Dumbed down MPFI talk

[S2X01]

Okay....so I just read through 3 pages of EFI talk that made my brain mad at me. A lot of it made sense...but I still feel like I have basic questions unanswered.

Let's say I have a 455 or a 425/401 that I wanna play with. Let's also say that fabrication is one of my stronger points.
Drilling and welding in bungs for injectors and rails....no problem.
Converting a carm to a throttle body.....I think I can do it. http://www.mez.co.uk/ms2.html

Now the electronics part is where my questions remain, and I am looking for some pretty simple answers. I've already read every word of this... http://www.v8buick.com/showthread.php?241628-Which-Fuel-Injection

Now....say the fabrication part is all done and I'm ready to do some wiring. Can I use a harness and computer from a MPFI V8(say a 454 just for discussion sake) and reprogram it to work for my application?

Second part of the question....is it worth it as far as fuel efficency gains? My next project is gonna be more "cruiser" based. I still want to be able to smoke the tires here and there, but I'd also like to drive it long distances. If this much work and effort is only gonna get me +2 MPG.....not worth my time for sure.

At that point....I begin wondering if a LT1 from a donor roadmaster would be worth the effort.

My ultimate goal is to take a 50-53 Super or Rivi, and make something that can both smoke some rubber when I want, but also not cost me a small fortune in gas. I already have this '72 455 in my garage that I'm willing to play with.... and I would definitely love to make her injected. But as I stated before....if the gains aren't substantial...I'm not sure it's worth my time.......

Discuss!

[bammax]

You want multipoint injection to see the biggest increase in mileage and power. That means the throttle body will just be used for air and the fuel lines will go to the fuel rails via a regulator.

The elctronics are the hardest part since your engine doesn't have a cam or crank sensor. The other sensors that you need can be found at the local junkyard and wired in easy enough to whatever ecu you want to use. The cam sensor can be snatched from a later 3.8 that has the distributor mounted cam sensor. Or you can use the distributor from a later vortec van and have it grafted to your Buick distributor shaft which will allow you to have the right sensor specs to run a vortec ecu. The crank sensor is all custom and that's the hard part. If you check someone made up a custom external one and it seems to be working. I know there's also a company that makes a factory style one that gets retrofitted to chevy motors to run the vortec ecu. I asked about a custom install and they said it was a matter of having enough clearence between the crank and the timing cover to fit the sensor wheel.

[S2X01]

Wouldn't a points eliminator kit help serve as an optical type crank sensor in theory?

http://www.youtube.com/watch?v=RuIislTGOwA

Since it's already optically sensing engine rpm, just not timing. Running a carb means that timing has to be set manually anyways. I can handle that.

If not....would a hall effect sensor mounted at the flywheel suffice? seems like in that case, a little fab work could have a mount on the bellhousing or even that dust cover.

[S2X01]

http://i8.photobucket.com/albums/a50/dezurtrat/Xterra/Crankshaft_Position_Sensor.jpg

According to this....on my Xterra the crank sensor is for diagnosis only and doesn't affect how things are running.

[TheSilverBuick]

Using a stock ECM from the junkyard will disappoint you. 90% of the people I hear complaining that EFI didn't work well for them was from using a stock or re-flashed ECM from a different engine. It's more of a headache than shortcut. I recommend going with an aftermarket ECM you have full control over. You've read my post in the other thread so no point in me re-hashing it.

You DO NOT need a fancy distributor or crank trigger to get most these to work. A simple wire from the points works just fine, and if you use the ECM to trigger the coil (or MSD box, etc) then there is almost no wear on the points contact because very little current will go through it to signal the ECM. If you get an HEI distributor you can run some/most ECM tach feeds right from the pickup coil. Or you can do like my TBI'd '67 T-bird and leave the distributor alone and just run a wire from the negative side of the coil and call it done (that won't control ignition, but is fine for fuel control).

If you have any sorts of carbueretor tuning skills, particularly with a Q-jet, then you won't gain much in the mpg department if your carb is tuned well. And overdrive transmission will net the most gains for mpg. Figure 5%-10% mpg gains with the EFI if the carb tune is good, 25+% if not tuned so well (Because it's a Holley =P and assuming you can tune the EFI, if you're good with a carb, you'll find the EFI tuning intuitive IMO). Overdrive transmissions are almost always good for at least a 20%-30% increase in mpgs right off the top.

That's my two cents.

[TheSilverBuick]

Wouldn't a points eliminator kit help serve as an optical type crank sensor in theory?

Yes, or you can stay with points, see my post above.


Since it's already optically sensing engine rpm, just not timing. Running a carb means that timing has to be set manually anyways. I can handle that.
Yup.

If not....would a hall effect sensor mounted at the flywheel suffice? seems like in that case, a little fab work could have a mount on the bellhousing or even that dust cover.

Flywheels from my understanding make bad trigger wheels for a couple reasons. One the tooth count is so high that few ECM's can be programmed to count that many teeth for a single revolution. Another is a missing tooth has to signal when a full revolution occurs, and the starter might not like that. Also the teeth are pretty narrow at the top, some hall sensors may not pick up the teeth that well. Then if a starter hiccup happens and the teeth get damaged then that has to be delt with and may lead to false signals if the above issues are already addressed. I think Jeep did this in the late-80's and then got away from it for some reason or another.



Answers in red.

[S2X01]

So basically with a well tuned carb, I could just beef up an OD trans and get similar results with a whole lot less work? (and of course rear end gearing)

Now what about this option vs the LT1 option? The car I'm looking at right now has no engine in it anyways, and the plan is a full front subframe swap, rear 4 link and lots of custom fabbing. So...at this point I'm trying to narrow down engine options. If an LT1 is a reasonable option, I might not be opposed to a full mid-90s Roadmaster frame swap.

[TheSilverBuick]

LT1's are good engines once the Opti-Spark upgrade,to keep water out, has been made. Reliable and make decent power right out of the box.

Quick starting and MPG gains are bad reasons for switching to EFI if they are the only reasons because they definitely don't justify the cost. My motivation was operating conditions. I start and drive my car in temps ranging from -20*F to 120*F, and in elevations ranging from 8,000+ feet to sea level. I have a deep overdrive transmission and wasn't happy with the mechanical/vacuum timing settings across the rpm range, and either was setting it for WOT or for cruise, but not quite ideally where I wanted the timing for both at the same time with the low rpm high load cruising (finding the appropriate vacuum can rate and total is a pain these days). Plus I can start a high compression engine with 5* of timing (spins like the ignition is off) and instantly get 20*+ of timing when idling for a smooth idle.

[S2X01]

This car, once completed, could very well see altitude changes from 10,000+ all the way down to sea level potentially. Ideally, it would take me from Denver to Flint. So that's why FI became a thought. There's no way a ballsy 455 is gonna be an economical means of conveyence to the mitten. The LT1 popped into my head, just based on the roadmaster's sheer size. Seems like it would be a reasonable match for an early 50s Buick, and if I found a donor Roadmaster, seems like a lot of parts would be quite useful. Beyond an engine even...suspension, transmission, drivetrain, etc..

[bammax]

Coming from the Gen2 world I can say that the LT1 is a good motor. I'm not a fan of that over a generic Gen1 sbc though. Basically the biggest benefit to the LT1 is the head design and the fuel injection. Now that you can bolt the vortec heads to the Gen1 there's not much value in LT1's anymore. The trans is junk that comes in that car too. Stay away from the 4l60e if you can help it. Also they tend to have issues when it comes to grounds and other electrical things.

I'd do as Silver says since he's been down the fuel injection retrofit road and speaks from experience.

