What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

EFII - the jury is coming in

Now, a "new" (not really) tech comes along, that has the potential to makes things better. What do you know? It is crazy nonsense again...at least until it becomes the new status quo.

Electronic fuel injection is certainly not crazy nonsense. In many ways it is better than a mechanical system, and in a few ways worse. I think readers desire an honest discussion of fact, regardless of their personal choice, as they prefer to know the truth about both the better and the worse.
 
Electronic fuel injection is certainly not crazy nonsense. In many ways it is better than a mechanical system, and in a few ways worse. I think readers desire an honest discussion of fact, regardless of their personal choice, as they prefer to know the truth about both the better and the worse.

It is better. The question is, is it worth the cost?

Magnetos are indeed very old technology but they work, not perfectly, but they work.

My take on it is when the first mag fails, it will be replaced with EI. I would have done it already if I could have sold a new mag at 2/3's its cost, but their were no takers so I did not do it.

To convert to EFI with a Lycoming is a big deal, just not sure it is worth it. EI is easy and its benefit is apparent and obvious from the git-go. How much better EFI is beyond EI is subjective and matter of personal choice.

My take on Rober Paisley is all positive, he is a stand up guy and if you go with his EFI system, it will be ok. If my money bucket was over flowing I would do it just for the kicks.
 
Electronic fuel injection is certainly not crazy nonsense. In many ways it is better than a mechanical system, and in a few ways worse. I think readers desire an honest discussion of fact, regardless of their personal choice, as they prefer to know the truth about both the better and the worse.

We know that. But a lot of people in the light GA segment see anything as crazy nonsense if it's not decades-old practice. It's like I said elsewhere, if a practice or method goes long enough without advancement or changes, it seems to ossify into "it's been done this way so long that it must be the best way it could ever possibly be done", and anything not that way is Inconceivable!

That's certainly not to say every new thing must be adopted hastily and completely; the proper testing and design work must still be properly accomplished. But it also shouldn't be automatically dismissed just because it's not the "traditional" way of doing things.

"There are two kinds of fool in the world. The first says 'this is old, and therefore good'; the second says 'this is new, and therefore better'".



It is better. The question is, is it worth the cost?

Magnetos are indeed very old technology but they work, not perfectly, but they work.

My take on it is when the first mag fails, it will be replaced with EI. I would have done it already if I could have sold a new mag at 2/3's its cost, but their were no takers so I did not do it.

To convert to EFI with a Lycoming is a big deal, just not sure it is worth it. EI is easy and its benefit is apparent and obvious from the git-go. How much better EFI is beyond EI is subjective and matter of personal choice.

My take on Rober Paisley is all positive, he is a stand up guy and if you go with his EFI system, it will be ok. If my money bucket was over flowing I would do it just for the kicks.

To me, EFII is worth it, despite the upfront complexity, because of the operational simplicity and the efficiency. The idea of not having to fiddle with a mixture knob every time I change the power is very appealing, and the system architecture addresses the majority of the occasional issues with running mogas/E10, or even fuel injection, with a fuel system that wasn't originally designed for it. I don't mind spending a little more time, money, and effort up front if it saves me those things later. Unless something changes between now and engine-buying time, I'm going with EFII.
 
Last edited:
Electronic fuel injection is certainly not crazy nonsense. In many ways it is better than a mechanical system, and in a few ways worse. I think readers desire an honest discussion of fact, regardless of their personal choice, as they prefer to know the truth about both the better and the worse.

Yes, I appreciate the discussion on this. I'm learning a lot (or at least seeing things that make me do more research).

Mods, I think this thread might be more appropriate for one of the engine-related subsections, since it's not just RV-10 specific. Would it be possible to move it?
 
I like to see the feedback here from users and those considering EFI. Obviously here are many reasons why someone might choose to use it, whether or not it provides any useful performance or fuel economy gains even.

Interested to hear any more ideas and comments.
 
I would be concerned about the Lycoming intake manifold.
Is it optimized for EFI? Probably not.
Another issue is 100LL and the O2 sensors.
I would be happy with EI but not sure EFI and Lycoming are good partners.
 
The idea of not having to fiddle with a mixture knob every time I change the power is very appealing

If that is a problem to you then you are doing it wrong. Let me explain by way of example. Recent flight doing an Instrument renewal. 6 Mixture functions in all and i barely recall doing them. I select radio and nav functions more often in a flight and they are far more complex. Programming the flight plan in a GTN is more involved.:eek:

1. Prime/Start > Mixture never got past half way then leaned back to see the RPM rise.
2. Mag check done aggressively lean as possible
3. Full rich for takeoff
4. Lean to target EGT in the climb at 3500-4000'
5. Big Mixture pull levelled out at 6000' and appropriately LOP. Simple as that.
6. Idle Cut Off at the hangar door an hour and a half later.

In that time the plane did an ILS with a hold, VOR-DME arc app, RNAV app, and all the climbs and repositions and missed approaches, plus unusual attitude recovery checks, all using nothing but throttle. The FCU took care of everything.

Unfortunately over time pilots have been taught by instructors who knew no better, and just made simple things very complicated and scary for the pilot, because the instructors were scared themselves by old wives tales.
 
Dan - Thanks for that exposition. My -8 is at 32 hours. Cruise CHT's within 9 or 10 degrees (320-330) on Lycon IO360M1B with 10:1 compression. EGT's vary more, with 1 and 4 being higher (1229,1180,1149,1241). All @ slightly below peak EGT, OAT 55F - the pattern holds @ widely different speeds and mixtures. Seems time to perform GAMI lean test and consider swapping/changing injectors. Would you do anything else first? Please pitch in, everyone. Thanks.

Sorry Bill, I think you got run over in the stampede. Let's look at what you have...

First, is this an EFII electronic fuel injection, a mechanical fuel injection, or a carb?

Second, are you "All @ slightly below peak EGT" on the rich side or the lean side?
 
OTS injectors are usually flow matched to about 2-3% variance. Along with this, each port flows slightly different amounts of air so this results in the same issues as with mechanical injectors hence the GAMI idea is needed to correct AFRs for individual cylinder.

GAMI spread information is not applicable to electronic injection - you already have balanced fuel delivery.

I mentioned in a previous post about the GAMI spread being essentially the same as with mechanical injection using standard nozzles. That's a reality.

