Mappable ECU's - The definitive thread

[Redracn]

It is never apples with apples. For instance just entering the same ignition advance numbers into two different ECU even the same brand and model will yeild different advance when measured at high rpm at the crank. 

The reason is delays in the electronic system components. The longer these delays the more the ignition retards as rpm increases. The delay starts with the crank sensor and ends with the voltage rise time of the coil itself and then depends on the arc over voltage at the spark plug which is also variable. The first delay in the ecu hardware is the filter and processing that  and responds to the signal from the crank sensor. The final delay in the ecu is in the output power stage that drives the ignition module. Some ECU have a compensation that can handle the major part of the delay but I am not aware of any that also compensate for temperature changes in the electronic components that also causes some change in ignition timing. Or are individually corrected at manufacture to compensate for the production variability of the components. So if you want the best you must optimise each ECU to the engine and then treat that ecu as if it is a part of the engine.

The tuner is doing what good tuners do. Compromising to deliver a good allround package. But there is no doubt some power (across the rpm range) is left untaped in doing so.

 

to be fair peak power there would be nothing in it, on the old V8's the upstream mixture and slight charge cooling effect leaves not much on the table compared to port injection throwing fuel at the back of the valve.  

 

But why use low mounted injectors?  How about sitting above and pointing straight down the runner? Just because an OEM chooses a compromise mounting position dosent mean you have to use it. Also the back of the valve is very hot and any fuel drops that touch it are vaporized. 

[SimonN]

like any purchase decision, it should be based on requirement.  An aftermarket ECU isn't needed on a road car with minor mods unless changing from carby.

That's a pretty sweeping statement that I cannot agree with, unless you can back it up.

Let me set the scene for my 3.2l. I want more power, say 270 bhp, significantly more torque across the board and better throttle response. How do you propose I get that? My route is to fit decent headers and a Motec. Is that "minor mods"? I don't believe that anybody can get close to that with Motronics without opening the engine up.

Or take the 964 3.6 engine. The "gold standard" for those is usually considered to be decat, cup pipe and hot film, plus live remap  of the Motronics. Most consider that to be "minor mods" and the very best I have seen is 310bhp with reasonable gains across the rev range. Again, with Motec, you can get 325bhp and a significantly better torque curve. I know from personal experience that driving those 2 cars is chalk and cheese. In fact, it's pretty mindblowing. Add a loghtweight flywheel, which many do, and you get even better throttle response from both Motec and Motronics, but the difference in idle is again, stunning. with motronics, it is possible to get it "acceptable". With Motec, you can get it as good as the factory. Again, this is from personal experience. What would you suggest to get the power, if not Motec?

Or is your argument against increasing performance unless you open the engine up?

 

Stock engine management is going to be unlikely to be able to be tuned for anything other thas minor fuel tweaks, and earlier stuff (90's and earlier) probably uses very VERY outdated and limiting technology such as air flow meters (particularly those horrible flappy gate units).  Aftermarket is good here.

I believe that is also not correct. As said above, Motronics can be remapped for MAF. It's basically what Porsche did with the early 993's and it achieves about 25bhp on its own. In terms of pure value for money, it is a great upgrade.

If going aftermarket, list your requirements.  Most people don't need anti lag and all the other fancy hard core stuff. Plenty of cooking level units are affordable and do everything that is needed.

Megasquirt used to be good when it was a kit as you could build it for about $300. Now it's a full deal and costs about the same as an entry level Haltech.

 

 

The differences between, say, Megasquirt and Motec (or other top line ECU's) is a lot more than unwanted features such as anti-lag. The cheap, DIY end of the ECU market has its place, but I personally don't go for it. Simply, they are not as powerful or fine grained. I have been trying to come up with an analogy that works. It's not quite right, but it is like using the standard Apple headphones that come with your iPhone and cost $30 vs some half decent headphones that cost, say, $200. Both play the music, but you can really tell the difference between the sound. I would argue the same is true with ECU's. For those who want an aftermarket ECU and doesn't want to put the time and effort into the sophistication of Motec, and there is some time and effort needed, an easy to use set up like Megasquirt has its place. But there is a very good reason why its at the DIY end of the market and pros do not use it. The results gained will probably be good enough for some or even many with Megasquirt, but, because of the diminishing returns nature of mapping, but for some, its just not going to get you to the point you want and for somebody charging money, the finished result would not be acceptable. DIYers accept compromises that no pro can afford to accept.

