Why must EFI be reprogrammed for elevation?

[LovingOffroadPain]

I'm not a mechanic, but I've read posts on TTalk that discuss the difference between a regular car/street bike EFI vs dirt bike EFI.  Why must we program EFI for elevation change on a dirt bike but not for a car? 

 

I can't find the thread with this info.  Something to do with barometer/oxygen sensors??

[zotis777]

I'm not good with electronics but it might have something to do with the fact cars have maf and o2 sensors, so its always reading how much air is goin into and leaving the engine, so if the car is getting less air, the MAF tells it so so it squirts less fuel

[ScootMcGoot]

The reason is that on dirt bikes they use open loop EFI.  A car or modern street bike uses a closed loop EFI system.  The primary difference between these two types of systems is that a closed loop will have O2 sensors in the exhaust stream to analyze what is coming out the other end.  The sensors in the air box which measure at the very least the air temperature is the only environmental input variable to the EFI map.  This will get the a/f mixture pretty close but in order for the system to be able to self adjust you need to be able to collect data after combustion.  With an open loop system there is no data collected after combustion so fine adjustments cannot be made by the ECU.

[LovingOffroadPain]

The reason is that on dirt bikes they use open loop EFI.  A car or modern street bike uses a closed loop EFI system.  The primary difference between these two types of systems is that a closed loop will have O2 sensors in the exhaust stream to analyze what is coming out the other end.  The sensors in the air box which measure at the very least the air temperature is the only environmental input variable to the EFI map.  This will get the a/f mixture pretty close but in order for the system to be able to self adjust you need to be able to collect data after combustion.  With an open loop system there is no data collected after combustion so fine adjustments cannot be made by the ECU.

 

I wonder how much complexity and weight would be added for a closed loop system on dirt bikes.  Before I ever consider a four stroke, how easy or difficult is the computer that adjusts the EFI for elevation??  Is it a simple push if a button, or is it a diffucult guessing game like jetting your bike for elevation?

[ScootMcGoot]

I wonder how much complexity and weight would be added for a closed loop system on dirt bikes.  Before I ever consider a four stroke, how easy or difficult is the computer that adjusts the EFI for elevation??  Is it a simple push if a button, or is it a diffucult guessing game like jetting your bike for elevation?

 

I don't have an EFI bike so I have not spent much time researching but I would be surprised if most EFI bikes did not have 2 sensors in the intake tract.

 

1. An air temperature sensor

2. A barometric pressure sensor (for elevation)

 

I want to say KTM does have both of these sensors in the air box but I could be wrong.  If that's the case the bike WILL adjust for temperature AND altitude by using those two input variables with the pre-programmed map.  Obviously my original statement still applies in that fine tune adjustments are not possible due to the fact that it is still open loop.  My guess is that they have not put closed loop systems on dirt bikes yet because of the cost and power requirements.  Many of the EFI bikes don't have batteries so the stator must be able to generate enough power when you kick it to light up the ECU and the fuel pump.  You would also need an exhaust pipe with an O2 sensor bung on it as well.

[SnowMule]

They don't have to be "reprogrammed for elevation".  There's an absolute pressure sensor (barometer) that measures ambient pressure, and that input is used in the fuel map along with temperature, throttle position, load, etc. 

 

The computer adjusts all that automagically. 

 

They aren't a closed-loop system, so there's no feedback about performance/efficiency/emissions like a car has (o2 sensors and EGT's in exhaust, for example). 

 

To do a closed-loop system would be possible, and include a lot of math; but for the most part the framework is there.  Adds a lot of parts/complexity.

 

Sleds have had EFI for about 10 years now, and the only thing you have to change over elevation with a FI'd sled is the clutching.

[William1]

The range of 'automatic adjustment' is very small on dirtbikes. The ECU does not calculate and output values for the injector. It reads the data it collects and applies it to a map. If the data it gets is outside its' range (like from big temp swings or large changes in barometric pressure-It does not take humidity into account at all) it defgaults to a limp map, using just throttle position and rpm.