[TheSilverBuick]

Have you ever seen this:
http://www.bangshift.com/forum/dto_garage.php?do=viewvehicle&v=1072

It's a 1956 Belair wagon on a early 90's Caprice frame complete with LT1, etc. drivetrain. The stock ECM controls the transmission and a MegaSquirt is piggybacked onto the stock wiring to control fuel and spark if I recall right. So I'd say it's a reasonable match :)


Carb's work okay at elevation, they just run super rich if set up for low elevation or lean out if set at too low of an elevation. An example I have, is pulling the Eisenhower tunnel pass outside of Denver, CO. 11,400ft elevation. I pulled it in my Centurion with a 455, Q-jet, points,no OD, 2.93 gears, and could barely run 40mph because the tune was so far off. I had all the barrels open, but there just wasn't enough O2 or air pressure to make up for the load. I've pulled steeper summits at lower elevation with zero issues before. Then last September, I drove my Skylark, EFI'd 455, fuel and ignition, deep overdrive, cruising at ~2200rpm, pulled that pass at 75-80mph no sweat. I was up around 50% throttle and AFR's stayed right where they should for good power climbing the summit. Logged the barometric pressure at 64kPa, which sea level is at 101kPa, so 36% less air pressure. So it just depends how often you'll be seeing those swings in elevation. A carb will function, but not optimally.

[S2X01]

Heading west from Denver is a definite. So yeah....Eisenhower and Vail pass will get me way up. As well as headed to Michigan and getting down to 500 feet and possibly lower. It's gonna see an extremely wide variance of altitudes and air densities.

From everything I'm reading, however, it seems that there wasn't really a Buick specific LT1. Just a generic GM package. I really don't wanna be the guy dropping Chevy stuff into a Buick!!

[TheSilverBuick]

For an LT1 you would become "THAT" guy =P lol.

Again, a carb won't leave you stranded on the side of the road, but they can get pretty cranky covering that elevation. I live at 6,500ft elevation, work above 7,000ft, Vegas is the next town south at 2,500ft and "back home" in California is around sea level. Several times a year I'm driving off this snowy hill through the blazing desert and to the sea :pp

[pmuller9]

One of the next engines Jim Weise will be completing is a high torque BBB for a 65 Buick station wagon with fuel economy as the main goal. The most important component of this project is the overdrive transmission with the lock-up converter.
This engine will also be EFI.

Paul

[TheSilverBuick]

I'll be interested to hear how that goes. I've been at an impasse getting to 20+mpg. Stuck at 19 =/ I just installed a narrower tire on the car with a more road friendly tread pattern, so hoping to see some gains next time I leave town. I had a lighter '69 Firebird with a weak 400, 700r4, 2.73 gears getting an easy 25mpg highway, but it was way slower and likely way lighter and definitely more aero. I did get a regular 23mpg with the Skylark with a low compression/small cam combination. It could simply be the cam, but I don't want to admit that, so that being said, I'd like to see what JW does!

[S2X01]

20?!?!? That's awesome! Is that on the injected engine or the q-jet??

[TheSilverBuick]

With the current engine, I got a best of 16 with a Q-jet (non-optimized) and 19 (repeatable) with the EFI. With the old low compression engine with small cam it was 23 (repeatable) with a Q-jet (it also had narrower tires). The low compression engine only makes around 300-320HP, so waaay docile.

[S2X01]

docile yes....but 300 would still be enough to have fun.

Anything getting over 17 MPG (what my Xterra gets now) is cool by be for driving cross country.
Starting to like the idea of using my spare 455 with low compression heads or even a 401 for an experiment in fuel injection.

added bonus.... not being that guy with the LT1

[TheSilverBuick]

300-320HP, 3.23 gears, open diff, exhaust piped down to a single 2" glass pack, and 23mpg highway, 16 around town and yes, still plenty of torquey fun!

http://www.youtube.com/watch?v=R_CCLSwdLhU
There is a slight break in the rubber for shifting to 2nd gear =D

[elagache]

Howdy Paul and V-8 Buick members, :)


One of the next engines Jim Weise will be completing is a high torque BBB for a 65 Buick station wagon with fuel economy as the main goal. The most important component of this project is the overdrive transmission with the lock-up converter.
This engine will also be EFI.

Is that so? . . . . Golly, I wonder which Buick wagon it is going to end up in . . . . . . :Brow:

Cheers, Edouard :beer

[TheSilverBuick]

Howdy Paul and V-8 Buick members, :)



Is that so? . . . . Golly, I wonder which Buick wagon it is going to end up in . . . . . . :Brow:

Cheers, Edouard :beer


HA! I should of known! Way cool!

[doc]

Well,,,, I did this on a jeep.... and it did not gain any mileage,,, the power stayed about the same... and it had a high speed miss that could not be tuned out.... finally took it back to a carb... back to entirely normal....
If you absolutely must go to efi,,, go to the ''wet'' systems where the injectors take the place of the carbureator....they are simpler to work with... but dont expect more power or more mileage....but do expect that working on it will be more complex....
Give me a good working carb set up any time.....conversion to efi just aint worth the hassle, time, and expense.....

[pmuller9]

Well,,,, I did this on a jeep.... and it did not gain any mileage,,, the power stayed about the same... and it had a high speed miss that could not be tuned out.... finally took it back to a carb... back to entirely normal....
If you absolutely must go to efi,,, go to the ''wet'' systems where the injectors take the place of the carbureator....they are simpler to work with... but dont expect more power or more mileage....but do expect that working on it will be more complex....
Give me a good working carb set up any time.....conversion to efi just aint worth the hassle, time, and expense.....

Doc

You forgot to tell the whole story on why EFI didn't work for you.
If I remember correctly it was a verification of what was stated in post #5 of this thread


Using a stock ECM from the junkyard will disappoint you. 90% of the people I hear complaining that EFI didn't work well for them was from using a stock or re-flashed ECM from a different engine. It's more of a headache than shortcut. I recommend going with an aftermarket ECM you have full control over. You've read my post in the other thread so no point in me re-hashing it.



Paul

[doc]

Yep,,, and i might add the brain box was one of those that could not be tuned,,, or at least I could not find anybody around this little wooden axle town that could do that for me..... That said,,,, the way I look at it ,,, if I add a sensor,, that sensor can go bad,,, that adds about 5 more possibilities of trouble... and 5 things that have to be checked out.... and then there is the complex fuel system and 8 injectors.... and on and on and on..... simpler is better.....

[bammax]

I have kept downgrading over the years. I went from a Gen2 L99 motor with multipoint efi down to an Olds 307 with the CCC system (computer controlled carb) down to a straight old fashioned carb. Now I'm going to swap the quadrajet for an edelbrock to make it even more simplified and easy to use. Eventually though I'm going to go in the other direction such as HEI, electric fans, 2004r trans, and eventually a home built multipoint sequential fuel injection.

I find that the systems in themselves are perfectly fine. It's the way the factory wires them together that's the big issue. If you put it together yourself it's easier to know what goes where and avoid any issues down the road. Plus GM has a habit of piggybacking systems together using wire that's not the right guage. Home built systems can avoid those 2 problems if you're comfortable with figuring out how to wire things together the right way.

[elagache]

Dear V-8 Buick electronic gear-heads,

Just a quick additional comment about EFI and fuel economy.


Second part of the question....is it worth it as far as fuel efficency gains?

The great big-block Buick wagon fuel-economy caper has a lot of components associated with it. Actually EFI is just icing on the cake and wasn't the original plan. Jim proposed a Quadra-Jet at first but decided that this was in a way "beneath the dignity" of the overall project. The car will get a hardened 200-4R tranny. It will have stage-2 cylinder heads, exhaust headers, Jim's "secret" cam and all sorts of other goodies. So the hope is that this engine will take advantage of all sorts of racing engine technology and direct that instead toward fuel economy.

There have been discussions periodically about fuel economy. EFI can help, but to make really serious gains you should take (as corporate slang would put it: ) "a systems approach" to the problem of better gas mileage. Even the best EFI system won't help much if you have poorly chosen rear-end ratio (just to pick one issue.)

Cheers, Edouard :beer

[superbuickguy]

Your question has so many variables, and I agree with Randall's point that an overdrive transmission and a carb will get you 90% of what you're after. EFI is awesome after it's properly tuned, especially in the cold-start and drastic elevation change scenarios.... it also may get you better mileage. That hypothetical 454 fuel injection idea - on a 454 the difference in mileage is about 25% (which sounds awesome until you realize it went from 9 mpg to 12 mpg - but still will get the same mileage when fully loaded).

