Intending to do fuel only TBI on 4 cyl aircraft engine

I don't recall if I ever posted an intro message but my account was not deleted so maybe so. Maybe I joined before the intro message requirement?

Can someone point me to a comprehensive thread on adapting a GM system to a non GM carb situation. I think I have a pretty good idea of what needs to be done but would be good to see it in writing. The areas I am least knowledgeable about are the bin files and tuning and also the initial setup.

Will likely use an Iron Duke system but those may be marginally small - I think my cfm requirement will be around 200 and I'd rather error on the large side. Does anyone know the cfm rating on the vanilla Iron Duke TBI? Would be nice to land a 1 bbl from the 1984 Fiero pace car or, more likely, the Holley knockoff. Engine is 235 c.i. and will never (hopefully) spin up to more than 3,000 rpm. Red line is 2,800 (prop tips should not be allowed to go supersonic) and most of its work will be done 2,000 to 2,200. In terms of engine competency (it's stock, freshly overhauled and in compliance all the manufacturer bulletins), I would compare it to a mild build.

Regarding ECM setup, I'll need to shut off the spark advance feature, the VSS input and, no doubt, other parts of the system. The exhaust is simply 2 cyl on each side feeding into their respect open exhaust pipe (no muffler). So I will need to noodle the O2 thing a little and figure out what works best. Perhaps installing on one side and assuming the other side is going to look pretty much the same will work.

Thanks in advance to all.
Larry

:welcome:

I don't have any experience with an aircraft engine. I'm a touch nervous even posting in the tread asking information for an aircraft engine.

I don't have a link to share for a TBI installation. The DeBiase EFI link (http://www.slantsix.org/articles/dib...conversion.htm) is an awesome overview of a GM EFI Conversion. Most of the information on the DeBiase link will apply to a TBI conversion. I would suggest using the under hood 1227727 (aka '7727) ECM for a TBI 4 cylinder application. I don't remember which years GM used the '7727 with TBI 4 cylinder vehicles, 1989 ~ 1992 most likely. The 1227730 is an under dash ECM that electrically identical to the 1227727.

To operate fuel only, I'm thinking only the MAP Sensor, TPS, and a Distributor Reference Pulse DRP are all that would be required. An MSD 6A box can supply the ECM with the needed DRP.

dave w

Are you sure TBI is appropriate? Wouldn't a system with port fuel, with better altitude correction for fuel pressure, be a good choice? Why not control spark as well? Wouldn't you want to keep the engine out of detonation?

Anyway, there were several threads years ago on diy-efi and gmecm by an individual (Steve Parkman?) who was putting a gmecm on Subaru engines for aircraft use and selling them. He went by "swagaero" for Swag Aeromotive. He also did a fair amount of work with other engines in aircraft and was published in Contact! magazine several times.

http://lists.diy-efi.org/pipermail/d...98-January.txt
http://www.retro.co.za/efi/DIY_EFI/1997/efi2-293.txt
https://groups.google.com/forum/#!to...lt/ggkMhwrCduI
http://www.contactmagazine.com/backissu.html

Quote:

---

NTSB Identification: LAX99FA052 . General Aviation
Accident occurred Friday, December 18, 1998 in TUCSON, AZ
Aircraft: Parkman VARI-EZE, registration: N81EZ
Injuries: 1 Fatal.

This was the maiden flight in the aircraft, and during construction,
the pilot made numerous design modifications to the airplane which had
changed the flight and performance characteristics. Instead of the
airplane's engines recommended by the designer, the pilot had
installed a GM Geo Metro automobile engine, which the
pilot altered by the addition of a Mitsubishi turbocharger. The pilot
also altered the fuel computer chip to adjust the fuel flow,
attempting to achieve a 2-gallon-per-hour consumption rate. The pilot
told associates that the engine produced 78 hp at 4,800 rpm. Engine
technical data showed the actual power output of the unmodified engine
to be 79 hp at 6,000 rpm. Severe detonation was found in the engine
during postaccident examination and it is believed that it most likely
would produce only 55 percent of rated power. Following takeoff,
witnesses saw the aircraft in a nose high attitude and it never
achieved more than 100 feet agl. Some witnesses said the nose attitude
was as high as 15 to 20 degrees just before the crash. The pilot
radioed that he was having a problem and said he had to put
the aircraft down. The airplane collided with a tree during the
attempted forced landing and a postcrash fire consumed the wreckage.