As previously noted, the GAMI test is really irrelevant to electronic injection, you already have the fuel balance goal achieved.

You guys should really talk more often ;)
 
(warning, beer-fueled rant to follow)

If that is a problem to you then you are doing it wrong. Let me explain by way of example.

Whether it's "easy" or "simple" or not shouldn't matter--I shouldn't have to do it in the first place. None of the cars I have ever driven, not even the oldest ones that were about my age, have required me to pay any attention to fuel-air mixture or any other parameter unless something was wrong. I've driven cars over nearly as wide a range of altitudes as any airplane I've flown, with constantly varying throttle settings, and never once had to give the slightest thought to fuel-air ratio. And all of these are cars that, brand new, still cost less than just the engine that I'm putting in my airplane.

There's no reason an airplane engine should be any different--and in many airplanes, it isn't; turbines don't come with mixture knobs, even the ones with all-mechanical engine controls. I want to get in and fly, not babysit an engine, and not play flight engineer. And since we don't have to worry about all the complicated emissions reduction stuff that cars and trucks do, it makes the control issue simpler. Simply program the controller to provide the lowest possible fuel burn for a given power setting while keeping temperatures within normal limits and not causing damage to the engine. As long as I'm not damaging the engine or causing a temperature issue, why would I ever want to burn more fuel? This is the kind of thing that automated controls are made for, and they do a lot better job of it than humans do once properly set up.

Besides, what advantage is gained from full-time manual mixture control? "Simplicity" is often claimed, but there's more than one way to look at simplicity--and operational simplicity, while sometimes ignored, counts just as much. What problem does it solve? What advantage does it gain me to be fiddling with something else in the cockpit right at the times when my workload is likely to be higher, like right after takeoff or in the pattern or in maneuvering flight with lots of power changes? To me, it seems like this level of manual control has a lot more potential for pilots to introduce problems that otherwise wouldn't exist (how many engines are driven to early overhaul, or fail in flight, due to improper mixture management? How many thousands of gallons of fuel get burned needlessly because pilots don't want to, don't know how to, or for some reason can't run at an optimum LOP condition? and for that matter, how many airplanes get totaled every year due to carb ice, which could have been obviated by not having a carb in the first place?) than it does to save a pilot from something they otherwise wouldn't catch. Yes, there are failure modes that might require pilot intervention to save the engine, but for those instances an override can be provided, and requiring full-time manual mixture control under all other conditions isn't going to make that problem any less likely.

The mixture knob, like the vacuum gyro, looks like another one of those solutions that was developed a long time ago (100+ years, in this case) because there just wasn't any better way of doing it at the time, and there continued to be no better way of doing it (or at least, no affordable way of doing so) for a few decades, up to and past the "golden age" of the light airplane. It's become one of those things that we have just accepted as normal despite being incredibly anachronistic, just because 90% or so of the light airplane fleet was built before any other feasible means of doing it was available. That still doesn't make it the "best" way to do it.

We in the light GA segment like to gripe about being stuck flying behind 1930's technology in our engines--but then if anyone suggests actually doing something about it, we gripe about that. IMHO, after modern glass EFII-like systems appear to represent the biggest bang-for-the-buck improvement in real-world operational efficiency and simplification that we're going to see for light airplanes, especially if we constrain it to things that can be relatively easily retrofitted to existing airframes.

I don't think it's too much to ask that my brand-new airplane be caught up technologically with 1980s automobiles, especially when you consider that the technology I'm going to be putting into the panel is quite recent, and has features that were bleeding edge, still-in-R&D level when I started my career testing integrated avionics packages for high-end jets.


Unfortunately over time pilots have been taught by instructors who knew no better, and just made simple things very complicated and scary for the pilot, because the instructors were scared themselves by old wives tales.

I wasn't taught to lean by my instructors, because in the tired old underpowered C150 I trained in, the throttle went to full rental power at takeoff and it didn't move again until you pulled the power on downwind, and the hourly charge was the same whether I leaned or not. The mixture knob was simply used to shut the engine off at the end of the flight.

That same old 150--the good one at the flight school, because it had a text-only GPS and was in marginally better shape than the other one--met its end between a pair of trees six weeks after I flew it for my checkride when that "simple" carburetor iced up and both the CFI and student forgot the carb heat. A perfectly good airplane ruined because of a simple oversight regarding a manual engine control.
 
If that is a problem to you then you are doing it wrong. Let me explain by way of example.

1. Prime/Start > Mixture never got past half way then leaned back to see the RPM rise.
2. Mag check done aggressively lean as possible
3. Full rich for takeoff
4. Lean to target EGT in the climb at 3500-4000'
5. Big Mixture pull levelled out at 6000' and appropriately LOP. Simple as that.
6. Idle Cut Off at the hangar door an hour and a half later.

Unfortunately over time pilots have been taught by instructors who knew no better, and just made simple things very complicated and scary for the pilot, because the instructors were scared themselves by old wives tales.

Glad you like the old ways there but not everyone feels the same way or wants the same thing in 2015. We could all drive Model T Fords with manual spark advance too but I doubt many would want to.

We have lots of choices on what engine, avionics, electronics, props, fuel delivery and ignition systems to use in our RVs. Let's respect those choices others make.

Let's also stay on topic about experiences with EFI in aircraft here...
 
Last edited:
ditto

"Glad you like the old ways there but not everyone feels the same way or wants the same thing in 2015."

No doubt...:D
 
Let's return to tech discussion please.

Whether it's "easy" or "simple" or not shouldn't matter--I shouldn't have to do it in the first place. None of the cars I have ever driven, not even the oldest ones that were about my age, have required me to pay any attention to fuel-air mixture or any other parameter unless something was wrong.

Bob, this isn't that system.

The EFI in your car runs closed loop. Short version; the O2 sensor feeds information to the electronic brain, which modifies mixture. which results in a new O2 sensor signal, and so on. System failures illuminate a check engine light.

The EFII-branded system is open loop. The O2 sensor (or EGT display) feeds information to the pilot's brain. He is the meat servo, expected to use the dial-a-mixture knob, or program a new default. He is also the error check system.

We have this report from Jesse regarding manual operation (post #4)...

The simple answer is that if you want to run ROP in some cases and LOP in other cases, you do so with the mixture knob. We never run the exact same settings on every flight.....I climb at 2500 rpm and usually cruise at 2400rpm. In climb, at 25", I'm making maybe 80% power and running 11.8-12.5 AFR. In cruise I was running 12-22", which is 20-65% power roughly, and 13.1-16.8 AFR.