If you're planning on a power adder - run, don't walk to EFI. If not, it's really hard to do the work to convert from the simplicity of a carb. Especially if it doesn't work right the first time (my first EFI was a holley projection that I couldn't get to stop flooding the motor).

Of course, should you take the plunge, Randall, and Paul are both excellent resources for all things geek. :)

[TheSilverBuick]

That hypothetical 454 fuel injection idea - on a 454 the difference in mileage is about 25% (which sounds awesome until you realize it went from 9 mpg to 12 mpg - but still will get the same mileage when fully loaded).

You caught on that I was using percentages for that reason. I've heard several people getting sub-10mpg going to 12-14mpg with an EFI swap, but you don't hear people that are getting 12-15mpg netting too much more and I think it's because performance carbs, namely square bore ones, tend to give too much fuel at cruise in order to satisfy the fuel requirements at WOT. Q-jet's and EFI do better in that department on trying to give the best of both worlds. My personal observations on my vehicles is that the OD trans nets an increase in mpgs almost exactly proportional to the drop in rpm. 30% drop in rpm, 30% increase in mpg, might not be exact but usually within 5%.

Another word of note for why I found it worth it. I drive my car an average of 12,000 miles per year or more. Not many here put that kind of mileage on their old car.

As for the sensors going out, or injectors, etc. Sensors really don't fail that often, especially considering how many millions of cars are running them now? I get mine second hand from the junkyard when ever possible. And with a standalone ECU it's a piece of cake to diagnose. Example being my shoddy wiring job made for my intake air temp sensor to intermittently quit reading, which then the ECU would default -40*F in this case. Driving I noticed a hiccup when it would go a tad rich from calculating the air as denser than it was, but as soon as I plugged the laptop in, five seconds of looking at the data inputs it was obvious where the hiccup was coming from, a sharp change in temperature to an unreasonable temperature for the day. Told me exactly where the problem was, which was installation error not the sensor itself. At no time was I even stranded, just a slight engine hesitation as the fueling calculation radically changed. People spin circles chasing mystery ignition and carb problems for simple hiccups like that.

Another favorite story of mine that would of been impossible with a carb was when I blew a head gasket into the lifter valley while racing in Phoenix, AZ, 800 miles from home. Based on the spark plug and by removing one plug at a time I determined which cylinder was hurt the worse (I actually thought I cracked a piston or ring until I pulled the head) by the decrease in blow-by, so I unplugged that injector and spark plug and drove the 800 miles home on 7 cylinders. Sure a bit of fuel may of been sucked into that cylinder, but no where near as much going into the crankcase and exhaust if it were a carb. Another time I wiped a cam 150 miles from home (underrated valve springs) and it started bucking and backfiring on me. I managed to successfully troubleshoot and guesstimate which TWO cylinders were the worse offenders, unplugged the injectors and spark plug wires and drove it home on six cylinders with out too much hassle. Pulling the intake verified I guessed which two were giving me problems.

35,000 miles on the EFI since 2009 and counting.

--Randal

[S2X01]

Well...I'm liking this more and more. My debate still lies on a couple factors right now, but all said and done I think the port injected rout is the way I want to go.

First thing.....can anyone tell me about these? http://www.stylintrucks.com/parts/crosswind_efi_conversion_intake_manifolds/1104000204_22689.aspx#
It says pontiac....but the also claims to fit a Buick 455 application. I'm still debating the cheaper route and doing it myself, but this could save me a lot of effort and minimize my "woops" factor.

Second question.... can injectors be mounted horizontally?
Lets say I can get my hands on a 401. Could risers be fabbed up so the injectors sit below the intake?

Third thing....thank the good lord the site is back up!! I've been losing sleep and going bananas thinking of all the possibilities!!!

[S2X01]

Also..... thoughts on s 454 TBI/harness/ECU?

[TheSilverBuick]

I plan on mounting the injectors horizontally on the OHC straight six I'm building. I'm possibly going to have a ~90* bend in the intake ports to the head and run the injectors horizontally shooting directly into the port. Run a straight length of pipe from the head roughly the volume of the cylinder. Even if I don't do that, I don't see any reason why it couldn't be done, they are sealed systems.

I wouldn't use a stock ECU. 454 harness, maybe and sure on the TBI.

[S2X01]

Okay cool.

So it sounds like regardless if I go TBI or MPFI, a standalone system like the Megasquirt is my best bet.

[S2X01]

I'm definitely open to other suggestions. I would definitely prefer a system that is: 1. Affordable and 2. well supported or known well by fellow board members. The last thing I need right now is to be a guinnea pig!:rolleyes:

[elagache]

Dear Stu and V-8 Buick EFI wannabees,


I'm definitely open to other suggestions. I would definitely prefer a system that is: 1. Affordable and 2. well supported or known well by fellow board members. The last thing I need right now is to be a guinnea pig!:rolleyes:

I know it is a chicken and the egg situation, but I found myself seriously restrained by fears of becoming a guinea pig. There are quite a few systems that have enthusiastic owners on them "other engine platforms." However, I just couldn't stomach being the first Buick guy on the block with a "Super-Zippy" EFI system. I sure wish the various EFI manufacturers would take the time to have a least one sample installation of their systems on every engine type they claim to support and a real-live car owner that someone like me could have asked to see if the claims made are really true.

So I'll wait out the sequential fuel injection route until more brave souls tell me about it. I also sure would like to see some hardware advances . . . alas in this economy, R&D is the area probably cut the most by everyone.

Oh well,

Cheers, Edouard

[TheSilverBuick]

O.k let me put this fear to bed once and for all, I hope.

Any EFI system has no idea what brand of application it's on, PERIOD! It's not human or living in any way shape form or fashion, it's a thing for crissakes, not a being.
Now do spark/fuel maps differ from say one brand of 350 to another? Of course they do, but that also has to do with camshaft grind, head flow, intake design and so on. But remember for the most part, it takes X amount of fuel to make X amount of horsepower etc.

What you should look for is ease of set up, ease of software navigation and overall support. I and other Accel dealers support this system, the software is some of the best/easiest in this class. But I tell everyone, download everyones software, use the help screens and make your determination from there.

Thank you.


I agree. EFI doesn't know what it's on. Just like a carb, it doesn't care what it's on, it just has to be set up right.


I'm hoping this weekend to go junkyarding to pick up some ignition coils, possibly a step towards to going distributorless on my Buick if I find the parts I want, and I hope to get some Buick 3.8 coil pack sets from the mid-80's for my OHC L6 Pontiac build.

[TheSilverBuick]

I kept meaning to post this link.

This is a pretty good book on the principles that EFI works under and covers a bit on how OEM's went about doing things one way or another. $13-$23 is well worth the money.
http://www.amazon.com/Performance-Fuel-Injection-Systems-HP1557/dp/1557885575/ref=sr_1_2?ie=UTF8&qid=1328797227&sr=8-2

[pmuller9]

One of the problems is that the thinking is still stuck back in the 20th century during the analog age of carburetors.
Think Digital where any chunk of data can be sent anywhere in the world, modified and returned in seconds.

If you purchase an EFI system that has data acquisition then the tune-up can be viewed by anyone, anywhere.
They can make changes and the new file can be downloaded and tried on the spot.

One of our engine tuners would sit at home in New York while we ran the car in the shop in Spokane WA,
be telling us what he wanted us to do by cell phone and he would make changes to the EFI system
while the engine was running using a wireless broadband connection to the laptop connected to the car system.

With the new cell phones, you could be driving your new EFI system and have someone else tuning the car for you
(long distance) as you drive down the road.

Some systems even store mulitple tuneup files that you can change between with the flip of a switch.
Could be handy if one of those files was setup for State emissions testing.

Anyway, the bottom line is that many of the new systems allow you to get support as fast as you can email files back and forth
and in some cases can also be real time support.

Paul

[elagache]

Hi Paul and V-8 Buick hi-tech types,

Let me apologize in advance because this going to seem rather harsh. Still I think this needs to be said because I fear that this is precisely the sort of seriously misguided thinking that pervades the "cloud" age.


One of the problems is that the thinking is still stuck back in the 20th century during the analog age of carburetors.
Think Digital where any chunk of data can be sent anywhere in the world, modified and returned in seconds.