To me, if you are going to do something, you might as well do it right. I want to fully understand every perameter and I want to be able to be limited by my ability, not the ability of the gear. If i paint my own car, I expect to get the same finish as if I paid somebody else to do it. If I rebuild my engine, I expect it to be to the same standard as any professional, otherwise it is false economy. So if i am mapping a car, i want the finished results to be the same as if I had paid the top people. I also believe that is a realistic expectation, although I realise it will take me a lot longer and I will need a lot more patience as I learn, experiment and slowly get it right.

For those who want to get their car running OK, who do not have the same attitude as me and infinite patience, there are great options like Megasquirt.

My other concern about some of the DIY stuff is that while there is a great online community and support, if everything goes wrong and you need to hand it over to a professional, you have problems finding people. if i get stuck with the Motec, there are many pros I can turn to who can sort it out.

[SteamWolf]

Did you start this thread just to start an argument and reinforce self perceived superiority?  It sure reads that way.  I am confident that not a single post in this thread will convince you of anything but the position you have already taken.

 

Having said that, I'll address some of the points you have picked up.

"Let me set the scene for my 3.2l. I want more power, say 270 bhp, significantly more torque across the board and better throttle response. How do you propose I get that? My route is to fit decent headers and a Motec. Is that "minor mods"? I don't believe that anybody can get close to that with Motronics without opening the engine up."

Like any purchasing decision, it should be based on requirements. You require 270bhp (or about 200kW in Science).  You chose extractors and an expensive engine management system to achieve it.  So what? That's one approach and you achieved your goal. Extractors are a minor mod. Standalone engine management is not.  I fear this is apples and oranges. Motronic as fitted to 3.2/3.8 is in technology terms as similar to a current generation MoTec (or Haltech or Adaptronic or any of the multitude of other aftermarket systems) as a potato is to a bottle of Grey Goose.  

If you gain massive power increases from an aftermarket engine management system alone, it's unlikely to be because the aftermarket engine management system is so much better than a different aftermarket engine management system. Rather it's because it has just replaced outdated or garbage technology.

MoTec is wonderful stuff.  I have no beef with MoTec. Your fixation with MoTec is the confusing part.  I would suggest though that the difference between a MoTec install and an equivalent, say,  Haltech install on a comparable engine package would not yield significant differences in outright power, power across the curve or driveability.  In a road car you'd never notice. In a racing car the difference may or may not be worth the money spent depending on class of racing, budget and how close to the front you are running

It WOULD however be light years ahead of a stock ECU made 30 years ago!!

"I believe that is also not correct. As said above, Motronics can be remapped for MAF. It's basically what Porsche did with the early 993's and it achieves about 25bhp on its own. In terms of pure value for money, it is a great upgrade."

Don't care.  Very few ECUs are tuneable. Motronics is one. Delco is another. They are not the norm.  Motronics is so damn old as to be a museum piece. At least Delco can be tuned to not use a MAF at all. I'm talking in general terms. You're talking in general terms until you find something you don't like and then you pick a specific strawman

As stated, megasquirt is not what it once was.  The benefits of MS are no longer there IMHO.  There are MANY good options that can deliver the basic features required that MoTec does, for less money and with no discernible difference to power and driveability.  MoTec comes into its own when you get deeper into what their systems are truly capable of, with things like anti lag, launch control, digital dashes, data logging and so on.  Last I checked, a 964 didn't have four variable camshafts.

 

[MFX]

 

"Let me set the scene for my 3.2l. I want more power, say 270 bhp, significantly more torque across the board and better throttle response. How do you propose I get that? My route is to fit decent headers and a Motec. Is that "minor mods"? I don't believe that anybody can get close to that with Motronics without opening the engine up."

