FI engines- Mixture

[Dan Thomas]

My 07 jeep JK works great with multi port FI. and fly by wire computer.

And it'll have electronic fuel injection. I was talking about the constant-flow mechanical fuel injection we have in airplanes now; FADEC can't do much with the injectors in such a case. All it could do is control the total fuel flow, not individual injectors. An electronic injector system can't have a mechanical override like SMA's diesel; the injectors need electricity and control signals to work.

Dan

 

[Ted]

Why do airplane engines have mixture control? Why not have a simple processor monitor and constantly adjust mixture considering that it can do a better job than a human.

The reality is that the processing power required to perform this operation is not particularly high. However, in order to program a system that will actually operate a piston reciprocating engine as well as an intelligent human with a mixture knob is extremely difficult to do to provide the balance of fuel economy, rich running for cooling purposes, avoiding detonation on higher powered engines, etc.

FAA certification is certainly a big part of it, but the harder obstacle is producing a system that will actually be something people will want to buy. Nobody will want to purchase a system that, when installed, has worse power, worse fuel consumption, and lower reliability than the previous system.

Why don't cars have mixture levers?

This is commonly used as the justification for why airplanes should have computers on them, but building a computer control system that will control a car effectively and give people what they want is significantly easier than doing it for an airplane. You would be surprised how difficult it is to replicate the mixture control movements of an advanced pilot, or even of a good but not-advanced one.

As to Dan's question: the Aztec has a Precision RSA-5 on both engines, and the Continental has whatever comes on the IO-520-Es. But if they were trying to get a million flight hours on something, I could help.

 

[peter-h]

The obvious issue with the Eagle EMS system is that it stops working when the electrical supply goes.

A battery only postpones the inevitable.

If they were offering a little vac pump mounting compatible alternator, that would be very different.

FADEC alone can be approved. Diamond/Thielert did it OK with the FAA. They unfortunately had some problems with the engines They also had reliability issues with the FADEC system...

 

[elmetal]

NFW...

Sometimes I wonder where you people come up with this stuff...

ever have a car towed up or down signifant altitude changes?

I have. 3000 feet, it took a good 15 minutes for the car to be back to normal, Fuel Injected.

if you're taking 15 minutes to climb to 3000 then that's ok, but most of us aren't.


To answer your question Where I come up with this stuff? Personal experience.

 

[Everskyward]

ever have a car towed up or down signifant altitude changes?

I have. 3000 feet, it took a good 15 minutes for the car to be back to normal, Fuel Injected.

if you're taking 15 minutes to climb to 3000 then that's ok, but most of us aren't.


To answer your question Where I come up with this stuff? Personal experience.

I managed to drain the battery of a car by leaving something on for about a week while it was parked. After it was jumped it ran pretty badly for a few minutes. I was told that was because it had lost it's memory and had to relearn where it was, which was at about 5500' MSL.

 

[Dan Thomas]

ever have a car towed up or down signifant altitude changes?

I have. 3000 feet, it took a good 15 minutes for the car to be back to normal, Fuel Injected.

if you're taking 15 minutes to climb to 3000 then that's ok, but most of us aren't.


To answer your question Where I come up with this stuff? Personal experience.

Lycoming's IE2 engine doesn't have such issues. We can't afford to take 15 minutes to get to 3000 feet, either. The automobile's system is designed to handle automobile-type scenarios, not aircraft operations, and it's far cheaper and more reliable for them to do that. There's no reason, other than money and the fact that there's no need for it, why GM or any other auto manufacturer couldn't design and build a system that responded almost instantly to environmental changes. Shoot, it's only a processor that reads temperatures and pressures and EGO and RPM and stuff like that to get the dope it needs to run injectors and ignition timing, and it doesn't take more than a few milliseconds to do it. After all, it has to respond instantly to throttle changes and make mixture and ignition changes to keep the thing performing as demanded, right?

Your car's system is probably cross-checking for a while to make sure it's getting the truth and not some bogus info from faulty sensors.