The National Transportation Safety Board determines the probable
cause(s) of this accident as follows:

The pilot's failure to attain and maintain an adequate flying airspeed
during the aircraft's maiden flight, which resulted in a stall/mush
condition. Also causal was the pilot/builder's decision to modify the
engine and the fuel system control microchip, which resulted in
detonation and a severe reduction in power output, and led
directly to the pilot's inability to attain and maintain airspeed.

---

Scary report 2MANY.

Unlike the unfortunate pilot-builder, my plan would be to evolutionary as opposed to revolutionary. Don't know all the details of the build, but the addition of a turbocharger, which would likely call for a lower compression ratio than what existed, would have been a major error. Further, there are "club" propellers that can be constructed that will act as a crude dyno (small power envelope based on rpms at, or near, redline) and, had it been used in this case, it would have 1) told the builder that the engine didn't have the requisite power; or 2) cremated the engine in a stationary safe spot on the ground.

I do know of a Continental O-200 (real aircraft engine) that has been converted to MPFI and DIS - it's a SDS system. Seems to run fine - I wish I had that installed on my (very similar) engine. But that's much more work than converting to a simple TBI system - moding the intake runner tubes to take the injectors; adding a signal wheel for ignition; etc.

Do we believe that a properly installed and tweaked TBI system would be superior to a carb system? Probably, but no guarantees. My problem is that my recently purchased aircraft has a carb that (Marvel Schleber MA-3), inexplicably, does not have mixture control (for altitude variations). So I'll have to change it out at likely a cost of $800 - $1,000. If we think that TBI is superior to carb and can do the proper altitude adjustments, then why no do that? Not only does it solve the mixture control problem (and carb heat problem and fuel primer problem, neither of which we have discussed), but there might be a 4% or 5% improvement in power (torque) and maybe even an improvement in fuel consumption.

Serendipity: removed the carb today and the throttle bore is 1.6875" which is very close, if you believe info on the WWW, to the size of the vanilla Iron Duke single bbl TBI. Again less work.

So, for now, I'll continue with the game plan. But I love the input. All ideas and cautions are very, very welcome.
Larry

This is one request I wouldn't even make a DIY type suggestion on, knowing what requirements there are for aircraft. I work at an upholstery shop and when we re-build an interior, we have to buy certified for aircraft use leather and other materials.

Even though I love DIY and re-purposing parts that were never meant to be together, I strongly suggest finding certified conversion electronics for your project.

Six_Shooter, I appreciate your thoughts on "certified" aircraft parts. It's an environment that I live with every day. That said, I pretty much agree with the guys over at DIY Auto Tune (MegaSquirt) who generally take the approach that an engine is somewhat like a rose - an engine is an engine is an engine. Through annual inspections over the years, I've almost come the the thought that certified aircraft components/parts are less durable than ordinary automotive parts. Lots of stuff gets DQed and has to be changed out. Are the specs too tight or is it just crap? Probably equal quantities of both.

Okay, looking at the fuel only TBI installation, we have an ECM, one or two injectors, TPS, O2 sensor, fuel pump, and one or two other sensors. One of the autos that I drive is a 96 Jeep Cherokee that has 225K on it (still runs great) and MPFI instead of TBI. The only part that has failed in 225K miles is the O2 sensor and that was intermittent until changed out. If believe the average speed of a ground vehicle is in the 45 mph range, that's 5,000 hours. I think that's a terrific record, certified parts or no!

The most dangerous part/possibility of/in flying is losing power on takeoff. That's why pilots, in preparation for flight, do WOT run ups checking power (static rpm - prop acts as a dyno) and as many systems as possible. Losing power on takeoff is really bad news - losing power in just about any other situation is manageable.