Like any purchasing decision, it should be based on requirements. You require 270bhp (or about 200kW in Science).  You chose extractors and an expensive engine management system to achieve it.  So what? That's one approach and you achieved your goal. Extractors are a minor mod. Standalone engine management is not.  I fear this is apples and oranges. Motronic as fitted to 3.2/3.8 is in technology terms as similar to a current generation MoTec (or Haltech or Adaptronic or any of the multitude of other aftermarket systems) as a potato is to a bottle of Grey Goose.  

 

I tend to agree, that MoTec is a fantastic aftermarket computer system, but I think ultimately we should talk about bang for buck. I have no idea what  MoTec costs to set up at the moment (but I remember when I had my Skyline that is was very expensive), and you also need to factor in how long you are going to spend on the dyno getting everything perfect. 

I think it is worth working to a $ per hp. As it always comes down to a budget.

[Redracn]

Simon

You should check out the latest Bosch stuff if you want all the tunability imaginable. 

If you believe one ecu makes more power than another you need to ask WHY. Is it incorrect AFR or ignition not optimal. Then ask what are the reasons for this differance and finaly can that differance be corrected.  Because once correct there should be no differance. 

These days a lot of road ecu are tunable/reflashable.

Take my EVO X for instance. I use the Evoscan with TEPHRA Mods

http://www.tactrix.com/

http://evoscan.com/ecu-mods

It is cheap I can adjust all tables as many or more than a motec and my version also has dual maps. The standard ecu has good knock sensing and calculates fuel octane and interpolates between hi and lo octane mapping. MAF sensor can be removed and a 4bar MAP sensor fitted. So why would you fit an aftermarket ECU unless it offered something that the standard one can not deliver or be modified to deliver. It has OBD logging and a wideband AFR sensor can be fitted. It also can disable a lot of the diagnostics such as the second 02 sensor ECUTEK http://www.ecutek.com/

 

Reflashing is available for most flashable ECU. You will easly find plenty of info on Ford and GM as well as Bosch.

For older cars another route is to use an OEM ecu from a newer car or even a different vendor and reflash it.

Then there is the piggyback industry where an user mappable electronic module intercepts signals and modifies them before sending them to the standard ecu they are cheaper than a full replacement ecu, not much rewiring is required and they are often plug and play. You can also just remove them and run the car as standard. For limited modifications these are a good solution if you can not reflash the standard ecu or have an older ecu as you can replace the vane AFM with a hot film unit (or MAP sensor) but make it look like a vane unit to the ecu.

But older cars have many issues that need to be resolved. Such as flap style flow meters no TP sensor and often rubbish crank signals. Add rubbish generation 1 injectors along with poor fuel rail design with excessive resonance. If you want more power from these early system you first need to sort out all the issues they have. Some was simple cost cutting as well as things that have been learnt over the years and some is technology advancing. 

And the PRO's often compromise the most as they have very specific objectives and place great importance on achieving them. Technology is helping compromise less but as soon as you get into the nitty gritty you will be compromising.

 

 

[Philbee]

Great thread topic - thanks Simon!  I have a lot of interest but little experience so look forward to slurping the accumulated wisdom here

 

ditto from me!! ?

[SteamWolf]

To get this thread back on topic; I firmly believe that while the motornic as fitted to 993 and earlier cars are too damn old to do a lot with.

Having said that, there are a few things that can be done as pointed out by Simon and others.  A change in the intake, exhaust or minor valve lift increases (camshaft replacement with a mild cam) all necessitate tweaking the fuel and ignition maps for both engine longevity and releasing performance gains.  The problem being that the tech is just too old. Relying on a MAF - which by its very nature is a restriction in the intake - won't see the best from an engine.  Nor will the lack of knock sensors or oxygen sensors (depending on what model we are talking about).