. . . . .

One of our engine tuners would sit at home in New York while we ran the car in the shop in Spokane WA,
be telling us what he wanted us to do by cell phone and he would make changes to the EFI system
while the engine was running using a wireless broadband connection to the laptop connected to the car system.


Sure enough, cloud technology allows us to do this - but is it really what we want? :Do No:

Before we get too excited, let us recall the problem to be solved: quite simply: the engine needs to be supplied with correct amount of fuel and air at the precise moment to provide best possible combustion. Now last time I checked, this is a matter of straightforward physics and chemistry. When I was getting my Bachelors at U.C. Berkeley, this was sort of a problem that a graduate student would be assigned as "punishment" before getting onto the glamorous particle physics (where the :dollar: was :() So if this is the sort of thing that should be possible to solve with a suitable amount science and engineering - why would be tuning your engine over and over again - never mind need to ask human experts from all over the world to do it? :Do No:

There is an saying that seems desperately relevant in this context: "too many cooks spoil the sauce." So tell me, how many experts (and how often) do you need to figure out how to solve the physics and chemistry of an internal combustion engine?

I've complained before about my unhappiness about the way aftermarket EFI systems use look-up tables to accomplish tuning. Clearly it is "good enough" - cars work fine after all. However, this precisely the worst possible instance of "digital" - in its original sense of the phrase: chopping up the world into discrete bits. Fuel should never been dispensed by measuring cups. A gasoline engine is an analogue device that should be controlled by analogue functions (like those functions I toiled over as an undergraduate.) The appropriate design for the electronics if a fuel injection system is curve fitting and the mathematics behind it hundreds of years old. Once more as a naive undergrad, I wrote software that did curve fitting - that's what physics undergrads did. Yet today I'm being told that the best electronic technology that can be put into my car is unable to handle the sort of simple mathematics that I did with a computer that had less power than the processors going into those very EFI systems of today . . . . . . . . . :confused:

Excuse me? ???? Somebody has applied a bit too much technology - and not nearly enough good science and engineering . . .

Sorry if this seems beyond what any particular person or company can deal with. Yet, I think we need to be truly honest about what technology has really accomplished and how its misapplication had led us astray. As I understand it, the aftermarket EFI systems rely on what auto-makers have done. Either they have kept the best of the technology for themselves, or worse, in the desperation to meet all the conflicting demands placed on them - they haven't done the basic science to develop suitable computer models of engine combustion.

I'll be blunt - human beings should never tune cars! The internal combustion engine is a physical system and it should be possible for us to create accurate computer engine simulations - down to reliabily predicting fuel-air mixture and injection data. With such models, an onboard computer should be able to determine in real time from available sensor technology what the fuel-air mixture to use and when to inject that mixture. At most, humans may need to be involved in adapting the computerized tune to fit the driver's personal preference. That is what technology should be giving us. If we don't have this sort of thing, it is because the greatness that allowed us to win World War II has been allowed to slip through our fingers. :(

I don't mean to belittle our technology that make it possible for folks all over the world to share things. However, any sober look at what this technology has actually accomplished should put things in perspective. To use another modern saying: "None of us is as stupid as all of us put together." The greatest hope for such technology is greater understanding and harmony among human beings. If you believe this is actually happening, check out this headline (http://news.yahoo.com/facebook-defriending-led-double-murder-police-014442236.html). Global connectivity shouldn't be the route to improved engine tuning. Sure there will be cases that it is helpful. However, if there is something we have largely failed to do with technology, it is apply it to problems of physical automation. Geeks don't like to get their fingers dirty, so don't tell them about how an internal combustion engine works. To them, it is better to write software to have a pizza delivered than to understand how to get 24 MPG from a 65 Buick Special.

For everyone who loves old cars this is very ominous indeed. :ball: These are the folks who see electric cars are the answer and that classic car lover are as obsolete as their cars.

Computer networks are simply a tool - not a panacea . . . .

Sincerely, Edouard

P.S. Those geeks also believe that they can recharge their electric cars at night using . . . . . solar power! :spank:

[pmuller9]

Edouard

My reply was simply to give comfort in knowing that if a person needs help it is available quickly and easily.
A support person can look at what is going on remotely and can either give advice or make changes by email file exchange or directly.
This is only if a person needs help in the first place.

Wouldn't it be great if all products had that kind of support
So there only needs to be one cook in the kitchen unless that cook needs help.

Your argument is based on a very narrow aspect of what an EFI system is expected to control.
You made incorrect statements that were used to support the analogies that comprised most of your reply.
Setting up your own premise that is unfounded then building on it for the sake of augument is called a "Cheap Shot"
Using an EFI system to accomplish a 24 mpg wagon is your agenda and is not the only task a system is designed to do.
How about the rest of us?

You only spoke about fuel control. I'm sure you meant to at least include spark. Right?
If you look at a list of all areas a system can control, you will see dozens easily!
Consider that these system can handle many types of fuels and fuel additions like nitrous and other gasses.
There are a pile of functions just to support competition.
Don't forget power adders and manifold pressures above 1 atmosphere.

The list goes on and if you look at the combinations, you either make hundreds of dedicated units
or a single unit able to control it all.

The only way a universal unit is going to know what job it needs to do is by someone telling it by entering the initial values.
At that point the system is operating based on what you expect as a resulting response by the entries in the software.

However you may want a different response and how in the world is the system supposed to know that unless you tell it by making changes in the tables that represent your intent?
Are you proposing a system with mind reading?

I might be using the system to control a twin turbo, methanol staged injected (24 injectors), 2400 hp, 2 and 3 step limiters that tie into the mag controller, and muti level algorithm traction control, while you can use the same system to control a 24 mpg street car with a TH200 lockout converter tranny.

Do you think that some human intervention would be required to make the same system work for both applications?
and do you think that my turbo application may require me to ask for help?

Your reply was irresponsible but it gave me the opportunity to return the bashing I'm sure you were expecting.

Sincerely

Paul

[BRUCE ROE]

Just a couple comments. Spark is VERY IMPORTANT, I proved that with tunable crank trigger conversion on a carb.
Second, fitting curves is good. The 70s ANALOG EFI tried, but it was strictly a line with a couple bends
transfer function. Making a multi input transfer function transfer (for spark & fuel) is only workable with
digital systems, which is why (in part) the change was made about 1980. Bruce Roe

[elagache]

Dear Paul and V-8 Buick gearheads,

Let me repeat my apologies. However, I think my point simply wasn't understood. My complaint was that what we see today in terms of EFI systems doesn't represent the proper "marriage" of the science and engineering with the "magic" of a global Internet. We have neglected the careful work of building good engine simulations with a heavy layer of information technology that requires way too much of the technical support you are advancing.


My reply was simply to give comfort in knowing that if a person needs help it is available quickly and easily.
A support person can look at what is going on remotely and can either give advice or make changes by email file exchange or directly.
This is only if a person needs help in the first place.

Hey, how can anyone disagree? Yet, I stick to my main point: shouldn't a physical system require such expertise -once? If a car requires repeated tuning - do you fully undertand this engine that you are trying to tune?


Your argument is based on a very narrow aspect of what an EFI system is expected to control.

No it doesn't, but I trying to make the point that is very difficult to make. Let me try to make that point more clearly. An engine is a physical system. If there is software to model the entire global climate - shouldn't we be able to construct a model of an internal combustion engine such that we can predict accurately all the parameters that must be controlled electronically. Yes I didn't mention the application of spark, but the point is analogous. If the simulation can predict the moment of optimal combustion - it can trigger the spark plug. Today, you can buy software to run weather forecast models that take actual weather data and generate forecasts on your PC. Why don't you run something analogous on your car to simulate the engine and determine fuel air mixture, spark timing and anything else necessary - not based on some lookup table - but via a accurate simulation of what the engine is actually doing?



You only spoke about fuel control. I'm sure you meant to at least include spark. Right?
If you look at a list of all areas a system can control, you will see dozens easily!
Consider that these system can handle many types of fuels and fuel additions like nitrous and other gasses.
There are a pile of functions just to support competition.
Don't forget power adders and manifold pressures above 1 atmosphere.