Dan

 

[Sac Arrow]

First time I've heard that FADEC can't make rapid adjustments. I guess the FAA-certified FADEC on the FAA-certified SMA diesel we have here in the FAA-certified Cessna182 can't do it either...

Honestly. The real reason we don't see electronic injection in light aircraft is due to the propensity of electrical systems to fail. Without a supply they die. It's the reason we use magnetos, too. Just ask any pilot of any experience just how often aircraft electrical systems make trouble. 90% of all problems tend to be electrical in nature, and that applies to cars, too. Unless any manufacturer of an aircraft electronic ignition or fuel control can prove to the FAA that loss of aircraft electrical supply won't kill the engine, they won't get certification. The SMA's FADEC has a mechanical override, whereby the fuel control is mechanically shifted by a separate lever from the computer's control directly to the power lever; that lever normally just works a potentiometer. Works fine with diesel but it ain't gonna work with electronic gasoline fuel injection.

Dan

Except that within the last ten years, EFI, electronic coil-over-plug spark control systems and their associated ECU's have become so reliable that they typically just don't fail. I have put about three quarters of a million miles on vehicles (cars, trucks and motorcycles) so equipped in the last 12 years and I have not had a SINGLE failure of the EFI or spark systems. Some mechanical transmission issues and emissions control issues yes. In the last 500 hours of flight, I've had at least one new set of mags and a handful of unscheduled timing adjustments.

Magnetos fail because they are mechanical, high voltage devices which fail through wear and consequently require periodic adjustments. The difference is that they don't tend to fail entirely at once so you can often nurse a sick engine back to a field. COP ignition systems are solid state, low voltage devices (up to the individual stick coils on the plugs) and can have an unlimited ability to be timing controlled withough the requirement of a complex and finicky mechanical advancement system. They also allow for the use of an automated spark retardation system based on knock sensing, something we don't have and can't easily incorporate in an aircraft engine with magneto ignitions. If we did have it in aircraft engines, they could be safely run at a lot higher performance envelopes than they currently are. And unlike distributor and magneto ignition systems, they aren't temperature or altitude sensitive.

People spend a lot of money installing GAMI fuel injectors to even the fuel distibution to individual cylinders. This issue completely disappears with EFI systems, and EFI systems can much more accurately meter fuel than mechanical systems, improving performance even if we were to retain a manually operated mixture control feature. But why would you want one - I could see maybe putting a manual override in an aircraft engine. EFI systems in most cars compensate for altitude just fine.

All these systems take millions of dollars in R&D to develop. Problem is that's cheap when you can sell millions of these cloned systems in automobiles world wide, but not so cheap when your market is maybe a few thousand at most domestically, irrespective of the FAA certification process. I think the real answer is that if we want these systems for our aircraft (I do) Lycoming and TCM aren't going to develop them. I can see Ford or GM or some other major automotive manufacturer developing a modern series of aircraft engines under a subsidized government contract.

 

[Dan Thomas]

I think the real answer is that if we want these systems for our aircraft (I do) Lycoming and TCM aren't going to develop them. I can see Ford or GM or some other major automotive manufacturer developing a modern series of aircraft engines under a subsidized government contract.

Except that then the poor taxpayer is seen to be subsidizing the manufacture of high-tech toys for the rich. I don't think any politician would touch it. Never mind that the demise of 100LL and the possibility of no replacement would require such technology to enable the thousands of commercial piston operators to stay in business...

Dan

 

[Dan Thomas]

Except that within the last ten years, EFI, electronic coil-over-plug spark control systems and their associated ECU's have become so reliable that they typically just don't fail. I have put about three quarters of a million miles on vehicles (cars, trucks and motorcycles) so equipped in the last 12 years and I have not had a SINGLE failure of the EFI or spark systems. Some mechanical transmission issues and emissions control issues yes. In the last 500 hours of flight, I've had at least one new set of mags and a handful of unscheduled timing adjustments.

Those systems do indeed work reliably, as long as they're getting power. It's the power issue that remains, and it requires redundant alternators and batteries. Not impossible, just takes up room and reduces useful load.