What you have to ask yourself is what you want from your car.  Do you have a random power target? Do you simply want it to be nice to drive and start every time no matter the weather?  Do you want maximum performance?

it's also worth noting too much focus is often on peak power.  Peak power is great for winning dyno days or the morning coffee circle-jerk. Optimum engine performance is far more complicated than that.  What we really need to consider is the "area under the curve" of the torque and power curves.  A long delay until the upper end of the rev range before a massive ramp up (ala early 930) to peak power won't get a car around a track as quick as one tuned to have power come on early, smoothly and consistently to a long plateau (ala  993) even if the maximum value is lower than the previous example. Also makes it a complete bastard to drive in traffic compared to a something with smooth delivery of power.

 

Examples: 930 stupidly steep ramp up. This is why lots of people used to go backwards through hedges.

GT4's much smoother ramp

http://cimg0.ibsrv.net/gimg/rennlist.com-vbulletin/800x638/80-f81107da4b9ebb84_4a72404abfdc513677413f476c0455d45e2a7c23.jpg

 

Drag racing, circuit racing, hillclimb, rally and oval track all demand different power and torque delivery characteristics. So do road cars for that matter.

A mappable ECU will allow a tuner to fix some basic corrections after minor mods, but if you start doing significant modification (I'm looking at you, turbo guys) then the stock stuff won't cut it, with only a few exceptions.  Delco has been hacked to death and is one of the few stock ECUs that can do a heck of a lot. Of all the cheap ECU options it's a good one to consider for a retrofit.

Stock ECUs generally won't allow for changes to inputs and outputs. The ECU can control more than fuel and spark. It can be used to switch cooling fans on and off, warning lights, turbo timers, and so on. 

 

 

In short; getting the stock ecu adjusted for minor changes is a good thing to do and achievable but for an engine that is to undergo more than an exhaust and throttle body change I'd suggest going aftermarket ECU in some way.

 

 

 

[Redracn]

Should also raise the issue of legality. As I understand it non compliance with the emissions standards that applied when the car was platted is illegal. (Buy pre 75 for the most fun)

 The authorities tend to turn a blind eye to a car with a standard ecu (assuming everything else looks normal) as they have limited to no ability to tell if it has been re-flashed. On the other hand if they see an aftermarket ecu and you can not show them that it is compliant then you will be wearing a new windscreen sticker.

I know that if you are smart about where you drive and when you would be unlucky to be pulled over by a cop that is having a bad day. But it could happen. In VIC we have had and I have been caught in an every car pulled over blitz attended by the sherif and a team of roadworthy people and just about every other govenment agency. The Cayman had no issues but many were over in the pits looking very unhappy. Not to mention the long line of cars going nowhere except on a tow.

[clutch-monkey]

i hadn't even thought of that, though i would suspect that the authorities managing to find an aftermarket ECU (or even knowing what one looks like) is way, way down the list of things i should be worried about

[Redracn]

Not to hard to find. They just need to plug in an OBD scan tool as no aftermarket ECU I know of responds correctly. This also detects if you have disabled the check engine light.

[MFX]

Not to hard to find. They just need to plug in an OBD scan tool as no aftermarket ECU I know of responds correctly. This also detects if you have disabled the check engine light.

I would think that may be more obvious with possibly newer watercooled cars (which have more recognisable engine management), but when you are talking 20-50 year old cars, I am betting 99% would have no idea. Hell I am a car nut, and I had never even heard of CIS before I started playing around with this.

[clutch-monkey]

Local plod is more content with looking for pod filters and blow off valves and bald tyres, that's about where their knowledge ends. 

[MFX]

Local plod is more content with looking for pod filters and blow off valves and bald tyres, that's about where their knowledge ends. 

Most wont even know which end the engine is  

[SimonN]

The legality issue is an interesting one. The first thing to consider is what are the chances of being caught. Not being from round these parts, I can't know for sure, but with an old Porsche and professional looking install, do the local roadworthy people really know enough to tell that where there should be a Bosch ecu, there is a motec or other. Would they even know what ECU to expect?. Wouldn't they have to take the seat out to confirm? Can, or more importantly, would they do that in a roadside stop? As for the OBD scan tool, do they really have those at roadside checks and are they allowed to plug them in? How far do they go with a well maintained, good looking old Porsche that looks totally normal and roadworthy? I understand what happens with the "rice boys", the street tuned Ford and Holden plus the hot rods, but is there a history of them going after old Porsches?