You can add anything you like, that's not the issue here. The issue is - do you understand the system you are trying to tune? It is better to throw up your hands and insist in effect - we cannot understand how an engine with twin turbochargers, nitrous, etc. actually works and return to a world of engine tuning as a black art?

This may seem idealistic until you consider that internal combustion engines are . . . . centuries old. Computer simulations of physical phenomena are over 50 years old. There are how many automobiles in the world today? What are the pressures for improving fuel economy, reducing pollution and so on? Shouldn't this sort of simulation be just as commonplace as say simulations of global climate?

My view comes from a computer perspective and my complaint is that EFI systems lack what is called an "abstraction layer" so that what is worked on with the computer emulates everyday experiences. Desktop computing was a revolution in user-friendly design. In the same way, no bit of computer technology should appear "naked" but should restore to the user their "workplace." So for your Paul and other professionals in engine design the "desktop" would be a simulated engine. That engine would have parameters based on actual engine components. Adjusting those parameters would no longer be a black-art because the effects on the engine could be simulated in advance. Just as engine design software allows one to compare the effects of different cams, heads, and so on, now the tuning of an engine would be based on an environment replicating the engine itself. No longer would it be necessary to have a "expert tuner" adjust particular lookup table values "because in the past that seemed to make the problem go away."

Is this idealistic?? . . . well that gets to the core of my complaint. Effectively, there is a whole "High Tech" world that couldn't care less about how an automobile actually works. So instead of having the sort of simulation software I just described - we have tens of thousands of software engineers writing computer games, Social media toys, and other wasteful matters that cost everybody money and even lives - as my example of a murder driven by Facebook activities illustrates. If 25% of that effort was spent developing good software for EFI systems, what I just described is well within reach.

My complaint comes from this perspective. I'm sadly left to say to you Paul, Mark and most of you: "you have been brain-washed into thinking you have been truly well served by the High-tech geeks." Sure you can get your tech-support remotely. What you really needed was well thought out software to start with. A few days ago, I ran into a bug at the core of the Apple script system: AppleScript. AppleScript is decades old - yet no fix. Everyone has to work around it - and that's how people are expected to live.

Don't "be thrilled" that the computer industry have given you a few crumbs like global networking. In the end the Internet has become the plaything of spoiled boys and the rest of the world is being sorely neglected. Performance engine technology being only one of many. My vision for a "more better" way to do EFI may see like a hopeless dream. If it is - it is because ALL OF US have been let down by those technologists who promise "not to do harm." Apparently neglect isn't considered harmful.

Hope you all can see high-tech for what it sadly is. :ball:

Sincerely,

Edouard

[pmuller9]

Edouard

I was hoping that you would take the time to carefully read what I previously wrote including some reading between the lines
and then compare it to the content of your conversation.

If you notice I said nothing about tuning in my last reply and your conversation is all about tuning.

I'm doing my best to give you more insight on how an Engine Management system works. The bashing was just for fun.

A high end system that is running in a closed loop mode does tune the engine exactly to the requested parameters from the user.

Your premise is that there is a universally desired tune up that can be technologically obtained and maintained by a software/hardware combination and that there is a failure on the part of the tech community because it ain't happening making us all victims.

I'm telling you that if there was a system that you could install on any engine and it would measure every parameter from every input sensor and do all the math and have the engine running perfectly to some specification, it would only be right for a narrow application and wrong for a lot of other users.

That is exactly what we have present day and because I like chocolate and someone else prefers vanilla is why the resulting response must be user defined and cannot be predetermined by some computer software.

The high end systems do measure some of the same parameters as a weather station along with the engine stuff and tunes the engine accordingly and will also automatically adjust the tables to reduce the deviation which increases system response time.
No one has to go in and continually retune as you where suggesting.
The experimenters that want to learn more by making changes and checking the results are the ones that are glad to have the option but again it is not a necessity.

When we are bracket racing there is a full blown weather station on the trailer connected to a laptop that does nothing but compare previous run conditions with present conditions. Every 2 minutes the system transmits the predicted ET (to the hundredth of a second) to a pager on my belt to be used as the dial in time for the next run. So the technology is certainly there and is far from being a "Black Art"

It is just not useful in the way you are envisioning it to be.
I will always want a different engine response than the next person even on the same engine type for the same type of racing.

The other piece of this that I was trying to explain is that when there is a universal device or system that is expected to control an engine that can have hundreds of different combinations of hardware both internal and external to itself, there is no way for that system to know what that combination is until the user defines the job by setting up the hardware configuration and the initial run parameters in the software and that is where the user may need help more than anywhere else. NOT the TUNING!
As you were suggesting once the hardware is defined, the software can develop a base tune up which is fine tuned as the car is run without the need for the user to do any further action.

In racing we need to control the tune up because we tune for different reasons and sometime mistune on purpose for the sake of creating a different ET or because of track conditions.

There are a lot of other exceptions that you are not aware of that makes any premise for argument invalid.

Also remember that everything that determines the operation of an internal combustion engine is variable including fuel.
That means that there can be no set tune up and only a base table to work from with the processor constantly calculating the deviation using the sensor input information to obtain the desired output.
Again that desired output is defined by a person in the initial setup and even if it software produced, the target values driving the calculations are still defined by a person.

You really need to work with a variety systems for a variety of applications before you understand where the shortcomings are and are not.

The systems I have worked with are designed both hardware and software by guys that race, not some geek that is detached from the sport.
That is the rule and not the exception so quit making statements about subjects that you have no back ground in just so you can support some argument. That is a cheap shot.

We asked for multilevel traction control so we could run 2400 hp with 10.5 inch tires, we got it.
We asked for an even misfire 2 step so all cylinders would be the same temp at launch, we got it.
We said we where switching over to a ProMag 44 and needed an interface for the mag controller, we got it.
We needed the unit to operate closed loop with A/F ratios below 4 for methanol, we got it.

We basically can have anything we can dream of when it comes to control and data aquisition.

So correct me if I'm wrong. You have yet to have hands on with any EFI system giving you only second hand knowlege at best.
You also have had limited hands on with engines and still are working to understand the basics.
Armed with this you start a discussion on the failures of our present technical community with the pretence of having some authority on the subject.
Then if you look at the structure, the arguments are based on unfounded statements followed by analogies that make the whole discussion a little off base. This is a typical stratagy or outcome when an author is out of his/her area of understanding.

Sincerely of course

Paul

[TheSilverBuick]

One of the problems is that the thinking is still stuck back in the 20th century during the analog age of carburetors.
Think Digital where any chunk of data can be sent anywhere in the world, modified and returned in seconds.

If you purchase an EFI system that has data acquisition then the tune-up can be viewed by anyone, anywhere.
They can make changes and the new file can be downloaded and tried on the spot.

One of our engine tuners would sit at home in New York while we ran the car in the shop in Spokane WA,
be telling us what he wanted us to do by cell phone and he would make changes to the EFI system
while the engine was running using a wireless broadband connection to the laptop connected to the car system.

With the new cell phones, you could be driving your new EFI system and have someone else tuning the car for you
(long distance) as you drive down the road.

Some systems even store mulitple tuneup files that you can change between with the flip of a switch.
Could be handy if one of those files was setup for State emissions testing.

Anyway, the bottom line is that many of the new systems allow you to get support as fast as you can email files back and forth
and in some cases can also be real time support.

Paul


Paul, do you work with the MegaSquirt systems? There is an Android app called "Shadow Logger" or something like that, for if you have a BlueTooth connection to the MegaSquirt you can use an android app to view the tune and data log, but it takes it one step further and automatically uploads the log to a server on the net to be viewed anywhere or time through the internet. It also records the phone's accelerometer and GPS data with the ECU data for even more information, particularly as the vehicle is going around a track. The app is being refined further, looking for more ways to tune with it, but it's pretty neat that it's grabbing the acceleration and location data as well as uploading to a central server so someone across the country can look at it just as easy as the person next to the car.