....and this from Ross...

There are enough small variables that LOP operation should probably be set via the mixture control using either EGT or the wideband.

The EFII installation manual specifies system setup at 13:1, the lean end of ROP best power mixture. That will require using the dial-a-mixture knob for LOP operations.

Don't want to set it up for full time LOP ops, because it can't make rated power that way.

In theory the tables could be set for transition to LOP operation at some predetermined RPM. That would give you two modes, climb/fast cruise and economy slow cruise. Perhaps 12:1 at 2500 plus, and 16:1 at 2250 and below? The actual transition would be at 2375. No extrapolation, so there would be a noticeable power sag at the transition between the two settings. Running right at at 2375 wouldn't work well.

4. Flying. Once you have a basic tune up, you don't have to touch anything to fly. The ECU will handle the entire fuel delivery task. After you have a little familiarity with the system, you can reduce the fuel numbers at cruise rpms to automatically lean. You can also go a little bit richer (12.5:1) at max power rpms.

Robert, what approach have you found most practical while flying your blue ex-Subaru RV-7, N461RP, with its new EFI Lycoming?
 
Let's return to tech discussion please.



Bob, this isn't that system.

The EFI in your car runs closed loop. Short version; the O2 sensor feeds information to the electronic brain, which modifies mixture. which results in a new O2 sensor signal, and so on. System failures illuminate a check engine light.

The EFII-branded system is open loop. The O2 sensor (or EGT display) feeds information to the pilot's brain. He is the meat servo, expected to use the dial-a-mixture knob, or program a new default. He is also the error check system.

We have this report from Jesse regarding manual operation (post #4)...

The simple answer is that if you want to run ROP in some cases and LOP in other cases, you do so with the mixture knob. We never run the exact same settings on every flight.....I climb at 2500 rpm and usually cruise at 2400rpm. In climb, at 25", I'm making maybe 80% power and running 11.8-12.5 AFR. In cruise I was running 12-22", which is 20-65% power roughly, and 13.1-16.8 AFR.

....and this from Ross...

There are enough small variables that LOP operation should probably be set via the mixture control using either EGT or the wideband.

The EFII installation manual specifies system setup at 13:1, the lean end of ROP best power mixture. That will require using the dial-a-mixture knob for LOP operations.

Don't want to set it up for full time LOP ops, because it can't make rated power that way.

In theory the tables could be set for transition to LOP operation at some predetermined RPM. That would give you two modes, climb/fast cruise and economy slow cruise. Perhaps 12:1 at 2500 plus, and 16:1 at 2250 and below? The actual transition would be at 2375. No extrapolation, so there would be a noticeable power sag at the transition between the two settings. Running right at at 2375 wouldn't work well.



Robert, what approach have you found most practical while flying your blue ex-Subaru RV-7, N461RP, with its new EFI Lycoming?

The ECU has closed loop capability but the standard AFR target is 14.7 (stoich). This would result in peak EGT which would be a bad place for an engine to operate at under high load.

In the works is new software to allow the user to target different AFRs at different load/ rpm ranges when fed info from a wideband O2 controller. We tested a single point, hard loaded version back in 2007 on a Subaru powered RV. Worked awesomely but we were concerned about how to deal with the O2 sensor taking a dump and possibly melting the engine. We were not comfortable in releasing the software without sensor diagnostics in place.

The next software evolution will use a couple of different strategies to check O2 sensor validity and revert to open loop if things don't add up plus turn on the check engine light to warn the pilot about it.

We may also add a LOP window to the programmer to target a specific cruise AFR.

This is one of many new ideas our overworked software guy has to write code for after he finishes numerous other tasks...
 
EFII tuning

13:1 air fuel ratio is a nice safe starting point to make a safe and happy engine. The next level of playing with EFII numbers, if you feel so inclined, is to add a little bit of fuel at takeoff power to say 12.5:1 AFR and take away a little bit of fuel at cruise power, say to 15 or 16:1 AFR. At that point, you have a fully tuned engine that leans itself in cruise, and richens a little for takeoff.
After that, you just have to worry about where you want to fly to lunch.
The engine will take care of itself.

Once you fly behind an engine that does what you tell it to do without having to re-tell it every time. You will wonder why we didn't do this 20 years ago. Actually, there are a lot of us out there that have been wondering this for a long time!

Robert
 
Let's return to tech discussion please.

Bob, this isn't that system.

The EFI in your car runs closed loop. Short version; the O2 sensor feeds information to the electronic brain, which modifies mixture. which results in a new O2 sensor signal, and so on. System failures illuminate a check engine light.

Fair enough. I understand that car engines will have more in the way of automation, and that the open-loop system is set up for a reason.

The EFII-branded system is open loop. The O2 sensor (or EGT display) feeds information to the pilot's brain. He is the meat servo, expected to use the dial-a-mixture knob, or program a new default. He is also the error check system.

Error check I can live with, and I don't think that's going to change because the consequences of failures are greater than in cars (though a good engine monitor can help with this) and programming a new default I can work with (and is actually what I'm thinking I'd be doing anyway). But with a programmable controller, I don't understand why we can't program that controller to meet different mixture settings for different power settings (which I assume is what you mean by "program a new default"), even if it means a fair bit of trial and error and tuning first. I'm assuming that the engine is repeatable--that the same inputs (RPM, MP, and timing) will yield more or less the same AFR and approximately the same EGT; if that's not the case then I guess things get a little more complicated.

We have this report from Jesse regarding manual operation (post #4)...

The simple answer is that if you want to run ROP in some cases and LOP in other cases, you do so with the mixture knob. We never run the exact same settings on every flight.....I climb at 2500 rpm and usually cruise at 2400rpm. In climb, at 25", I'm making maybe 80% power and running 11.8-12.5 AFR. In cruise I was running 12-22", which is 20-65% power roughly, and 13.1-16.8 AFR.

....and this from Ross...

There are enough small variables that LOP operation should probably be set via the mixture control using either EGT or the wideband.

The EFII installation manual specifies system setup at 13:1, the lean end of ROP best power mixture. That will require using the dial-a-mixture knob for LOP operations.

That's not what I had been led to believe in previous discussions, which is why I'm confused now.

Don't want to set it up for full time LOP ops, because it can't make rated power that way.