If it actually calculated the values, the ecu would be more power hungry. Add a few more sensors, a 30 watt wideband O2 sensor, then a larger stator and bigger flywheel...

 

EFI as on the current crop of MX bikes is as sensitive to air density changes as the old FCR. It is easier to adjust (if you have the software) and more accurate (no 'transition' points like with jets) and easiy to add a little fuel on decel to quench the popping.

[rdlsz24]

Same reason it won't re-learn when you add intake and exhaust

 

I had a 98 Arctic Cat 600 sled that was EFI way back and was scared to modify it because you actually had to swap out the chip on those early ones. I was too scared of melting it down.

 

Rob

Edited September 27, 2013 by 03YZ85

[adam8781]

i have no idea why they have not designed afr sensors into the system and allow the ecu to make fueling changes, it really would not be that hard.

 

pipes, cams, different fuels, compression ratios, would all fall into line.

Edited September 28, 2013 by adam8781

[SnowMule]

It's not that it's hard, it's a cost issue.

 

AFR gauges aren't cheap.  And especially in a 2-stroke, they foul out easy. 

 

EGTs give you some indication of fuel ratio, almost a requirement once you start modding sled engines, a hell of a lot cheaper, but not quite as "honest" as an AFR is.

[adam728]

As said, cost. A performance mx bike isnt going to run at stoich in most areas, and a wideband ads a ton of cost to the system. Plus power needs.

Ad for the c omputer and calculation limitations due to power consumption, doubful. Not like adding 25% fuel vs 15% fuel changes the power draw any real amount.

[87XR250R]

Dirt bikes are lacking MAP, O2, MAF, and charge air temp sensors. Like it has been said the run in open loop. Very unlike a car. It is essentially a computer controlled carb. But not really.

[ScootMcGoot]

As said, cost. A performance mx bike isnt going to run at stoich in most areas, and a wideband ads a ton of cost to the system. Plus power needs.

Ad for the c omputer and calculation limitations due to power consumption, doubful. Not like adding 25% fuel vs 15% fuel changes the power draw any real amount.

It's more about the compute resources needed to make those types of real time calculations by the ECU that would be a power drag. I don't know how many times per second a typical closed look EFI systems polls the associated sensors for data but I would have to guess it would be in the hundreds if not thousands of times per second.

[Chuck.]

Even a closed loop system on a car returns to open loop anytime the throttle position changes.  So consider how many times during a lap  you change throttle positions before wishing the complexity of a closed loop system.  Also Diesel electronic injection engines don't have closed loop systems and they perform well in a variety of conditions.

 

Any system that measures intake temp and absolute pressure, if properly programmed, will do a good job of adjusting AF ratios  for a wide range of conditions (I have  a 11 year old CAT that runs like a modern EFI gas engine).  Just saying that it all depends on the details of the system.

 

I have a friend who owns a Montesa 4RT; it is a Honda 4T Unicam engine with a no battery EFI system.  Always seems to run the same; any engine temp, ambient temp,  or altitude.

Edited October 3, 2013 by Chuck.

[yz250f_spud]

Also Diesel electronic injection engines don't have closed loop systems and they perform well in a variety of conditions.

 

Eh, not quite.  The EPA/EU has almost indirectly made this a requirement, at least for certain actuators.  The closed loop systems are mainly monitoring NOx emissions, while AFR runs open loop.  And diesel engines don't so much care about AF ratio, as long as it's above a certain level to keep from being incredibly smokey making it fairly easy for them to run in a variety of conditions.  Two different systems.  

Edited October 3, 2013 by yz250f_spud

[Dragon67]

I seem to remember closed loop systems on some calibration modules I was looking at for the KXF before I remembered we weren't qualifying for Loretta's and I need to retire someday. They had a sensor that tapped into the exhaust. In any event, I was an FI skeptic before the KXF. Have to say it has been almost disappointingly trouble free. Nothing to tinker with or dial in. I did put in a second injector and swapped the ECU and harness, but it only ran better. Pretty cool stuff that FI, even if it is kind of primitive. Did not know I would have to reprogram for elevation. Maybe something to look forward to.