Also consider cars that are personal imports, because I am not sure there is anything they can do anyway if those have been modified. So long as the key safety features meet the requirements of ADR and it is roadworthy, I think you are OK. I looked at this when coming to Oz, because I was going to have a motec fitted to my RS. I also had no cat and a cup pipe. I was told it was all OK. I also asked the guy who did the plating about the process, and there was no records of the exact spec (such as the decat etc). Therefore, if it was a personal import under the 12 months rule, you could have a motec or similar without any issues. Does anybody know if personal import compliance plates look different from "normal" import compliance plates? I think they are the same. Doe this suggest that those who want to modify their cars should buy an import because they an then simply say "that is how the car arrived in Oz"!

I don't know if I have been lucky or whether it is normal, but in 10 years, I have never been pulled over for compliance checks. RBT'ed 3 times in 10kms on December 30th, but never had the car checked beyond a look at the tyres and lights. Lucky or normal? I had assumed that so long as the car looked right, wasn't too noisy and you weren't hooning, all would be good.

I should add that I know of 2 road cars that have recently been fitted with non standard ECU's, ETB's, specail fuel rails etc. Are they really taking a big risk?

[Redracn]

I can say from experience that when doing a blitz in VIC they bring the EPA and their roadworthy mates. I was crusing mid afternoon on a weekday along the Calder freeway and they were pulling everybody over no exceptions. I doubt you would fool the guys they bring that know what to look for. The average hwy patrol would only only look at the basics like lowered, pod filter, noise etc. I do know P car drivers that have sent for a noise test. 

I doubt they would hassle club cars or pre 75 about an ECU. 

I doubt calling it a personal import would allow a free for all in relation to mods. You would need to enquire further.

Surprisingly the UK is more lenient than Aus. You can even do a limited production run with minimal emission requirements. 

There is a risk and you could get done but it is low. They could crack down at any time and I have been predicting the nannies whould have had a go by now but so far there has been no major push. 

[SimonN]

Did you start this thread just to start an argument and reinforce self perceived superiority?  It sure reads that way.  I am confident that not a single post in this thread will convince you of anything but the position you have already taken.

SteamWolf

Just because I don't agree with you doesn't mean i don't change my view based on what others have said. Look at the issue of the use of MAP vs Alpha-N. 2 of us questioned Redracn's views on this and because he then provided correct technical input to the thread and i was prepared to get in touch with a top tuner, we all gained an understanding of how those 2 are used and I certainly have changed my mind.

The difference is that you come on and pass opinion. You made a bold claim without any evidence, namely that unmodified cars do not need programmable ecu's. You also incorrectly implied that the Motronics were limited to "use air flow meters (particularly those horrible flappy gate units) and that aftermarket is good here" when it is a common and well known to replace the air flow meter with a MAF and modify the chip. In that particular case, it would be a waste of time and money to change to an after market ECU.

Do you expect me to simply say "thanks, great information?" Instead, I asked you to suggest a different approach to the scenarios i presented. I want more power. I don't "need" it, but that is my goal for my car. It is that simple. So with a 3.2 engine, with only decent headers, a well mapped aftermarket ECU (in the examples I know, it was a Motec but I totally accept it could be another make) has been proven to achieve 270bhp, have a significantly improved curve and throttle response, all the things I think you have suggested in your later post. Yes, there are cost implications, but that is also true of any modifications. You say I don't need to go to an aftermarket ECU because my engine is standard. Please advise me how I should reach my goals without going for the aftermarket ECU.

