Quantcast

Forum made EFI?

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
1,251
Location
Jackson
Which Honda engines?

As to auto-correction, please see my previous post, and explain how the ECU can tell what mixture I need, for the various flight conditions, all of which are 21/21 (as just one example).
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,757
Location
US
Using the OEM ECU requires a full exploration of the failure modes. The power produced in factory "limp home" mode may not be sufficient for safe flight. Descending into the woods just because the MAP sensor went bad isn't a great outcome. Also, most OEM systems now use a MAF sensor which frequently depends on a large intake assembly for proper control of air turbulence through the sensor. It can be hard to find enough real estate for this under a typical aircraft cowl, though sometimes it is possible.
Mixtures appropriate to meet EPA emissions or MPG targets aren't necessarily appropriate for all flight conditions, so a reflash of the OEM system would be necessary, at a minimum.

Folks aren't eschewing the automotive ECUs out of ignorance or a desire to spend more time and money.
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,492
Location
Rocky Mountains
and explain how the ECU can tell what mixture I need
That is kind of the point of all of this.

From an engineers point of view we don't care what you 'think' you need/want. We are replacing you with a system that 'actually' knows what is needed.
From a pilot's perspective we need/want a way to take control if the system finds a set of conditions that the programmer/engineer didn't anticipate of plan for well enough (737 Max)

Somewhere in between there is a sweet spot that reduces overall death and destruction with a good balance of eliminating operator error without introducing a significant new hazard.
 

Toobuilder

Well-Known Member
HBA Supporter
Log Member
Joined
Jan 19, 2010
Messages
5,037
Location
Mojave, Ca
...From an engineers point of view we don't care what you 'think' you need/want. We are replacing you with a system that 'actually' knows what is needed...
Not sure what you are trying to say here so lets map it out with a real world scenario:

Im cruising along at 8500 feet on my way to Phoenix. As usual, the throttle is wide open against the stop, I have the prop set to 2300 RPM and Im LOP. At this DA and TAS, spinning the prop faster only makes more noise, so my primary interfaces with the powerplant are already maxed out. My buddy calls and says there's a Haboob (dust storm) approaching my destination airport from Tucson. I need to go faster to try beat the storm so I need to go ROP on the mixture to get best power.

How does your ECU "know" whats needed to go faster?
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,492
Location
Rocky Mountains
Push the little detent on the side of the throttle and advance it to the balls to the wall mode?

Not trying to be cute, but I really don't understand why the software can't figure out a situation like this. Other than when the system is told that max performance is needed it should be operating in a mode to maximize effeciency and engine life at whatever rpm/manifold pressure is chosen.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
1,251
Location
Jackson
Toobuilder clearly stated the operating condition; it is *already* 'ball (meaning throttle; no 's',or plural, adding mixture & prop) to the wall'. Advancing the prop will not help to increase speed, as stated. With manual control over mixture, fuel flow has been reduced to lean of peak, which reduces power produced. As Toobuilder implies, moving the mixture forward will increase HP, increasing speed. (Remember, throttle is already wide open.) Taking the mixture control away from the pilot means he cannot choose operating mode.

While Toobuilder has a constant speed prop, and therefore, independent control over rpm, the same principle applies even with fixed pitch props, to a lesser degree, as I have stated repeatedly.

Charlie
(Toobuilder, sorry about speaking for you...)
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,492
Location
Rocky Mountains
clearly stated the operating condition; it is *already* 'ball (meaning throttle; no 's',or plural, adding mixture & prop) to the wall'.
Maybe that is the root of our 'disagreement". To me the condition stated is NOT balls to the wall because there is still some power left in the system (the other ball) that is not being used. That is by the pilot's choice and it matters not if that choice is carried out by levers and knobs or a cpu and software.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
1,251
Location
Jackson
But with the cpu having control, there is no mixture knob. That's the point; there is no possibility to operate the a/c the way Toobuilder describes, if the cpu controls replaces the red knob.
 

blane.c

Well-Known Member
HBA Supporter
Joined
Jun 27, 2015
Messages
4,492
Location
capital district NY
The thing with closed loop is that you don't know if O2 sensor data is valid without another one to compare to continuously. If you start erroneously trimming lean on some fragile engines, you'll melt the pistons. Need more code to compare sensors and then more still to decide how far to go before disregarding and jumping back into open loop.

Using EGT for closed loop is possible but don't forget that EGT cools on either side of stoichiometric so you need to know if you're ROP or LOP. You have to start from the rich side to be safe and trim leaner, register peak EGT in each cylinder and then trim to the desired temperature. Not all cylinders will peak at the same time. A lot of complexity for simple engines which spend most of their time at constant power.

There are lots of things you can do, but should you? Is there real benefit in the end for the user or are you adding complexity and impacting reliability with these features?
It is standard practice on the Pratts to enrichen and re-lean every hour or 1,000 foot of altitude change. It is good to start at a baseline periodically with them.
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,757
Location
US
Luckily, engines will keep running and making power over a fairly wide AF ratio. Open loop EFI is capable of reliably staying on the rich side of the curve and manually leaning as far as we want isn't very difficult. That's the simple way to avoid detonation, etc. It does require that we pay attention if we advance the throttle again.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
1,251
Location
Jackson
Blane,

The whole point is that you'll be *unable to do that*, if the cpu controls mixture.