[William1]

It is not elevation really, it is air density, a total package made of temp/air pressure/humidity. How much air is in the box at a given point in time.

Air pressure must be taken in a location that does not get affected by ambient conditions (breeze/vehicle movement for example). Temperature of the airflow and unaffected by ambient engine temperatures. Humidity too. You take those values, and calculate (based on RPM/throttle position/relative change in throttle position (speeding up/slowing down/steady state) thrown in data regarding flow characteristic of the engine, mix in ignition advance and culminate with the reading from the wideband. Then come up with a spray time for the injector. 13,000 a minute. Remember, one spray is when the intake valve is closed, the next, when it is open, so fudging (this was done on 2S back in the day before reads and the engine often sucked and blew the same charge back and forth through the carb) needs to be factored in.

[adam728]

It's more about the compute resources needed to make those types of real time calculations by the ECU that would be a power drag. I don't know how many times per second a typical closed look EFI systems polls the associated sensors for data but I would have to guess it would be in the hundreds if not thousands of times per second.

Once per engine cycle is typically good enough. Under some stratagies/conditions maybe even less than that. On a 4 stroke at 14,000 rpm that's about 116 times per second.

 

Even a closed loop system on a car returns to open loop anytime the throttle position changes.  So consider how many times during a lap  you change throttle positions before wishing the complexity of a closed loop system.  Also Diesel electronic injection engines don't have closed loop systems and they perform well in a variety of conditions.

 

Any system that measures intake temp and absolute pressure, if properly programmed, will do a good job of adjusting AF ratios  for a wide range of conditions (I have  a 11 year old CAT that runs like a modern EFI gas engine).  Just saying that it all depends on the details of the system.

 

I have a friend who owns a Montesa 4RT; it is a Honda 4T Unicam engine with a no battery EFI system.  Always seems to run the same; any engine temp, ambient temp,  or altitude.

No, they don't. Closed loop can be ran under an amazing amount of conditions, including throttle transitions and even WOT. Of course, application, engine design, etc factor into this heavily.

 

Otherwise I will agree with you, open loop can run really well if done right. Seat-of-the-pants there's a lot of wiggle room for AFR, you might not be able to tell the difference between 11.5:1 and 13:1. But emissions can, that's the biggest reason for needing closed loop in other applications. Also, keeping in mind this is for MX bikes, running at stoich is going to be rarely the best in most situations, so closed loop with a narrow band O2 won't bring much improvement. Doing it with a wideband O2 would be possible, but again, cost is going to be much more. And in the end the rider likely won't feel any difference.  It's one of those things that's "good enough" for most end users now, and the factories aren't going to push added cost and complexity unless something like emissions regulations force them to.

 

It is not elevation really, it is air density, a total package made of temp/air pressure/humidity. How much air is in the box at a given point in time.

Air pressure must be taken in a location that does not get affected by ambient conditions (breeze/vehicle movement for example). Temperature of the airflow and unaffected by ambient engine temperatures. Humidity too. You take those values, and calculate (based on RPM/throttle position/relative change in throttle position (speeding up/slowing down/steady state) thrown in data regarding flow characteristic of the engine, mix in ignition advance and culminate with the reading from the wideband. Then come up with a spray time for the injector. 13,000 a minute. Remember, one spray is when the intake valve is closed, the next, when it is open, so fudging (this was done on 2S back in the day before reads and the engine often sucked and blew the same charge back and forth through the carb) needs to be factored in.

The system wouldn't need to be setup to inject every revolution. Injecting just on the intake stroke, and timing it before/during the intake valve opening is possible.

[William1]

 

 

...

The system wouldn't need to be setup to inject every revolution. Injecting just on the intake stroke, and timing it before/during the intake valve opening is possible.

It does not have to but the engines do not currently run a cam position sensor and therefore have no idea if it is on a intake stroke or a power stroke.