In theory the tables could be set for transition to LOP operation at some predetermined RPM. That would give you two modes, climb/fast cruise and economy slow cruise. Perhaps 12:1 at 2500 plus, and 16:1 at 2250 and below? The actual transition would be at 2375. No extrapolation, so there would be a noticeable power sag at the transition between the two settings. Running right at at 2375 wouldn't work well.

That seems a perfectly reasonable strategy, and it was my understanding from previous discussion that this does indeed work--that is, preprogramming the ECU to run LOP under certain conditions and ROP/best power when needed. However, would it be better to program this by RPM, or by MP?


The ECU has closed loop capability but the standard AFR target is 14.7 (stoich). This would result in peak EGT which would be a bad place for an engine to operate at under high load.

In the works is new software to allow the user to target different AFRs at different load/ rpm ranges when fed info from a wideband O2 controller. We tested a single point, hard loaded version back in 2007 on a Subaru powered RV. Worked awesomely but we were concerned about how to deal with the O2 sensor taking a dump and possibly melting the engine. We were not comfortable in releasing the software without sensor diagnostics in place.

The next software evolution will use a couple of different strategies to check O2 sensor validity and revert to open loop if things don't add up plus turn on the check engine light to warn the pilot about it.

We may also add a LOP window to the programmer to target a specific cruise AFR.

This is one of many new ideas our overworked software guy has to write code for after he finishes numerous other tasks...

Interesting! That would be icing on the cake :)

The next level of playing with EFII numbers, if you feel so inclined, is to add a little bit of fuel at takeoff power to say 12.5:1 AFR and take away a little bit of fuel at cruise power, say to 15 or 16:1 AFR. At that point, you have a fully tuned engine that leans itself in cruise, and richens a little for takeoff.
After that, you just have to worry about where you want to fly to lunch.
The engine will take care of itself.

That's what I've been wanting this whole time:
an engine that does what you tell it to do without having to re-tell it every time.
 
Last edited:
In the works is new software to allow the user to target different AFRs at different load/ rpm ranges when fed info from a wideband O2 controller.

Excellent, and I know it will be good. However, right now, as stated, the EFII-branded system is purely open loop, delivering only what its base tables dictate for any particular RPM and density combination. In-flight variation requires pilot input via the mixture knob, or, if the change is to be permanent, a new value for the table(s).

Operationally, this system is precisely analogous to a mechanical fuel injection system. The default ("the base table") is full rich. In-flight variations from the base requires pilot input via the mixture knob, or if the change is to be permanent, by swapping brass parts.

Both systems require pilot input if anything other than the preset base fuel flow is desired.
 
Excellent, and I know it will be good. However, right now, as stated, the EFII-branded system is purely open loop, delivering only what its base tables dictate for any particular RPM and density combination. In-flight variation requires pilot input via the mixture knob, or, if the change is to be permanent, a new value for the table(s).

Operationally, this system is precisely analogous to a mechanical fuel injection system. The default ("the base table") is full rich. In-flight variations from the base requires pilot input via the mixture knob, or if the change is to be permanent, by swapping brass parts.

Both systems require pilot input if anything other than the preset base fuel flow is desired.

Pretty much yup. Just to be clear for users, any single point in the map is adjustable in flight via the panel mount programmer.

The mixture knob adds or subtracts the indicated percentage from the final injector pulse width.

I guess we should mention that this system automatically leans in the climb as MAP drops off and richens in the decent so it's a bit smarter than than mechanical injection in that regard. It also takes no action from the pilot with regards to cold/ hot starting techniques, especially hot starting. Idle/ taxi mixture is not a matter of guesswork as with mechanical injection either.

The ignition tables are fully user programmable as well to suit any unusual mission- maybe high altitude, low power settings for example. Or low and fast at Reno. Mike Dacey and Kevin Eldredge used it to good advantage racing.
 
Last edited:
I guess we should mention that this system automatically leans in the climb as MAP drops off and richens in the decent so it's a bit smarter than than mechanical injection in that regard. It also takes no action from the pilot with regards to cold/ hot starting techniques, especially hot starting.

There we go...good solid points for EFI.

The next level of playing with EFII numbers, if you feel so inclined, is to add a little bit of fuel at takeoff power to say 12.5:1 AFR and take away a little bit of fuel at cruise power, say to 15 or 16:1 AFR. At that point, you have a fully tuned engine that leans itself in cruise, and richens a little for takeoff.

So please, tell us how how you handle the transition in your own airplane. At what RPM or manifold pressure do you make the transition? Do you program for the automatic switch from rich to lean in one jump, like this example....

Perhaps 12:1 at 2500 plus, and 16:1 at 2250 and below? The actual transition would be at 2375. No extrapolation, so there would be a noticeable power sag at the transition between the two settings. Running right at at 2375 wouldn't work well.

...or spread the transition across a wider range of RPM?

Doesn't seem like we would want to use a manifold pressure point to trigger a transition, as it would automatically reduce climb power by going LOP above a certain altitude.
 
We set it to 11.8 at 2750, which held from 2625 and up for takeoff and initial climb. Then we set it for 13 at 2500, which worked great for climb. Then for cruise we could set it at 16.5 or so for lop economy at 2250, which would work up to 2375. If you want ROP and max speed in cruise, you could run 2400+. For economy, pull back to 2370 or lower and it automatically goes to economy.

In our testing, as we already talked about, there is no interpolation, so you can work the system like this.
 
The consensus at the Egg Factory, with regard to an ECU other than the Subby unit, was that the average pilot was not capable of managing AFR - in time he would screw up and destroy the engine. To the best of my knowledge they did not offer that option, but I don't know for sure. I never ventured away from the stock ECU.

That view is not universal as we now have this discussion of the pilot managing AFR rather than the ECU. Not a big deal but to say it is less work than pulling a mixture knob does not ring true. Misuse of either could harm the engine.

I liked closed loop operation as it went to the tables and really tweaked the operation. But it was available only at very low power settings. At high power requirements the system ignored the tables and poured the fuel in to protect the engine. I believe managing AFR is just as critical as pulling the mixture, you have to know the operation to not mess up your engine.

Setting the knob for correct AFR for take off is just as critical as full rich with the mixture.
 