[pmuller9]

Paul, do you work with the MegaSquirt systems? There is an Android app called "Shadow Logger" or something like that, for if you have a BlueTooth connection to the MegaSquirt you can use an android app to view the tune and data log, but it takes it one step further and automatically uploads the log to a server on the net to be viewed anywhere or time through the internet. It also records the phone's accelerometer and GPS data with the ECU data for even more information, particularly as the vehicle is going around a track. The app is being refined further, looking for more ways to tune with it, but it's pretty neat that it's grabbing the acceleration and location data as well as uploading to a central server so someone across the country can look at it just as easy as the person next to the car.

I have not had the opportunity to work with the MegaSquirt system.
Thanks for the information on the phone app!
That's what I was talking about! How cool is that!

The great part is there can be multiple people anylizing the data at the same time which gives a better chance to catch a problem that might have been missed otherwise.

The trouble shooting aspect is something I didn't include in the above discussions along with many other uses for remote data access.

The best reason is, it is just plain fun.
It's even more fun when you analyze the G meter results and give the driver a bunch of grief over his driving skills.

Paul

[GNandGS]

There was a time when you had two basic types of folks... ones that could tune awesome but didnt like or pretended not to understand the computer controlled setups. The other guys "got" the interface just fine and maybe even wrote their own code as a hobby but fell short on the tuning.

Those days are gone!

You don't have to look any further than the Buick turbo guys for examples. Adding and removing fuel where you wanted as well as playing with timing to see what changes in spoolup took place was actually getting results for us on one car. If WE could do it a lot of people could - and did.

Lock/Dont lock converter. Lean cruise for hway. Antitheft. Big injectors that don't behave like it... were all common place years ago.

Todays systems are VERY capable and much more so than the 80s GM 8bit setup. The guys that can both tune AND make changes via laptop are common now. Boost control, closed loop fuel/timing, traction aids, and even a valet mode are all available without getting dirty (once the dirty work is done).

The performance numbers don't lie!

PS: I once got 22mpg w/455 and Qjet no OD. Had to drive at 50mph though :)

[TheSilverBuick]

However, this precisely the worst possible instance of "digital" - in its original sense of the phrase: chopping up the world into discrete bits. Fuel should never been dispensed by measuring cups. A gasoline engine is an analogue device that should be controlled by analogue functions (like those functions I toiled over as an undergraduate.) The appropriate design for the electronics if a fuel injection system is curve fitting and the mathematics behind it hundreds of years old.


How about this. Consider the intake valve. The gate way to the cylinder. It's either open (on), or it's closed (off). Sure there are varying degrees of "on" but there is only one type of "off" that ends the sequence. Sounds pretty digital to me. That's the reason fuel puddling is a problem, the valve get's turned on and off and on and off and on and off, rather abruptly too. In either case, short of direct injection both EFI and carbs are supplying fuel when the valve is off simply to ensure the supply of gasoline is sufficient for the power expected.

[elagache]

Dear Paul and V-8 Buick engine gurus,


So correct me if I'm wrong. You have yet to have hands on with any EFI system giving you only second hand knowlege at best.
You also have had limited hands on with engines and still are working to understand the basics.
Armed with this you start a discussion on the failures of our present technical community with the pretence of having some authority on the subject.
Then if you look at the structure, the arguments are based on unfounded statements followed by analogies that make the whole discussion a little off base. This is a typical stratagy or outcome when an author is out of his/her area of understanding.l

Okay I'm willingly admit to all the limitations you mention here. However, I'll have a corner of expertise of my own and my hopes were that you'all would see what liberating possibilities are possible if you got the right kind of support from that corner of the world.

Your right, I'm arguing in abstractions. However, I don't think my abstractions are unreasonable. It is possible to create computer models of something as complex as global climate and intersellar systems like black holes. Does your expertise allow you to conclude that it is beyond information technology to come up with a universal architecture to model the physics and chemistry of any internal combustion engine that can be built? My understanding of the science and technology sees no such limitation. Should we be truly so pessimistic as you insist?

I've tried to make my point clear. I'm not faulting those who actually work on engines and indeed I've tried to make it clear my profound admiration. I thought my complaint was clear: you haven't been given reasonable tools from the software world to do your job properly. Let me try to make my analogy of the computer desktop again. Before the Macintosh, people typed obscure commands to get computers to do what they wanted to do. After the Mac, what they wanted to do was reflected in the very experience of using a computer: there were pieces of paper, folders, and places to put them. The office worker was given in software an emulation of the world they physically worked in.

So change the context to using electronics to control an engine. We are back to days of MS-DOS. Yes, I haven't looked in detail at the various bits of user software. However, the core of all speed-density EFI systems is a set of lookup tables. Exactly what do those values represent? Do all these values represent parameters that correspond to physical properties of the engine? If they did, nobody would call it "lookup table." You Paul and everyone else working with engines have been forced to become computer programmers. As someone who has done this sort of thing, it goes almost all the way back to programming a microprocessor in assembly language. It is about as crude a user-interface as computing can be.

What I'm proposing as an alternative is a layer of software that becomes the "desktop" for the engine world. Instead of lookup tables, software uses physical properties of the engine itself and presents the user (you Paul) with software generated emulation of that the engine is actually doing for the given parameters selected for the EFI and other electronic control systems. If the engine isn't performing as well as it could - the simulation shows it - and simulates the underlying physical processes that are degrading its performance. Now you are back to designing engines and when you try to get the best performance out of an engine - you are attempting to understand the mechanics and engineering of that engine. Isn't that what you are really interested in?

To put it in a nutshell I'm asking if the present capacity to communicate with the EFI hardware (effectively the language to communicate) truly allows you to express your expertise as an engine builder and optimizer into that electronics? Okay, call me ignorant, but it sure seems to me like everyone in this arena finds themselves unable to express what they want the systems to understand. That's the sort of communication problem information technology experts are supposed to solve. Perhaps you disagree, but I feel that information technology has failed you Paul and everyone who provides internal combustion engines for whatever the applicaton.

Now you can insist that what I've describe is an impossible dream - congratulations, by that very insistence you gone a long way into making it a self-fulfilling prophesy.

"beware of what you wish for . . . you might get it."

With a heavy heart,

Edouard :(

P.S. There seems to be sense "personal insult" here. What I'm describing here is well-beyond what any person or company that caters to the V-8 Buick crowd can do. So I'm not "blaming" anyone here. Nonetheless, we all suffer when things could be better and aren't. Whether it is classic cars being junked because people can't afford the gas or people pushed into poverty because the country spends too much of our GNP on oil, we all have a real stake in questions like this. Steve Jobs is a good role model for this. Sure we cannot wish miracles into existence. Nonetheless, if we don't set standards of greatness for ourselves - what hope do we have of ever reaching them?

[TheSilverBuick]

Simulations and models are not substitutes for actual tests and real life. Global climate models are b%& $h!!, and don't predict jack. Model's on black holes are being revised all the time. Way too many people actually believe models are real life, and research scientist are the worse, when in fact they are meant to emulate real life and more often than not can't because the number of influential variables are so large, uncountable even, that they are limited by that. I hate to use the term garbage in, garbage out, but it practically comes down to that. Maybe model's are like baking a cake with a couple ingredients missing? If engine models were perfected, then GM, Ford, Honda, Dodge, etc wouldn't be spending multi-millions of dollars on dyno's and engine designs, because engines and fuel management would of been perfected in the computer software and all would have the same engine. Basically it would be nice to have Star Trek level of computing power, but all we have today is look up tables because so far millions of dollars (billions?) and decades of research haven't generated the perfect model for air flow and internal combustion dynamics.

Signed "Your Local Geologist and Car Enthusiast"

[pmuller9]

Edouard

What you are saying is appreciated and I understand your view point.
There is no personal insult taken.
The only reason for the rebuttal was to keep from blowing out readers that might consider EFI as a possibility.
You yourself are an advocate of EFI but you couldn't tell that from your initial remarks.

As far as your point, no matter where you are technologically you are always behind the possibilities and what tools you have today, even if they are the latest and greatest, become obsolete in months.
However that is not an issue here and I'll explain in a moment.

What you are describing is called the dyno room. Equipped with pressure sensors and laser tools and even transparent cylinder heads the poppet valve piston engine has been analyzed to death. From there we can see that certain indicators can be used to determine engine performance and that is what the present day EFI systems use. We also know that using those indicators gets us close enough to the optimal response that whatever is left on table is not worth going after.