E-mag is working on certified electronic magneto replacements for us little-airplane folks. They have uncertified units for homebuilders; these just go into the magneto's hole and make the normal spark. They've an alternator within them to power them independently and have variable timing, too. A good start. http://www.emagair.com/

Dan

 

[Capt. Geoffrey Thorpe]

ever have a car towed up or down significant altitude changes?

I have. 3000 feet, it took a good 15 minutes for the car to be back to normal, Fuel Injected.

if you're taking 15 minutes to climb to 3000 then that's OK, but most of us aren't.


To answer your question Where I come up with this stuff? Personal experience.

OK, your car obviously didn't have a barometric pressure sensor. Most cars don't now a days because cars don't change altitude very fast and we can make a pretty good guess over time - and leaving the sensor out saves $3 to $5 - that's a lot of money. If I were designing the controls for an aircraft instead of a car, then I would be more likely to include a barometric sensor.

It isn't a question of basic capability, but one of design / cost trade offs for a particular application. Someone saved a few bucks on sensors for your car.

But then, it also depends on how the air flow is measured. Most current production automobiles use one of two systems - a Mass air sensor, or a speed density calculation. The mass air sensor uses a hot wire exposed to a portion of the air flow and determines the mass (and I mean mass, not volume) flow rate of air into the engine - from that I can determine how much fuel to supply. Speed density is a system where you measure the absolute pressure and temperature of the air in the intake manifold and thus calculate the density. Given the density of the air in the manifold, the engine speed, and assumptions about the volumetric efficiency, I can again determine how much fuel is needed. Those two systems are relatively insensitive to altitude changes. Your vehicle may have used a vane based sensor or a throttle angle based air flow estimate that is much more sensitive to altitude. It would make sense to design the sensor set to match the application and expected operating conditions.

You may have noticed that when you step on the gas pedal, the engine keeps running. That is because the calculations in the powertrain control module are updated on a time scale of tens of milliseconds. And, as fast as you can change the air flow by moving the throttle, I can change the fuel to match (and even make adjustments to compensate for the fact that some of the fuel injected ends up on the walls of the manifold as a "puddle".)

Keeping up with a 2000 feet / minute change in altitude would be trivial compared to the other things that I have to compensate for (fuel vapors from the evap canister and fuel tank vent, for example).

 

[Capt. Geoffrey Thorpe]

Except that within the last ten years, EFI, electronic coil-over-plug spark control systems and their associated ECU's have become so reliable that they typically just don't fail.

The auto industry has been through the school of hard knocks...

People spend a lot of money installing GAMI fuel injectors to even the fuel distribution to individual cylinders. This issue completely disappears with EFI systems, and EFI systems can much more accurately meter fuel than mechanical systems, improving performance even if we were to retain a manually operated mixture control feature. But why would you want one - I could see maybe putting a manual override in an aircraft engine. EFI systems in most cars compensate for altitude just fine.

The (existing) EFI systems in cars rely a lot on the oxygen sensors in the exhaust system. But those won't work for aircraft using leaded fuel. On the other hand, one could design the system to control based on measured exhaust temperatures - you could periodical adjust the mixture to "peak" then lean to the desired temperature drop (just like a pilot). You could also learn what mixture corresponds to peak or the target lean of peak to reduce how often you need to "touch base".

All these systems take millions of dollars in R&D to develop. Problem is that's cheap when you can sell millions of these cloned systems in automobiles world wide, but not so cheap when your market is maybe a few thousand at most domestically, irrespective of the FAA certification process. I think the real answer is that if we want these systems for our aircraft (I do) Lycoming and TCM aren't going to develop them. I can see Ford or GM or some other major automotive manufacturer developing a modern series of aircraft engines under a subsidized government contract.

I would look to one of the tier 1 suppliers (Bosch, Siemens, Vistion, Delphi, etc.) instead of an OEM. They currently supply "turn key" powertrain controllers to the smaller auto manufacturers.

They would also be starting from a proven reliable hardware / software base - something you won't get from a small start up company.