Last edited:
The consensus at the Egg Factory, with regard to an ECU other than the Subby unit, was that the average pilot was not capable of managing AFR - in time he would screw up and destroy the engine. To the best of my knowledge they did not offer that option, but I don't know for sure. I never ventured away from the stock ECU.

That view is not universal as we now have this discussion of the pilot managing AFR rather than the ECU. Not a big deal but to say it is less work than pulling a mixture knob does not ring true. Misuse of either could harm the engine.

I liked close loop operation as it went to the tables and really tweaked the operation. But it was available only at very low power settings. At high power requirements the system ignored the tables and poured the fuel in to protect the engine. I believe managing AFR is just as critical as pulling the mixture, you have to know the operation to not mess up your engine.

Setting the knob for correct AFR for take off is just as critical as full rich with the mixture.

Actually over 140 of the same SDS ECUs were fitted to Egg EZ30/36 engines since the later OEM ECUs (post 2006 MY)were not user friendly due to chassis tie ins which were no longer there in the aircraft. Jan stated that users would not be given either mixture knobs or programmers and the ECUs were locked out at his request to prevent tampering. This was counter to our advice at the time.

The idea was to establish a safe map through testing and burn the same values into every ECU. Where this went wrong was the cold start mapping was never completed outside warm climates (he was warned about that one too) and that several years of engines with slightly different specs were used without repeating the mapping validation.

Later it was also found that true ignition timing values were not always the same between engines due to variations in timing disc placement. This MUST be verified in all cases on aftermarket ECUs.

Additionally, initial mapping was carried out on 100LL fuel where many people were later running 91 unleaded. You can probably see where this is going. After some piston failures, many ECUs were hastily reprogrammed with reduced total timing. Engines in colder climates than Florida had to have ECUs returned for more guesses at cold start values etc. There were other issues where the work was not done right with regards to magnet mounting and hall sensor alignment to which customers had to deal with later.

Most auto engines won't tolerate too much timing with their higher CRs on 91 fuel nor will they tolerate running leaner than mid 12ish AFRs for very long before the pistons start to be distressed at high power/rpm. This was all well known by us but not everyone would listen.

After Jan left the Sube scene, we continued to support users with our ECUs but still had to beat it into people to check timing, limit timing on 91 fuel, buy a wideband AFR meter to verify that the AFRs were safe etc. We sold numerous mixture knob and programmer kits along with unlocking ECUs so users could check and reprogram properly. Those who listened had generally good success, those who didn't usually melted pistons at between 2 to 50 hours of engine operation. This stuff is really critical on auto engines with cast pistons. The auto OEMs really pull back timing at WOT around torque peak rpm and throw the fuel to it to reduce chances of detonation and keep piston crown temps within limits. You have to do the same with any aftermarket ECU to get similar longevity.

Robert had years of experience with the Sube engines and the SDS ECU and applied all that experience to adapting it properly and successfully to the Lycoming with the EFII. I think you can see the results of that experience today. He did extensive testing on his own mule aircraft before product release. His kit has all the critical parts ready to bolt on for the most part. This is key in my view and he is giving good customer service by all the accounts I read here- also key.

There will be better documentation, base maps and new advances in the future to improve out of the box performance and save tuning time.

The high time EFII user has around 1600 flight hours on the system and has provided much useful feedback for mapping and feature improvements.

It's great to see people like Jesse and the others here report on their experiences. It helps the whole community and helps to improve the product. Thanks again to Jesse for starting the thread. Judging by the number of views, many people would appear to be interested.
 
Last edited:
We set it to 11.8 at 2750, which held from 2625 and up for takeoff and initial climb. Then we set it for 13 at 2500, which worked great for climb. Then for cruise we could set it at 16.5 or so for lop economy at 2250, which would work up to 2375.

A two-step transition...thanks Jesse.

A 16.5 AFR would be 50 LOP. You previously reported a GAMI spread of 1.4 GPH. Even 0.5 results a significant difference between first and last cylinder to peak, so you probably had some cylinders very lean and some just below peak. Was the system in fact stable enough that you could pull the prop knob below 2370 at any altitude, and have every cylinder end up comfortably LOP, without touching the mixture trim knob?
 
Would be interesting to know if these are the same 2 cylinders which peak early with a mechanical injection setup and std injectors?

What kind of flow rate variance does GAMI offer for the the 540?
 
What kind of flow rate variance does GAMI offer for the the 540?

Don't know about GAMI, but I balanced injectors on my 540 by putting in different sized nozzles from Airflow Performance. These are balanced closer than .4 GPH, may be as good as .2 (testing without a vernier on the mixture has its limitations). I can run very lean of peak (9 gph at 23 squared at 3000' ASL) without any engine stumbling, however there is a very noticeable loss of power there - you get a pretty good kick in the butt when pushing the mixture back in.
 
Don't know about GAMI, but I balanced injectors on my 540 by putting in different sized nozzles from Airflow Performance. These are balanced closer than .4 GPH, may be as good as .2 (testing without a vernier on the mixture has its limitations). I can run very lean of peak (9 gph at 23 squared at 3000' ASL) without any engine stumbling, however there is a very noticeable loss of power there - you get a pretty good kick in the butt when pushing the mixture back in.

Do you recall which cylinders were peaking first before?

What flow rate differences were required to even things up?
 
Do you recall which cylinders were peaking first before?

What flow rate differences were required to even things up?

The stock injectors are .028. The initial tests included one at 8500 PA, 38* F, 2300 RPM and 21" MP.

CYL...FF
2......11.3
6......11
1, 4..10.8
5......10.7
3......10.3
4......10.1

Note that #4 had a double peak.

I put in new injector bodies with .024 restrictors, #2 was still peaking early so I put in an .025 restrictor in #2. This was the final result:
7500 PA, 32* F, 2300 RPM and 21" MP

CYL....FF
1, 2...12.1
5, 6...11.9
3, 4...11.7

For good comparability I should have tested at the same PA and temp, but I wasn't looking for comparability when I was testing, just narrowing the peak margin.
 
Last edited:
Ross, don't know if this will help, but here's a peak EGT fuel flow spread (which we've learned to call a GAMI spread), recorded on the 10th of this month somewhere over southern Virginia. The total fuel flow spread within which all cylinders reach peak is 0.1 GPH or less (red, 8.9-9.0), and that's within the dither of fuel flow indication.



Data from Phase 1 back in 2010, similar conditions. Spread is 0.8 GPH. Needed two restrictors changed to bring cyls 1 and 2 into line, one smaller, one larger.