I'm going to put this statement out in the open so you don't miss it this time.

We don't always want the optimal engine performance tune up and there are so many variations in engine response wanted by different users that there has to be a way of telling the engine management system what that desired response is!!

After that the present day systems get us to where we want to be and no matter what advancements there are in software you aren't going to get any closer because there ain't no closer to be had!

What the new technology does offer is the ablity to investgate and control better engine mechanical configurations which is the arena that is so far behind it.. (Add your favorite conspiracy theory here).
I don't intend to start a mechanical discussion here so please don't either. Thanks in advance.

Going back to software real quick. When code was written in machine lanquage it was very compact on almost never locked up due to open ends. I worked for GE back then. As soon as our software group started with higher level language they started asking for more memory because the higher level languages required libraries as part of the source code and took up a lot of memory space.
Sounds a lot like look up tables to me!

Paul

[S2X01]

Dumbed down.


I am not a phisics or chemistry major. I am not an engineer(although this fall I'm going back to school for that).
I'm just a dude who wants to find an early 50s Buick with no motor or trans, and make a cross country cruiser that can still smoke some tires.

I realize that a Q-Jet can do this for me, but part of me wants the satisfaction of accomplishment, and the other part of me doesn't want to tweak a carb every 200 miles.

I plan on building either a 455 or a Nailhead, and adding some sort of EFI. I AM BROKE! I am a laid off construction worker who is going back to school in the fall. So $1600 on an Accel system is not happening. I'm selling my Rivi to fund this project, so funds will be evtremely limited.

I was just turning to you guys for suggestions on how to efficiently accomplish this, on a budget, as well as creating a reference point to someone else in a similar situation.

I like the megasquirt system because I can build it myself and save a ridiculous amout of money. Also, since there are other people on this board that are experienced with it, that helps me a LOT.

I didn't mean for this to get out of hand. I just wanted simple suggestions and ideas to keep the brain working.

[pmuller9]

Stu

Actually it isn't the way it appears.

Edouard and I like to pick on each other in a friendly dual.

The hopes are that there was still a lot of usable insight within the content
of the replies from all who participated in the discussion.

As you can tell there are some frustrations because both of us are also in the middle of engine projects
that are on hold because of parts supplier issues.

So what better way to spend time than here on V8Buick.

Aside from the EFI purchase there is also the fuel pump, regulator and fuel tank as additional cost.
If it is a port injection then there is the cost of adding injectors to the intake manifold.

Focus on the engine first and possibly start with a carb setup, then as the budget allows go to EFI.
This will also let you compare the differences between the two setups.

What type of engineering are you looking at doing?

Paul

[bammax]

Every single time we talk efi on this site it goes to heck.

How bout this, new cars have way too many electronics and not enough brains. I can show you a fundemental flaw in "reliable" oem application that can make you question the sanity of the designers. Jump in an L99/LT1 b-body and try to drive it with a loose connection on the coolant temp sensor. The car will quickly alternate between a low rpm stall condition and lunging forward at max torque. With the brakes fully locked the car can actually surge forward as it outpowers the brakes. The issue is that the car is seeing a negative value on the temp sensor (defualt reading is minus 20 or so) and the car instantly goes full rich to try and heat the car up to allow good emissions readings. When it goes full rich the car launches forward. When the o2 reads full rich it cuts fuel and timing which makes the rpms drop to a point where the car almost stalls. This pattern happens about once every 2 seconds. Shifting into park will cause a pure power spike and may actually show you where the rev limiter is set.

Someone in the GM engineering department should have asked themselves why a coolant temp sensor would cause such a dangerous situation, especially when the iat sensor is reading normally.

I'm going out on a limb and saying that elagache is basically just saying that at this point in our technological history we should be able to count on a system that is designed and tested on our specific brand engine and we should also have faith that it doesn't have the same fundemental flaws that have popped up in the past. After 30 years of experience we're still at a point where we have to guess at a starting point and then hope we can adjust it to where it should be without running out of adjustment or blowing something up before we get it adjusted. How many of those aftermarket units can only raise pressure or flow a certain percent before the ability to tune is maxed out. That's a problem that should have been solved back in the tbi days, but it still pops up from time to time. How many systems use fuel pressure increases to cover up a fuel supply problem just so that they can say it fits both small and big blocks. That's an issue where the sales department has more clout than the engineering department.

I personally love efi and think it can do great things. I'm also in the process of working on both a '94 efi car and a '63 straight 6 and the less electronics an engine has the easier it is to track down shorts and troubleshoot slight problems. If efi was as basic as the old mechanical fuel injection systems, or even the bosch analog systems, than I believe people would be far more forgiving of the small flaws that are common on these new expensive systems.

[pmuller9]

Thanks Bammax

Aftermarket EFI is one of the areas that I feel has not caught on as fast as it should
and can cause a person to become overprotective of the subject.
My apologies for contributing to the fiasco.

The issues you brought up are certainly a contribution to the problem of making
the move to EFI something that a person can feel confident doing.

It would be great to be able to purchase a system for a Buick 455 configured closely to the specific engine build,
put it on and have it fine tune itself from there. The system would include the timing wheel and trigger sensor along with the cam sync
signal hardware. I added this because it is one of the main obstacles in installing an EFI on an older engine.

This shifts a lot of the responsibilty for the success of EFI to an engine specific forum like V8Buick.
What we need more of is specific installation examples and less debate.

A great example (and forgive me for not mentioning others) is TheSilverBuick.
The thread had continuing progress reports and pictures showing important detail.

If every time a new EFI installation was posted especially of different brands
then a person could see what was involved and could make a lot better decision
and feel a lot more confident.

And yes, the next Buick EFI installation I do will be posted

Paul

[TheSilverBuick]

Was an engine/system that hasn't been used in at least 15 years cited as new?

[GNandGS]

The crazy thing is we have this same argument in other forum about why points are preferred over elec. All very similar points made.

Fact is that EFI is proven as very capable AND flexible. Sure there are more point of failure but for the most part cars today are gas-n-go. No point debating if it should be that way...

The turbo regals were good examples of leveraging tech to make power. Coil packs supporting 4x the orig HP, distributor less, FI, overdrive, cheap HP gains without opening the hood, etc etc. Would not have been the same with carb and HEI !

Remember when transmissions didn't have computers? Sure, there is a downside to requiring a brain on the box, but look at what folks have been able to do with the 4L80. Can get with or without brain iirc

[doc]

Dang, Bammax dont look like he is that smart....:Brow::laugh::laugh:

[bammax]

I'm not that smart. Just poor and cheap and constantly fixing things. It gives good insight into the shortcomings of various vehicles. Like the light switch catching fire on the 80-90's T-birds, or the temp sensor fiasco on the Gen2 sbc, or the fact that an s10 blazer uses different rear glass shocks if it's a 2-door vs a 4-door because for some reason GM decided to use a different mount on the glass for the coupe version. I can also point out that GM mounted trans lines and vacuum switches right up against the floor on most of their cars and trucks, which means the parts can't be replaced without lifting the body or dropping the trans slightly. That's something people with new car warantees don't know :laugh:

I'm either looking for an efi system that is tailored and tested on the specific engine family I'm using, or a system that needs some work (universal fit) but is at least safe to start with and affordable to buy. Right now all that's out there is universal fit parts at $2-3 grans and then need to be professionally tuned at the cost of a few more hundred. You wont see the fuel savings being worth the initial expense until you put 200k miles on the engine. So until one of those dream systems comes along I'm still going to keep trying to figure out which combination of junkyard parts can get the job done safely and affordably.

I'm still leaning towards a GenVII 454 setup with the 24x trigger wheel and a vortec van cam sensor retrofitted to the 455 distributor

[TheSilverBuick]

Any ECM worth it's salt doesn't ever need to be professionally tuned. No one but me has ever tuned my Skylark or T-bird. I put a ton of miles on them and am pretty satisfied with the results. Have I mentioned I'm a Geologist not some computer or engineering whiz? I can get good drivability and mpg's without a professional. For max HP, that requires track time and/or a dyno. Professionals just make that process faster, which of course is where the cost is. Same deal for a carb isn't? Lots of people screw up the simple bolting on of a carb, and lots more rarely ever get them optimized for drivability, mpg's and max power. If you can tune a carb, you can tune a programmable EFI system (even easier).