I thank you for your concern but, based on my personal experiences, the "certified" label is somewhat down the priority list. But keep your thoughts coming- everyone's input is extremely valuable.
Larry

my thought is, are these ECMs going to be alright where atmospheric pressure is ~20kpa or whatever?

im no pilot, is that even a reasonable threshold? 10+km in the air does push you into that range, right?

personally, i'd be scared just to run all this ECM code at high altitudes where atmos. map goes to weird places a car would never go

wouldn't you lose a lot of VE resolution?

hell, this is probably a place the ecm has never been tested, would that low of an atmospheric pressure be clamped in the ECM, cause an error, and hit a default value that'll throw your AFR out to lunch?

might it just crash?...

you have to make damn sure your ecm is updating barometric pressure constantly -- an ecm with an external pressure sensor for atmospheric would be a great idea (not sure any GM ecms do that, but i know there are some out there? i've seen it in some foreign cars..)

a guy like robert or someone else familiar with disassembly could do a bit of an audit on the code you're using to make sure it wont hit some weird threshold like that, you know, dropping to whatever KPA doesnt throw a CEL and go into limp mode assuming you're at sea level.

i'd even throw someone like robert a bit of cash to do that auditing so it's less likely something goes sideways on your test runs if you're really going to use a gm ecm for it.

just throwing it out there, flying is one thing i haven't done.

this is definitely something i would attempt.

but if you do it, can you pleasepleaseplease build one of my raspberry pi dataloggers and build it into your airplane instrument panel, just for the hell of it?

The 1990 Chevrolet Lumina 2.5 Liter was offered with TBI engine using the 1227727, 16197128, 16198260 ECM. According to RockAuto.com the 2.5 liter Memcal part # is 16152936. The BCC is APTW, AUAJ, and AUSS. RockAuto.com is showing they have a one 2.5 liter TBI Memcal #16152936 available ($62.79).

I've been thinking, GM would have considered some vehicles would be driven in the Rocky Mountain Rage of the Western United States, with some roads reaching elevations of 12,000 feet above sea level. Seems plausible to me a 1 Bar MAP sensor could be safe for an aircraft flying up to an elevation of 12,000 feet above sea level. I'm thinking GM has / had a testing facility in Arizona that reached elevations to about 7000 feet above sea level.

dave w

most gm map based code does a key on baro check before startup and then anytime the throttle opens far enough past a set value of throttle position.now the throttle position this is allowed at would have to be at a big enough throttle opening to have no vacuam in the manifold or you could massively offset kpa readings to the ecu.if you were to climb without enough throttle you wouldnt get a reset of baro as you climbed.theres a lot of things to consider

Quote:

---

Scary report 2MANY.

Unlike the unfortunate pilot-builder, my plan would be to evolutionary as opposed to revolutionary. Don't know all the details of the build, but the addition of a turbocharger, which would likely call for a lower compression ratio than what existed, would have been a major error. Further, there are "club" propellers that can be constructed that will act as a crude dyno (small power envelope based on rpms at, or near, redline) and, had it been used in this case, it would have 1) told the builder that the engine didn't have the requisite power; or 2) cremated the engine in a stationary safe spot on the ground.

---

Steve was a decent guy and really appeared to know what he was doing. I seem to remember discussions about dyno testing as he had a specific rpm and power level he was targeting. He'd developed his own detonation sensor and knock control which I believe may have been part of the issue. The articles in Contact! include one of his gm efi conversions on a Lycoming 0-235 engine.

I'm sure you understand that anything discussed on this forum is not to be used without a complete evaluation for its appropriateness for a particular purpose, and that despite what may appear at times to be very in depth technical knowledge, we are all hobbyists and should not be considered experts in any way.

Quote:

---

Serendipity: removed the carb today and the throttle bore is 1.6875" which is very close, if you believe info on the WWW, to the size of the vanilla Iron Duke single bbl TBI. Again less work.

---

That is close, and the TBI may present less restriction to flow than the carburetor due to no venturis. I'm somewhat concerned about what happens if a restart attempt is made in the air. The carburetor will introduce fuel based on air velocity through the carb. The TBI introduces fuel based on computer signals generated because the crank is turning. If the engine is at a substantial angle will this cause fuel to flow where shouldn't? It's possible to delay the onset of fuel delivery after cranking is initiated so the chance for a problem could be reduced, but I'd still consider this strongly.