AFP FM200. I could look up the injector size changes, and I'd bet serious money that Don Rivera could tell me the change in fuel delivery, per nozzle increment.
 
Last edited:
Thanks guys, interesting data.

Dan, would be good to know how the fuel flow rate to each nozzle was affected in percentage to have some idea how much different the airflow rates are between cylinders. This would allow some correlation to what is being seen with the EFI.
 
Having the GAMI spread small is desirable but like I said before if one cylinder is 16.1 AFR and another 16.7, how much does it really matter?

It absolutely does not matter when all cylinders fall into the 16.1~16.7 range.

However, the GAMI spreads reported here indicate a somewhat wider range of AFR for various cylinders than you're offering.

Jesse, you there? With a 1.4 GPH GAMI, how many degrees LOP was your leanest cylinder when the richest peaked? 90F, 100F?

Not far off. Here's some O-360 data borrowed from our own Alex Peterson; a 1.3 GPH GAMI is about 75F.

BTW, that's an O-360, not an IO...a carb.



While we're waiting, consider the question from a GPH view. Here's an example from the power side of peak rather than the economy side.....since we've heard so much gnashing of teeth over horsepower ;)

Best power cruise means mixture must be set so the leanest cylinder is not leaner than 100 ROP. With a 1.4 GPH GAMI, that forces the richest cylinder to be quite rich. Balance them to a 0 GAMI spread (adjust all the individual flows to the average), and you get the same or more power on 0.7 GPH less fuel. You're able to lean all of them to 100 ROP.
 
Last edited:
Just thinking out loud but I suspect you could balance cylinder HP (fuel flow) based on 1/2 order vibrations which are primarily a result of the combustion cycle gas pressures which includes such things as compression, mixture, and induction losses to name a few. Here are a couple of plots showing how the 1/2 order (0.5x on the chart) vibs can vary on good running engines. The only problem with this method is you can't tell which cylinder or cylinders are not carrying their share of the load.

This was an extremely smooth running engine, .103 IPS on the 1/2 order:

Steve%252520Christo%252520IO390%252520Vib%252520Survey.png


While this one had .347 IPS on the 1/2 order:

BH%252520RV4%252520spectrum.png
 
Last edited:
If we use that 1.4 gph number and total FF was say 14 gph, that's a pretty nasty imbalance. Can we attribute that mostly to different air mass flow and a bit to production tolerance of the injectors? We don't really know the answer without more data.

Let's assume that cylinder #1 peaked first at 14.7 AFR and cylinder #6 was richest then at around 13.4 AFR. Assuming linear spread, with #1 at 16.5 AFR, #6 would be at around 15 AFR, still LOP, but barely.

We could strain gauge the crank and use a Hall Effect sensor to identify #1 cylinder, data log from the scope throughout the rpm and MAP range to see what torque each cylinder is producing. Would be fascinating I think.

I'm thinking back on data from the Questair Reno racer with EFI and the Conti 550 which has a known, fairly decent intake manifold. Of course we were running close to RBT or a bit richer but EGT spread there was typically around 35F if memory serves me correctly. This was running OTS Siemens 83 lb. injectors. This does not seem too far off Lycomings with adjusted low GAMI spread.

We know some here on VAF are running LOP with carbs and some carb heat with no way to adjust individual cylinder AFR directly. Maybe not optimal but it works.

I was on an Australian forum today discussing EFI for the Jabiru engines. One fellow running the standard Bing carb reported a 120C difference in EGT on his 3300 which is not uncommon. He also said the hot cylinder was a different one depending on throttle angle. These engines are known to have uneven mixture distribution with the carb and have been drastically improved when switched to EFI. Just a bad manifold design for use with the carb. I see guys on the Jab forums discussing how to correct this, ad nauseam.

In the automotive world, when designing plenum style intake manifolds, we always attempted to have equal lengths, angles and bends from a very large plenum with the runners tipped with velocity stacks and often used tapered runners to increase airflow and velocity at the port. I never really saw big variances of EGTs between cylinders using OTS injectors during dyno runs. The point being you can't make a silk purse outta...

Personally if I flew a Lycoming, I'd probably be building a proper intake for it. It would just bug me otherwise. OEM auto ECUs don't individually trim cylinder AFR because the engineers designed proper manifolds in the first place that flow close to equal air, squirt in equal amounts of fuel and Voila!, you get near equal AFRs and near equal work from each cylinder. Simply the way it should be done in my view, coming from my background.

Anyway, enough ranting, let's see more real world data and learn something...
 
Last edited:
Jesse, you there? With a 1.4 GPH GAMI, how many degrees LOP was your leanest cylinder when the richest peaked? 90F, 100F?

I don't have the data I. Front of me, but it was around 100F or more lean of peak on the leanest when the richest peaked.
 
There are no plans in the near future to go to a full timed injection strategy with the injector outputs be able to be individually trimmed. This would involve a completely new CPU, ECU, crank sensor, wiring and software and probably quadruple the system cost. We'd be competing against existing designs like the $5k (ECU only) MoTec M800 series, which does not make much sense for a small company like us. It would take us years to develop and validate to the same level of hardware and software we have today. With over 50 different programmable ECUs on the market today, you need to pick and choose your markets and battles to survive. Going head to head against the big player's products would be a sure way to ruin IMO.

I assume by "timed injection" you mean that the injectors are individually fired according to each cylinder's point in its cycle. Would that really be necessary just to change the pulse time for each injector in a batch fire system? I'm thinking you could still fire all the injectors at once but just vary the time each one is open, maybe as a percentage of the nominal "open" time.

Of course, if the spread is different over different throttle settings, then it gets even crazier. That's probably too much to manage, especially if the primary points we're concerned about are just full power and/or cruise.
 
I assume by "timed injection" you mean that the injectors are individually fired according to each cylinder's point in its cycle. Would that really be necessary just to change the pulse time for each injector in a batch fire system? I'm thinking you could still fire all the injectors at once but just vary the time each one is open, maybe as a percentage of the nominal "open" time.

Of course, if the spread is different over different throttle settings, then it gets even crazier. That's probably too much to manage, especially if the primary points we're concerned about are just full power and/or cruise.

