throttle is 5% open and you crack it open to 10% you will see a much larger delta change in Air Flow (% wise change) than if you where at 80% and went to 85%. The calibration of the TPS switch really matters most at low throttle angles. This is why if the TPS switch is slightly moved from it's calibration point it will be most noticeable at light loads and idle.

Try this setup procedure: Once engine is fully warm idle it at say 2000RPMs and then set fuel mixture to lambda=0.95 (AFR=14.0) to hit the target rotate the TPS switch till you richen the mixture to the target. Once this is done rev it to 3000 then 4000RPMs does the AFR stay put at the target 14.0?

[TimT]

I'm not putting a goofy plenum over these carbon fiber works of art!

Unless you put carbon fiber art over the other carbon fiber art

The plenum from our LeMans liveried GT3RS

[preston_brown]

Originally Posted by scarceller

Try this setup procedure: Once engine is fully warm idle it at say 2000RPMs and then set fuel mixture to lambda=0.95 (AFR=14.0) to hit the target rotate the TPS switch till you richen the mixture to the target. Once this is done rev it to 3000 then 4000RPMs does the AFR stay put at the target 14.0?

that totally makes sense. I will give that a try on Friday.

What do you think about the air bypass/correction screws? Do you think I should screw them all the way in and then back them out a turn or two, and then rebalance later? or leave them totally alone?

Thank you for the suggestions. As an aside, here is a picture of the car's sister at LeMans in 1995. My car also has a LeMans tech inspection sticker but didn't compete in the race.

[scarceller]

Originally Posted by preston_brown

What do you think about the air bypass/correction screws? Do you think I should screw them all the way in and then back them out a turn or two, and then rebalance later? or leave them totally alone?

I suspect these screws simply bypass air around the the throttle plate to help setup idle. My suggestion is don't mess with them yet, my bet is these where setup with a special air meter device that slide into the trumpet. Then at idle they simply set the bypass screws so each cyl ingests the same exact air flow. These bypass screws most likely only help you achieve idle. I suggest maybe someone else much more experience with ITBs can share some thoughts here.

My suggestion is don't worry about getting idle perfect yet, concentrate on a higher idle speed like 2000 or 3000RPMs to set mixture in the ballpark of AFR=13.8 - 14.0

[LJ851]

scarceller has good advice, i agree with all of it.


As someone that has extensive experience with ITBs, i will offer what i would do to sync / etc.

1) close all the air bleeds.

2) At idle, find the the cylinder with the highest vacuum in each bank (Left and Right) and sync them at idle.

3) Take those same 2 cylinders and check the sync at 3000 rpm. If they vary, i like to split the difference. For example, L bank slightly stronger at idle slowly changing to right bank slightly stronger at 3000. This step is less important on a race car, but does improve part thottle running.

4) Take one bank and using the air bleeds sync the lower 2 barrels to the strong one you chose in step two. Do the same to the other bank.

5) The throttles are now sync'd and you get to choose how to control idle speed and mixture. Having a wideband 02 will help here. The system will idle with X amount of airflow from either the throttle plates being opened or the air bleeds being opened or a little of both. What you are going to use to your advantage is the fact that the ECU recognizes throttle movement (and the adds fuel) and the air bleeds are purely mechanical and will lean the mixture.

The best result for your system will be achieved with TPS adjustment for the best AFRs at full and part throttle and correcting the idle mixture as much as possible by your choice of idle speed control.

[TimT]

You mention the car had not been run for six years... was the car running well before it was put in storage?

Regarding the air bleed screws, they are very easy to set.... and once set they will very rarely if ever need to be adjusted again...

The bypass screws are the least of your worries.

Most of them are screwed out a fair way, and when we checked with a synchrometer, balance was quite good (when pedal is held at 2k RPMs so we don't stall)

On carbs your still working on the idle circuits and transitioning to the mains at 2k With ITBs with no venturi I think as soon as you crack the throttles open the bypass screws job is over...

Does the car have any sort or OBD type connector? perhaps you could get the Bosch software and monitor what is going on live?

[preston_brown]

Originally Posted by TimT

You mention the car had not been run for six years... was the car running well before it was put in storage?

...

Does the car have any sort or OBD type connector? perhaps you could get the Bosch software and monitor what is going on live?

The seller said the car was running ok when stored but that idle was becoming more difficult to maintain. But trust me the info I have received from him has been spotty if mostly accurate.

The car has a round port and is OBD1. Sorta. I have a Durametic cable and software and can monitor many of the sensors in realtime.