If I have misunderstood what you have said, because it's always hard on forums and all to easy to misunderstand, then I truly am sorry. I might also have been over sensitive because the idea of the thread was to discuss the technical side of ECU's, for those who don't understand or for those who are about to embark on fitting one rather than carbs. On the US forums, these sorts of thread fall apart because of people coming on with zero technical input, expressing opinions and trashing the thread. Maybe my vision for the thread wasn't clearly enough stated, in which case, agan, I am sorry. Hopefully we can move on and discuss the technical side of this area of tuning.

I doubt calling it a personal import would allow a free for all in relation to mods. You would need to enquire further.

I don't know about calling it a personal import after the event, or fitting it and saying it was on the car before it arrived in Oz, but I do know that there is a loophole with the personal imports on emigrating, having owned the car for at least 12 months and driven it. It seems that all mods done, so long as they don't compromise safety, are "grandfathered" in on arrival. So, even if I am not allowed to fit, say. Motec to an Oz delivered 1985 Porsche, if I bring one in under the above mentioned provision, it can have a Motec and be approved. I was surprised when I checked it out, but those are the rules. The extreme example is that a Singer probably cannot be brought into the country, but if you have lived overseas, are emigrating and have owned it for a year, you can bring it in without issue. Sometimes I don't understand the rules over here

[Redracn]

Alpha N mapping. Alpha for throttle blade angle and N for RPM

These days this is the preferred mapping for high performance NA engines especially multi-butterfly. Simon noted that some are also connecting a MAP sensor for air density compensation. This is not required to be connected to the manifold to perform this task as it can be left open to atmosphere for air pressure/altitude correction. And yes you will need altitude correction even in AU. Where the MAP sensor helps in these engines is if it has an air valve that bypasses air around the throttle for idle up when cold or AC activates or other variable engine loads that the driver is not meant to feel or correct.. Every car that does not have an electronic throttle falls into this catagory along with some early versions of electronic throttle. In this case the additional air bypasses the throttle so the ECU sees no change from that sensor so does not add additional fuel to go with the additional air. Having a MAP sensor  connected to the plenum goes a fair way to correcting this. You still get pulsation in the ign and fuel from the MAP sensor but it is a lot less than if MAP was used as the primary load input by itself. This is of course a compromise. If you want the most stable ignition/fuel then you idle up when cold with the right foot. A/C what A/C. Also ignoring Power steering at lock and variable alternator loads like seat heaters etc that need to be compensated for as they are in any modern road car.

[SteamWolf]

I only post on the internet because it seems less crazy than shouting at clouds.

[bumble]

Currently I feel a bit like Dick Cheney - there are things I know;  There are things I don't know;  There are things I know I don't know;  There are shed loads of things I don't know I don't know.  Sooner or later I will know enough to start asking questions and become opinionated

 

Following are comments from a site that I came across during my research regarding Alpha-N.

 

"This is going to be rather brief because Alpha-N is rarely used as the primary strategy for engine management today.  You should at least know what it is because it is often used as a “sanity check” for primary control strategies.

Alpha-N is also sometimes called “TPS maps” because the only sensor that is used for determination of fueling is the Throttle Position Sensor.  (And measured RPM, or how fast the motor is spinning)  Fuel and timing requirements for the engine are expressed as a function of RPM and TPS.

Alpha-N is used most of the time in tricky situations:

1. When the MAP sensor or MAF sensor has failed and the primary control strategy is deemed to be invalid.  Something-is-better-than-nothing is the idea.  (“Load with Failed MAF” is an example from Ford-land)
2. In conjunction with ITBs (Individual Throttle Bodies) due to the extremely low vacuum created by them (making Speed-Density tricky) and the desire to avoid needing to fit a potentially restrictive Mass Air Flow sensor (making MAF impossible).  Again, something-is-better-than-nothing is the idea.
3. In conjunction with ITBs and MAP as a load multiplier. (PowerFC D-Jetro for GTR Skyline, most notable example)  ITBs + Boost – Alpha-N output is multiplied by a MAP sensor to come up with a composite load index.
4. In conjuction with Speed-Density and some kind of blending algorithm.  This approach is often used with very large camshafts that pull little vacuum at idle.  Basically, TPS and MAP are allowed to contribute varying amounts to the overall load calculation.   Net result: more stable and meaningful load index close to idle when MAP sensor readings are unstable.  Found on the Electromotive TEC3 among others.