I don't like the industry term "sequential" since it's not accurate. All EFI is sequential by nature- one injection event is completed and the subsequent ones follow. By timed, I mean, injection events are timed to valve position. People might be surprised to know that some "sequential" OEM systems actually open the injector before the valve opens, others only start spraying after intake flow is established however at high rpm, the amount of time available to inject is reduced and in order to get enough fuel into the cylinder, you must either have a very large injector and/or spray almost continuously- so technically the "sequential" part becomes little different than the batch fired strategy at that point. When the injector duty cycle gets to 100%, the injector is spraying continuously.

We usually size the injector to limit duty cycle (injector open time) to less 80% at full power. This helps the drive transistors and injectors stay cool. On the Lycoming setups, we use a rather large injector due to the large cylinder volume (compared to automotive). This is done to aid cold starting. As such, the injector duty cycle would not exceed 50%. By the continuous nature of aircraft engines, this also ensures maximum reliability of the components by keeping things cooler.

"Sequential" systems came along about 20 years ago mainly as a response to tougher emissions regs, especially during cold warmup and showed some gains at light load as well but power and economy at high load is little different. One guy proved this running a 500+hp Nissan SR20DET on both SDS and MoTec systems back to back on the same dyno. Hp was within 2 (experimental error).

To answer your question, yes, technically you would not need timed injection to modify individual injector pulse width but you would need separate injector channels which is something we don't have and would require an all new design to do plus all new software.

You've also guessed right that at different rpms and MAP, mixture distribution may not be the same and trim might have to change so programming might become much more complicated.
 
Last edited:
Ok, so what does a 100F spread between the first and last to peak translate to in terms of actual fuel-air ratio delivered to the individual cylinders...in the case of a 540, six little engines in close formation.

Zoom back to the 1930's, radial engine development, at the NACA. A 100F spread on the lean side of peak was a AF ratio spread of almost 3 points:



Jump to a more recent Lycoming document. Again, a 100F spread is almost 3 points of AF ratio:



As noted previously, when running LOP a rather wide FA spread isn't a plus, but it doesn't actually hurt anything either. However, the spread is still there on the rich side of peak, when we want to make power.

FAA Hughes Test Center, detonation test mixture sweep, with RPM and MP set a little bit oversquare, a pretty good approximation of a fixed pitch takeoff on a hot day. Here a 100F spread equates to 2.2 points of FA ratio, more than enough to span the detonation range.



Given the current inability to adjust FA for individual cylinders, it appears that EFII users really need to pay careful attention to EGT and mixture management. Setting the system to an initial 13:1 AF per the instructions could have unfortunate consequences when things stack up wrong. For example, assume the wide band sensor is installed on a cylinder which, by pure lucky chance, happens to be the one exactly in the center of the AF range. That's a good thing....but the range of delivered FA ratios is still wide enough to park a cylinder or two in the detonation zone:



It doesn't matter if the lack of FA equality is due to engine porting or fuel delivery. It's a pointless debate. Reality says FA's need balanced, or some EFII installations will be no better than a carburetor. A carbed system, with its typically wide GAMI spread, avoids the detonation zone by pushing the red box above even further to the left, the target being 200F or so ROP at full rich.
 
Jump to a more recent Lycoming document. Again, a 100F spread is almost 3 points of AF ratio:



As noted previously, when running LOP a rather wide FA spread isn't a plus, but it doesn't actually hurt anything either. However, the spread is still there on the rich side of peak, when we want to make power.


An important takeaway from the Lycoming chart above is the flatness of the power curve on the rich side of peak EGT. Even at a whopping 300F ROP, we're only giving up 3% power over best power mixture. This is why Braly et al. advocate setting up engines with very rich takeoff mixtures. Rich takeoff fuel settings give up very little in terms of power but result in significant reduction in cylinder head temperature.

Skylor
RV-8
 
I agree that we should try to get spreads under 50F if possible. Since there is little fundamental difference between this and a standard mechanical FI setup with standard nozzles, the EFI is no better and no worse than that in this regard and people have not been too worried about it for several decades before GAMI and digital engine monitors ever existed nor have engines been failing with carbs or mechanical FI in great numbers from detonation. I don't think the sky is falling.

It is more ideal and makes you feel better while at the same time possibly increases engine roughness. Take your pick. The crux of the issue is poor manifold design, possibly coupled with a stack up of tolerances on injector flow rates so no point in blaming the EFI system for delivering equal pulse widths- that's what it was designed to do. We expect near equal airflow rates with EFI but simply don't have it with Lycoming engines. The very different ACE manifolds copy good automotive design practice (Toyota has used CFD for intake and cooling system design for over 30 years now) and would be a good place to start to solve distribution issues IMO. Existing solutions are available if you want them:

https://www.facebook.com/AdvancedCo...2956441867781/239863309510427/?type=1&theater

https://www.facebook.com/AdvancedCo...3522474811175/453522474811175/?type=1&theater

https://www.facebook.com/AdvancedCo...2956441867781/450833581746731/?type=1&theater

https://www.facebook.com/AdvancedCo...2956441867781/399361706893919/?type=1&theater

I am not affiliated with ACE in any way.

Here is how a leading aftermarket company develops improved intake manifolds using CAD and CFD (BTW AMS is home of the 2000hp Nissan Skyline so are no amateurs): https://www.youtube.com/watch?v=MU1od1_kx5g#action=share

Dan, a project like this would seem right up your alley after what we've seen you do with cooling mods. You have the design and metal/composite fabrication skills to make it happen.

Anyway I'll leave it here since it seems well thrashed at this point. Those who desire a closer spread with existing manifolds can hopefully apply the same fix as mechanical injection with matched injectors.

I welcome any other EFII users to post their experiences and GAMI spread data here so we can learn what is typical.
 
Last edited:
I've been looking at the intake/sump on my IO-540-D4A5 for the Rocket with a critical eye. From a design standpoint it's a mess, and there seems little hope of feeding 6 cylinders an equal slug of air. I like the tuned tulip induction tubes on the angle valve engines so I have purchased a few of these sumps for experimentation. I intend to install one on my engine and play with runner length to move the torque curve around to better align with my typical cruise profile (altitude and RPM).

I thought I was pretty clever installing a tuned angle valve sump on my parallel engine until I discovered that Robinson helicopters do exactly that. Interestingly, this hybrid engine retains the same 260 HP of the "straight" parallel sump.