Would you be using an oxygen sensor to monitor fuel delivery? The oxygen sensor really limits the amount of change to air/fuel ratio to 14.7:1. It's possible to use a wide band sensor which provides a much greater AFR range but using it for fuel control will require custom code.

Quote:

---

most gm map based code does a key on baro check before startup and then anytime the throttle opens far enough past a set value of throttle position.now the throttle position this is allowed at would have to be at a big enough throttle opening to have no vacuam in the manifold or you could massively offset kpa readings to the ecu.if you were to climb without enough throttle you wouldnt get a reset of baro as you climbed.theres a lot of things to consider

---

Exactly. Automobiles are expected to make relatively slow changes in altitude so barometric updates are not frequent or automatic. Any increase in altitude is expected to be tied to a throttle change, and a substantial change in altitude should require substantial throttle opening. So baro isn't updated unless throttle angle rises above a set value, generally 80% or greater. Is this going to be a substantial concern? Probably not if fuel usage isn't a concern. But if obtaining altitude adjustment to mixture control is part of the goal for the conversion then it's a major setback. Two options to defeat this are to read baro updates from a dedicated sensor and switching to a mass airflow sensor which doesn't require barometric pressure changes but both options will require custom code.

The automobile ecm itself is not weathertight although I would not expect altitude by itself to cause any troubles. They tend to be fairly reliable although I'd want to use a truly weatherproof unit for added security. The MEFI controllers, built by GM for marine use, are lightweight and completely sealed and have a long history of reliability. Aircraft engine operation is probably similar to marine operation. The engine will operate at a relatively fixed rpm for fairly long periods of time, and rapid changes in engine rpm are few. Automotive ecm code handles these conditions while staying within emissions requirements which really adds a degree of complication that may not be needed in the air. The MEFI code is much simpler.

More thoughts to follow...

Steve, Delcowiz, Dave and 2Many - hey guys, this is really great stuff - just the kind of input I need.

1. Yes the plan is to definitely use an O2 sensor, most likely a wide band.
2. 99.9% of operation would be at 12,000 altitude or less and I agree that GM must have this altitude range built in - else how could one drive over Loveland Pass in CO. But what happens on the rare test/adventure at 15,000 feet? Good questions.
3. Yes, the engine would operate with less throttle changes than an automobile. If one is practicing landings (doing touch & goes), then there will be a lot of changes to throttle similar to an auto but other than that, it would be pretty stable in the throttle settings. For cross country cruise (the opposite extreme of practicing landings), I could see the same setting for maybe an hour.
4. Weight is number one in an aircraft. If the marine ECMs are somehow lighter than the auto equivalent, then that's the way to go. And if weather tight, all the better - but not required unless the auto versions are prone to corrosion.
5. One issue may be the TBI would be inverted (the carb is an updraft). I've read that, with the normal auto application, holding the pedal to the floor will "shut off" the fuel. Question: on an inverted application, would giving it full throttle prior to cranking and then backing off as it cranks be an effective procedure for starting?
6. The baro pressure thoughts are very illuminating and very important. Lots of considerations: BTW, the engine is a Lycoming O-235 and it is designed for some lead in the fuel for valve lubrication. The original fuel (87 octane AV fuel with lead) is no longer available and now one has to use the 100 octane which has more lead. One can mix that with pure auto gasoline (no alcohol) to lower the lead count back toward the 87 octane but the punchline is: need to keep the AFR close to 14.7:1 to keep from fouling the plugs. So I'm all ears on a dedicated sensor or any effective way to keep the ECM ready to do a good job of leaning the AFR to what is needed at altitude. With a dedicated sensor or whatever changes that are made to overcome the problems, can the programming be done without too much problem? Could the ECM just be reprogrammed to not look for throttle changes and simply change injector metering rates based on feedback from the O2 sensor?