More generally, air cooling is one big factor making 'full authority' very difficult. A friend once set up a rotary using a Megasquirt controller, and programmed it so that it was operating at max lean until the throttle butterfly was wide open. The throttle *control* was configured to reach full butterfly opening at about 1/2 the total throw, and the rest of the throw began to richen the mixture to just past best power fuel-air ratio at the end of the control's throw. This can work with a normally aspirated rotary, which can operate safely over an incredibly wide mixture range. But if you try it with an air cooled engine, you will be buying another engine really soon. And please note that even his technique keeps mixture under the control of the pilot.

CPU controlled open loop EFI is great; I'm installing one on my rotary. But open loop requires manual mixture (unless you can afford to waste 20% of your fuel). My recent responses are to post #352 & following, which discuss *closed* loop in a/c CPUs

Charlie
 

Hot Wings

Grumpy Cynic
HBA Supporter
Log Member
Joined
Nov 14, 2009
Messages
7,492
Location
Rocky Mountains
But with the cpu << snip>>
I think this is one of those time where trying to hash this out via keyboard, when an in person meet with hand waving and napkins would have things resolved in less time than it takes to order another beer.
Probably time to move on..............
 

gtae07

Well-Known Member
Joined
Dec 13, 2012
Messages
2,077
Location
Savannah, Georgia
I guess part of our confusion here is terminology... are we talking something like SDS on a Lycoming, where the prop and throttle are still direct mechanical linkages? Or a true single-lever FADEC ("throttle by wire") where the pilot commands power/thrust output instead of changing airflow? The latter is more complex to implement but obviates the "how do I tell it what my flight condition is?" question because it doesn't care; all it wants to do is give the commanded power/thrust with minumum fuel burn.
 

rv7charlie

Well-Known Member
Joined
Nov 17, 2014
Messages
1,251
Location
Jackson
This:
"Some sort of closed loop will be optimal for Aircraft.. Just add some limits for results for an O2 reading that is out of range . Auto cruise settings, full power settings, all adjusted for actual density. I see no reason to limit the ECU and have manual input while flying for lean. The AFR chart does that all by it self if you build it correctly."

Is what I've been responding to. The AFR chart (actually, the cpu) cannot directly handle all operating modes, as Toobuilder and I have been saying. Flying under the various conditions, especially with a constant speed prop, will demonstrate itself to you.

Charlie
 

jbiplane

Well-Known Member
Joined
Jan 29, 2014
Messages
276
Location
Krasnoyarsk, Russia
Exist self tuning chinese FAI EFI which completelly autotune by narrowband O2 and cylinder themperature without any user interaction. Easy to install by non ecperienced user on 4stroke 50...250cc engines. It is lean on middle and reach when engine overheat. Throttle bodies range 22,24,26,28,30mm corresponding 7...24hp range per intake. They use combined pump-injector which greatly simplify system compare classical. I got few of this set and change ECU and wiring on mine. But i have not resources for this project and it is goingto be Msc project of one of local student, cannot be sure if he sucseed.
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,757
Location
US
Exist self tuning chinese FAI EFI which completelly autotune by narrowband O2 and cylinder themperature without any user interaction.
If it is using a narrowband O2 sensor then it is probably operating very near stoichiometric mixture. I don't think that's going to work in airplanes, it results in max power about 10% less than would otherwise be available at WOT with a richer mixture.
With manual mixture control (ideally for each cylinder, or else a set trim so the jugs are getting the mixture desired by the pilot) we can get better maximum power at WOT and better cruise flight fuel economy than it is possible to get at stoichiometric AF ratio.
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,757
Location
US
I remember reading about those combined injectors/pumps a while back. The "regular" pumps and injectors are reliable and a known quantity. The combined unit might be okay for acft use if they are similarly reliable or so cheap that redundancy is affordable. If I recall correctly, they have a small electric draw (good for V-twin installations with limited amps available). Also, no high pressure fuel lines and no return lines required. I wonder if vapor lock would be an issue, seems like it could be.
 

jbiplane

Well-Known Member
Joined
Jan 29, 2014
Messages
276
Location
Krasnoyarsk, Russia
vapor lock would be an issue
They resolve this problem

For classical - I got pulmsized brushless pump assemblies which consume about 12w (2...3 time less than typical moto) much more reliable with weight ~300 grammes (Yanaha MIO R3 125cc) and good and fast fuel injectors 28...240cc per minute
brushl.jpgme.jpg
 
Last edited:

Toobuilder

Well-Known Member
HBA Supporter
Log Member
Joined
Jan 19, 2010
Messages
5,037
Location
Mojave, Ca
Maybe that is the root of our 'disagreement". To me the condition stated is NOT balls to the wall because there is still some power left in the system (the other ball) that is not being used. That is by the pilot's choice and it matters not if that choice is carried out by levers and knobs or a cpu and software.
Modern cars go rich/lean based on throttle position as a primary system logic. Naturally aspirated airplanes are typically flown at WOT in cruise to achieve the highest MP possible so throttle position is not a valid logic signal for the ECU to "know" what the pilot wants. In this case, the pilot has set the throttle full open to minimize pumping losses - that's the only message the pilot is sending to the CPU. And in many cases (high TAS, high altitude), maximum RPM on the prop will actually REDUCE airspeed - so the blue knob is also a poor logic source. So we are left with power being influenced directly by mixture, and that manifests as a discrete ROP/LOP selection.

In my airplane the decision is eased significantly by the LOP function of the SDS ECU (flipping a switch cuts fuel flow and advances timing), but it remains a discrete pilot action to enter that mode. The switch could certainly be tied to a detent on the throttle, but activation is still using the biological ECU (pilot) as the logic signal.
 
Top