Short of a CNC porting and flow balancing the heads, I'd think the use of a Lycoming "tuned" sump would be the most effective method of getting all six cylinders fed the same amount of air.
 
Last edited:
I intend to install one on my engine and play with runner length to move the torque curve around to better align with my typical cruise profile (altitude and RPM).

Keep us informed!!! i would be interested in some data on this
 
EFII GAMI Spread

I'll post more of my thoughts and experiences with the EFII system after I've had a better chance to tweak everything and then get some substantial time behind it. But for now, since folks are obviously interested and I have it handy, here are some data from a low-power flight a couple of weeks ago.

8000'
20.5"
2200 rpm
EGT 4 peaked at 9.5
EGT 2 peaked at 9.1
EGT 6 peaked at 9.0
EGT 1, 3, & 5 peaked at 8.5
GAMI spread 1.0 gal

EGT 4 peaked at 1288 deg and was at 1245 when the last EGTs peaked (43 deg LOP)
My AFR sensor showed between 16.0 and 18.0 from first to last peak, respectively (though since I created the sensor definition, I can't vouch for its absolute accuracy--it seems high).
When the last cylinders peaked (at 8.5 gph) I was doing 154 KTAS.
 
Chris, what was your percent power reading at that setting? I'm guessi about 45-50. That's amazing economy.
 
Dynon was showing 47-50%, I believe. Skyview's lean function was showing GAMI spreads between 0.8 and 1.1 at those low power settings.
 
..here are some data from a low-power flight...

Thank you Chris. Could you (or Jesse) tell us which engine model and intake manifold system is installed on this airplane?

Jesse, was the 1.4 GAMI spread at some higher power setting?

Then for cruise we could set it at 16.5 or so for lop economy at 2250, which would work up to 2375.......For economy, pull back to 2370 or lower and it automatically goes to economy.

Gentlemen, we realize that you're dialing the mixture knob to see GAMI spread, but let's return to a previous question regarding operation based on programmed mixture.

Is the system in fact stable enough that you can pull the prop knob to any RPM below 2370 at any altitude, and have every cylinder end up comfortably LOP, without touching the mixture trim knob? Or are you finding it necessary/prudent to tweak mixture manually?

Chris, what was your percent power reading at that setting? I'm guessi about 45-50. That's amazing economy.

154 KTAS at 8.5 GPH sounds pretty good for an RV-10. However, let's remember that the ability to run very lean is an ignition function. There are several brand choices available in high energy ignition, all of which can be installed with a mechanical injection or a carburetor for pretty much the same result. That includes EFII-branded ignition (SDS ECU + Subaru coil) without the fuel injection option.

If all six cylinders had a tight AF ratio spread, economy would be even better. Instead of some being near the lean limit, and some being just under peak, they would all be near the lean limit.
 
Last edited:
I have been following this thread with interest, as I am really hopeful that a reasonably priced non mixture knob system can be made available. In 1998 I started balancing my injectors so that I could run my Rocket LOP. This was based on the work done by the folks at GAMI. Don Rivera was able to get me some inserts that would allow me to do this rather then use the expensive GAMI nozzles. I am not sure if anyone else had been doing this at that time. Since then I have balanced all my airplanes using this method. They run better at all fuel flow settings if they are balanced as close as you can reasonably get them.

And so after all these years, how do I fly my airplane?

If I am in cruise, at altitude, 65% power or less, I run my engine using the best power Lycoming charts.
If I am trying for an economy cruise for maximum range then I will go LOP. It is more efficient but I do lose at least 5% of my speed, 210 down to 200 or less knots. BTW at 65% power or less going from best power to LOP is about one gallon of fuel flow, which is an easy number to remember for the 540.

But what I do get with that loss of speed, is 5% MORE engine tach hours because I am in the air longer.
That is 5% more of more wear and tear on the engine. If one gets 2000 hours on an engine, then running LOP will meant that you get there 100 hours sooner then if you were running at best power.
I have not calculated what the savings in fuel LOP are vs the extra TBO engine costs because by the time everything is said and done, I don't care, I just want to go faster!
 
Last edited:
I have been following this thread with interest, as I am really hopeful that a reasonably priced non mixture knob system can be made available. In 1998 I started balancing my injectors so that I could run my Rocket LOP. This was based on the work done by the folks at GAMI. Don Rivera was able to get me some inserts that would allow me to do this rather then use the expensive GAMI nozzles. I am not sure if anyone else had been doing this at that time. Since then I have balanced all my airplanes using this method. They run better at all fuel flow settings if they are balanced as close as you can reasonably get them.

And so after all these years, how do I fly my airplane?

If I am in cruise, at altitude, 65% power or less, I run my engine using the best power Lycoming charts. This puts me in the 50 to 100 ROP area, which as I understand it is quite safe.
If I am trying for an economy cruise for maximum range then I will go LOP. It is more efficient but I do lose at least 5% of my speed, 210 down to 200 or less knots. BTW at 65% power or less going from best power to LOP is about one gallon of fuel flow, which is an easy number to remember for the 540.

But what I do get with that loss of speed, is 5% MORE engine tach hours because I am in the air longer.
That is 5% more of more wear and tear on the engine. If one gets 2000 hours on an engine, then running LOP will meant that you get there 100 hours sooner then if you were running at best power.
I have not calculated what the savings in fuel LOP are vs the extra TBO engine costs because by the time everything is said and done, I don't care, I just want to go faster!

Interesting thought about engine life. I run efii dual ignition on my Murphy w the io-540, . I fly much slower than an rv-10, but what I really like about LOP is my CHT temps are 40 degrees lower(360f)when LOP. I would guess for my application on a slow airplane where cooling can be a challenge, this probably more than makes up for the engine tach hours /life. I will use the efii system in my rv8, I am very happy with it.
 
Dan ? I'm sorry that I missed your reply. Thanks for getting back. My injection is mechanical. The figures were for the rich side of peak, probably about 100-150?. Thanks in advance for your thoughts... P.S. I have not been sure if my engine is completely broken in, and I've also been not sure how to lean with the 10:1 compression and what detonation dangers exist. At any rate I've been very tender with my leaning. With previous certified aircraft/engines I was used to operating (the leanest cylinder) at peak EGT (below 75% power) which seemed to give good results. Then learned that the richer cylinders might be operating in a bad area about 40-50 degrees ROP. Now am like the robotic character in "Short Circuit" - I need input... Thanks.
 
Back
Top