Thanks again for the input. I'll try to locate the article in Contact.
Larry

The typical automotive ecm might occupy a space of 6" by 8" by 3". The first and second generation marine computers are about 3" X 5" X 1". Third and fourth gen are smaller. 5th gen seems to be a bit larger.
http://www.aldlcable.com/graphics/mefi1-4.png

http://store.custombuiltmotors.com/e...tbNrQ60_57.JPG

Leaded fuel poisons oxygen sensors. How quickly will depend on the amount of lead that is in the exhaust. A pyrometer could be used as a feedback device though. The ecm could target a manifold temp vs oxygen level, and with some diligence a person could engineer a variable input to represent manifold temp.

TBI in updraft? That will be interesting. Well, it's possible to alter the number of revolutions of the crank before fuel delivery is started. "Clear flood" mode is intended to clear the intake manifold if the system is overfueled. It is actuated by pressing the throttle to the floor before or during cranking. I would program the system to wait longer before delivering fuel rather than requiring input from the operator.

GM does not always design systems for all environments the vehicle can be used in however, the MAP sensor range is from 2.9" Hg to 14.7 psi. If you're attempting to fly where atmospheric pressure equals 2.9" vacuum using a GM EFI powered aircraft, well, I'm not sure what advice I can offer for success. ;)

Okay, probably a marine ECM (or perhaps two of them for redundancy) if not too expensive. Not likely to find a lot of them at Pick-n-Pull. I may start with a salvage yard application and then convert to marine when this thing gets ready to fly.

I can go with 91-93 octane auto gas and use a special lubricant additive. Fuel will be cheaper and will save on plug maintenance. Engine has an 8.5:1 c.r.

Yes, I also would opt for the delayed fuel delivery - an extra 2 or 3seconds of cranking?

I have the back issue of Contact on the way. Copyright laws will probably prevent me from posting the article but I will be able to share some of the data. I'm anxious to see how S. Parkman overcame the altitude compensation hurdles. Here is what I read in another thread on SWAG Aero's EFI:
a Ford Escort throttle body, a Saturn ignition module and an ECU from a GM Geo.

MORE INFO: I obviously don't have any kind of coolant temperature information. But I do have a gauge that shows cylinder head temperature (selectable by cylinder) and another that shows oil temperature. The latter is probably going to be the closest to what GM ECMs are going to be looking for in a coolant temp. Only downside to the oil is it normally takes oil some time of running hard to get up to temp. Cylinder head temps come right up to the 325* + range. Maybe Cyl head is a better signal for TBI.

Anybody want to speculate on the easiest GM ECM (for TBI) to program and whether or not it has the dormant "highway cruise mode" capability - I think that is where I should start.

Thanks,
Larry

cylinder head temp should be just fine. will probably just require more calibration. most air cooled fuel injected bikes i've played with are cyl. head temp...

I don't believe Steve tried to overcome the altitude correction issues. His effort happened in the early days of EFI hacking and I don't believe the community was aware of such things at that time.

Highway cruise might be useful but it provides for no operator input to adjust mixture and does not rely on the O2 sensor. You might be better off using an old trick... bias the O2 sensor signal using a potentiometer. The ecm would increase or decrease delivered AFR based on sensor bias.

It seems to me the criteria for ecm selection might be a little more in depth than just TBI and lean cruise. Like home computers, GM's engine control computers became more sophisticated with each generation. The last production TBI ecm for light duty vehicles is the 7427. It provides the fastest data updates when using a laptop and is used by many people. It's primarily used on V8 applications so adapting it to a four cylinder might take some thinking and some patience. It does not have lean cruise although there are ways to force the engine to run lean. The 1227730 mentioned by Dave is also a good choice. It was available with a program designed for a four cylinder engine although it may not have lean cruise code. The program for the four cylinder engine could be migrated to a weathertight ecm, the 1227727. But keep in mind that most people do not use this code so you may be on your own for help if you select it. Older computers such as the 1227747 and 1227165 that were available with lean cruise have significantly slower update rates and are less flexible if the computer program needs to be modified for a specific use. In stock form with commonly available tools they are not my first choice although there are modified versions available from Bob Rauscher of Dynamic EFI that can be made to do almost anything one needs.