Alpha-N is very poor at dealing with hills (think about engine load going up and down hills at a constant throttle position), temperature variations and just about anything else that you’d care about except close to wide open throttle where it does fine."

[Redracn]

 

Alpha-N is very poor at dealing with hills (think about engine load going up and down hills at a constant throttle position), temperature variations and just about anything else that you’d care about except close to wide open throttle where it does fine."

Absolute rubbish in relation to hills. So could you please back that up with some theory. Also Inlet air temperature and air density are usually compensated for. For any NA engine that has been fully mapped Alpha N will do an excellent job at all throttle positions and RPM. As I said before its biggest issue is that it does not provide any correction for any air that bypasses the throttle blade. Or where the air pressure on the upstream side of the throttle blade varies and is not compensated for.

MAP,TP,RPM is often used in multi runner turbos particularly ones with individual butterflies that have a dramatic effect on the runner resonance when not fully open. Also used on single and twin cylinder applications.

Road cars these days almost always use MAF due to its simplicity and other benefits. You can even do boost control without a MAP sensor.

So ITBs create a low vacuum.  I assume that you can back that incorrect statement up.

 

 

 

[bumble]

As I've stated before, I don't claim to have any expertise in this area.  I only quoted what someone else published in the spirit of presenting a range of opinions on the topic which, as I read it, was why Simon started this thread in the 1st place.  For me ATM, all opinions carry equal weight.

"So ITBs create a low vacuum.  I assume that you can back that incorrect statement up."  Nope, but if anyone has any theory as to why that statement was made, I'd be interested to see it.

K&N have an interesting page on MAF stuff - http://www.knfilters.com/MAF/MAFDevices.htm

 

[Redracn]

False information propagates when those who know better allow it to  go unchallenged and those who spread it fail to correct it.

Not to bad an overview by K&N. The biggest mistake is to say that the ideal stoichiometric AFR is 14.64:1. It varies and is both fuel and air composition dependent. Humidity being one factor. It is the amount of oxygen in the air that actually matters or oxygen mass flow.

Humidity also plays havoc with MAF and MAP sensors and every other engine load sensing as these are only used to estimate the amount of oxygen being consumed.

In most aftermarket systems the actual air mass flow is not calculated. The ECU simply uses a few tables to model the entire engine and goes straight from the input values through one or more tables straight to an injector pulse width. 

Another problem is that injectors inject volume but it is a mass that is required. For a constant fuel temperature this is ok but the trouble is that fuel temperature varies a lot and it is not uncommon to have over 50C in a tank that is close to empty. As its temperature increases fuel expands and its density decreases resulting in less fuel mass being delivered by the injector unless compensated for. 

[bumble]

I would have thought that every issue raised could be measured and compensated for?  The effects of humidity and fuel temperature on air density and fuel volume are not exactly a mystery.

Which brings me to a casual observation.  From what little is publicly available from the ECU guys, it seem that even the best processors in use are relatively slow by today's standards.  Some boast of processors running at 120Mhz!  What?  An average phone is 10x faster than that.  Looking at some of the processors in use (some of which I've worked with) indicates nothing special.  I can understand that when ECUs were being developed, it made sense that not everything could be calculated fast enough and it was necessary to directly access parameters stored in a matrix to augment some basic measurements.  As time went on, the matrix grew, more calculations were done and so on.  But today?  An engine, even at 10,000rpm is relatively a sloth compared to a cheetah in electronic terms.  I am moved to question the need for parameter matrix (map) at all.  But Motec et al are not idiots so there's obviously something behind that big stand of trees I can't yet see.  Or maybe it's just inertia - there's so much invested in current knowledge and techniques that it's too hard or expensive to change?

Another puzzle is why, some cells in the fuel map are much richer than stoich? 