 

[gismo]

I would look to one of the tier 1 suppliers (Bosch, Siemens, Vistion, Delphi, etc.) instead of an OEM. They currently supply "turn key" powertrain controllers to the smaller auto manufacturers.

They would also be starting from a proven reliable hardware / software base - something you won't get from a small start up company.

I suspect that none of those suppliers would want to enter the aviation market because of the liability exposure.

 

[Ted]

I suspect that none of those suppliers would want to enter the aviation market because of the liability exposure.

Also, none of the off-the-shelf computers they have will pass FAA certification testing, so they will need to do something new.

 

[gismo]

Also, none of the off-the-shelf computers they have will pass FAA certification testing, so they will need to do something new.

JOOC, what makes you say that? It's my understanding that most if not all automotive ECUs are designed to standards similar to FAA requirements, save possibly the code inspection part. They certainly operate in similar environments other than atmospheric pressure.

 

[Ted]

JOOC, what makes you say that? It's my understanding that most if not all automotive ECUs are designed to standards similar to FAA requirements, save possibly the code inspection part. They certainly operate in similar environments other than atmospheric pressure.

The environmental requirements are similar in general (at least with regards to temperature), and the ECUs themselves may operate acceptably. The biggest thing has to do with electrical requirements, focusing especially on radiated emissions. There are other details.

If you go to any of the airshows and look at the iE2 engine on display and compare that to your car, you'll notice a lot of similarities, but you'll also notice a lot of differences in terms of additional hardware and the like. Such items wouldn't be added if there wasn't a reason why they were required.

 

[Sac Arrow]

The (existing) EFI systems in cars rely a lot on the oxygen sensors in the exhaust system. But those won't work for aircraft using leaded fuel. On the other hand, one could design the system to control based on measured exhaust temperatures - you could periodical adjust the mixture to "peak" then lean to the desired temperature drop (just like a pilot). You could also learn what mixture corresponds to peak or the target lean of peak to reduce how often you need to "touch base".

This is true, although most EFI motorcycles and some high performance car engines without O2 sensors use a fixed map, with corrections for either atmospheric or manifold air pressure to account for changes in altitude. And even O2 sensor equipped motors use a fixed map on acceleration and transient conditions, and revert to a fixed map in cruise if there is a sensor malfunction. An EGT sensor feedback control loop isn't nearly responsive enough for automotive use, but there is no reason why it couldn't be incorporated in an aircraft engine (well, actually it is I guess in FADEC systems.)

 

[gismo]

The environmental requirements are similar in general (at least with regards to temperature), and the ECUs themselves may operate acceptably. The biggest thing has to do with electrical requirements, focusing especially on radiated emissions. There are other details.

If you go to any of the airshows and look at the iE2 engine on display and compare that to your car, you'll notice a lot of similarities, but you'll also notice a lot of differences in terms of additional hardware and the like. Such items wouldn't be added if there wasn't a reason why they were required.

I guess I was thinking specifically about the ECU, external sensors, and the injectors. The ECU's I've had apart appeared to be well shielded for emissions and susceptibility to EMI although I can't say for certain they'd meet FAA requirements. All I meant is that it just doesn't seem like a complete redesign would be needed. I do agree that significant changes would be required with much of that due to the different operating requirements.

 

[Ted]

I guess I was thinking specifically about the ECU, external sensors, and the injectors. The ECU's I've had apart appeared to be well shielded for emissions and susceptibility to EMI although I can't say for certain they'd meet FAA requirements. All I meant is that it just doesn't seem like a complete redesign would be needed. I do agree that significant changes would be required with much of that due to the different operating requirements.

A total redesign might not be required on the hardware, although there would still be some changes necessary. So we're on the same page there.

The software would probably be the bigger issue to adapt a standard EFI system for a liquid-cooled automotive engine to our air-cooled aircraft engines. You probably wouldn't have to make many software changes for a modern electronic system to put a Chevy 350 in an Aztec and make it fly, but that wouldn't work very well for a number of other reasons.