[Redracn]

Yes you can measure pretty much everything but at a price. Given closed loop self learning based on the Lambda sensor it is not necessay to do so as most errors can be trimmed out including manufacturing variance in all electronic components.

For those with older cars with no lambda sensor or aftermarket where one is not fitted humidity and fuel temp can be an issue. This is one reason to ensure a tune has a safety margin in it. If you run at the limit and want max power/torque the small items start to matter.

Why some operating areas of an engine are richer than stoic is no mystery. If you want max power you will need to be at lambda around 0.9 but this is engine dependant. If it is turboed you could need to go as rich as  0.7 or even more as you need to tune the engine so that it will survive. Adding fuel to reduce the chance of detonation and keep the turbo and exhaust valves cool enough to survive. For max economy you would run a bit learner than stoic.

The only reason to run stoic is for emmisions

The processors used are specificly designed for the task with powerful timing co processors and other features. They have plenty of computing power to do the maths including 32bit floating point processors and look up tables as long as the programmer is not lazy (which many are these days)  In fact floating point is not required. 16 bit integer storage with 32 bit intermediate values is more than enough. You will also have plenty left over for some comms and logging. 

There is a differance between processors designed for consumers and those designed for industrial or automotive use.  How many of the fast consumer processors are specified to operate at 125C but this is off topic.

 

 

[SimonN]

Wow. A couple of days away from this thread to get some paid work done and it's interesting to see what i have come back to.

The comments about Alpha-N are at best misleading and at worst, simply wrong. I would love to know what site it is on. I am going to guess it is in some way connected to a particular engine management system, because it seems to talk about limitations that not all have.

In addition to the points Redracn has made, i would point out the following

Alpha-N is also sometimes called “TPS maps” because the only sensor that is used for determination of fueling is the Throttle Position Sensor.  (And measured RPM, or how fast the motor is spinning)  Fuel and timing requirements for the engine are expressed as a function of RPM and TPS.

 

 I believe this is incorrect. I am researching motec, because that is the easiest for me, so please excuse me using that as an example. I also don't know the exact way the algorithms work behind the tables, but when you use MAP with Alpha-N, you set up what is called "MAP fuel correction". As such, the TPS is not the only sensor determining fueling. MAP can also be used as a secondary sensor for other things, such as altitude and natural air pressure changes, when combined with the air temperature sensor.

Something I found out this week is that Singer use motec with Alpha-N plus MAP for their top power engines, which were (and maybe still are) designed, built and tuned by Colin Belton of Ninemeister. As Redracn has educated us, there are other ways, but, IMO, it says something about that way of tuning. Having said that, when I start to play, I will probably just go with Alpha-N until I feel that I cannot get the car to run the way I want, such as with idle. 

I have been trying to find the time to write something on AFM vs MAF vs MAP, because I believe there are a number if people who would like a basic overview of them to better understand some of the above. In crude terms, they are all ways of measuring the air going into the engine. The air flow meter operates with a flap which opens in response to the amount of air passing it. This is 'bad" for 2 reasons. First, it is a restriction for the airflow and second, it creates unwanted turbulence (swirl). however, don't panic. It was used very successfully for about 10 years and does no harm, other than limit power. Porsche went from there to a MAF (Mass Air Flow) sensor, often called a "ht film" sensor. This basically is a sensor that is stuck into the airflow. It has a wire that is heated (hence "hot wire") and the change in resistance due to it being cooled by the incoming air is measured and this can be used to determine how much air is passing it. Because of the way it sticks into the airstream, it disturbs the air far less. It is estimated that this change alone is worth something like 20bhp on a 964 engine. Finally, the MAP (manifold absolute pressure) does what its name suggests and measures pressure in the manifold which is used to calculate density and air flow. 

While there is a big difference between how much the airflow is disturbed with the AFM and MAF, there isn't a big difference between MAF and MAP, although there is a gain.

I hope that helps people and  doesn't, in itself, cause too much debate (I know what they are, but writing it clearly might have been beyond me!!)