I really should just EFI the 455 in the Centurion already. I think I can get the MPI conversion down to less than $800. I only have around $800 in the Thunderbird's TBI system. The only expensive thing on my Skylark related to EFI was the Buick specific part, the SPX intake. It cost as much as the MegaSquirt, wiring harness, WBO2 sensor/controller and relay board combined. I plan on drilling my own intake for the Centurion and using a junk Q-jet for a throttlebody.

[Jim Blackwood]

Here's what I'm thinking, and I guess the auto-tune feature on some of the newer efi brains sort of do this. You've got all these lookup tables and some of them can get VERY confusing (Like trying to tune a Ford system using the TwEECer for instance), but as an example just taking the fuel map since that's usually where it all starts, or maybe a better example is ignition advance since the general requirements are pretty specific and we aren't dealing with a variable like injector size or pump pressure.

Some things we can agree on for any application. Advance for easiest starting is never going to be outside a range of, oh I dunno, just for giggles lets say 0 to 10 degrees. So why do we need to put that in there? Shouldn't it automatically start out at the average best figure and we can move it if we want to? Advance needs to increase with engine speed following a curve that gives best economy under light throttle. This curve is pretty similar for most engines. Shouldn't it start out as an average? And when we change it by for instance dragging a point on the map, shouldn't the map automatically blend with the surrounding areas of the map instead of us having to manually change each and every surrounding data point? It's a computer for chrissake, why can't it do that job? For that matter, why are we specifying target AFR's? Like that's inherently understood or something? No, it is not. Why can't the danged old computer take the throttle opening and other load factors and figure out for itself what the AFR should be for that speed and load? I sort of agree with Edouard on this one guys. If you've been born into tuning with EFI maybe it's no big deal but I gotta tell ya, this stuff is way more user unfriendly than it ought to be. And he's exactly right on his analogies. I diddled around with DOS, and IBM punch cards before that. When you have to spend 3 days making a stack of cards 3 feet high just to get a readout that tells you the time it's enough to make you pull your hair out. Now you just glance down at the lower right corner and there it is. With EFI we're somewhere in the middle. Not punch cards, but certainly not a full featured desktop by any means. It's not a question of putting the studio on the desktop, it's a matter of getting the machine to do all the hard work and computations in the background so that all I have to do is click a button on my shifter when a problem crops up, let the computer analyze it, tell me the cause and give me options for fixing it. WHY in heaven above would I want to do the equivalent of punching cards to have the luxury of doing all the analysis and computations myself? Answer: I don't. What I want are results, and he's right. There's no good reason why the computer could not be set up to do just that.

Call me ignorant if you like, but for one given scenario, how many options do you really have? Leaner or richer? More or less advance? A few more peripheral choices and then a few special situations. Can you really say that the computer is incapable of eliminating the least likely candidates correctly? I say it is capable of doing it more accurately and consistently than even the best tuner around once it's set up to do that job. Even if it can only eliminate a few reliably and then gives you a choice of the rest, doesn't that make it easier? And that bit about smoothing the map. That at the very least should be something that is in every package, should be automatic, and adjust to the way you drag your map. JMO. Of course, I'll still take what we have now over a carb any day, but that's a far cry from saying it's good.

Jim

[GNandGS]

This is available now. You may not see it so common because often folks use them for performance where its on/off. Closed loop feedback for fuel and timing is already available and most systems either come with or can be provided with a "safe" start base program.

So if you want to give it a target fuel ratio and let the system do the work - you can. Want timing backed off automatically or change when the air temp changes - you can. OEMs have even allowed now for years systems that learn driving habits for shifting behavior and timing (but only within a range).

The critical comments in this thread would be gone I guess if we hade a fully automated Buick specific system that did everything on auto but with a full manual mode available. But such a specific setup is not needed nor will it ever happen at a tempting price point. You will NOT be able to get away from lookup tables as you have to have a ref point or target. "Tables" has become a bad word? The same process is used to tune carb and points. If THIS then do THAT. Every once in a while a new idea comes along and experimentation begins until you find what works for "THIS n THAT"

If you are capable of tuning a points and carb setup you ARE capable of tuning EFI, the providers can walk you through the interface. However the same issues apply to preferences. Some love the Holley while others swear by the Demon or QJet. Windows vs Apple. etc. Those that learned on EFI may have a harder time going old school IMO but I bet this forum could sort them out. The key is to have known good stuff to begin with. If you have worn throttle shafts, bad fuel pump, mismatched parts, or the wrong cam it can be just as frustrating to sort as an EFI setup that is faulty or mismatched from some homebrew mashup.

Also, if you are able to screw up traditional tuning EFI won't save you. I can screw up both and know just enough to be dangerous!

[Jim Blackwood]

I don't think the point is to make it an auto-tune with manual override, I think the point is to simplify the interface and make it more human-friendly. Maybe the way to do that IS to use multiple tuning levels, starting out with a simple overall interface that automates almost all of the tuning functions and have a way to "drill down" to more in depth tuning parameters.

As noted elsewhere, the brain could care less that it's a Buick, just like a carb. So there's little need for a car-specific tune. An average generic baseline tune would run 99% of the engines it's ever connected to, and automated routines could fine tune that in reasonably short order to eliminate the worst behavior using the sensors it was equipped with. More sensors equals a better tune, and drilling down could take care of the special cases. In terms of what the tuner has to do, isn't that exactly the way we do it with carbs? So why does efi have to be so complicated? Especially when we have a computer we could use to simplify it.

I realize this still means inputting parameters like injector size and engine displacement, but just using a generic start-and-run routine, how long would it take for the WB-O2 to tell the brain how much fuel is needed? For that matter, does the brain even need to know how big the engine and injectors are? No it does not. All it cares about is pulse width, O2 level, throttle opening, egt, and things of that nature. The engine itself could be big or small and it doesn't matter in the least.

Jim

[GNandGS]

Sorry forgot winky face. If we had a perfect Buick only setup we wouldn't have a topic :)

[bammax]

Not to stir the pot but I wonder who's going to be the first to try out the msd tbi setup and let us know how well it's programmed.
http://www.atomicefi.com/home.aspx

I'll never consider tbi to be a worthwhile upgrade over a well running carb, but the programming could be harvested and used in a multipoint setup. At the very least it would swap to a 4 cylinder multipoint directly (assuming 4 injectors in the tbi unit) or could be used just for the tables and charts and plugged into the programming for the system of your choosing.

[TheSilverBuick]

If you are capable of tuning a points and carb setup you ARE capable of tuning EFI, the providers can walk you through the interface. However the same issues apply to preferences. Some love the Holley while others swear by the Demon or QJet. Windows vs Apple. etc. Those that learned on EFI may have a harder time going old school IMO but I bet this forum could sort them out. The key is to have known good stuff to begin with. If you have worn throttle shafts, bad fuel pump, mismatched parts, or the wrong cam it can be just as frustrating to sort as an EFI setup that is faulty or mismatched from some homebrew mashup.

Also, if you are able to screw up traditional tuning EFI won't save you. I can screw up both and know just enough to be dangerous!

Amen!

As one guy I know put it, "A carb has a few variables that each change a whole lot and EFI has a lot of variables that each change very little." AFR and timing tables can't be defaulted with out leaving a lot on the table for improvement. You'd never (want to) enter enough variables into the EFI program to make a perfect tune. An LS based 454cid engine versus a Buick 455 engine for example, you'd think would be easy to just set a good "average" AFR table and ignition table and everything would be good, but the reality is that the LS engine will tolerate and make more power at a leaner AFR and far less timing advance because a better cylinder head design, better intake and exhaust manifold flow, and that's assuming the cam specs are the same. Likewise change the displacement to a 4-cylinder with a narrow bore and the timing requirements can be radically different. Knock sensors sound like a good idea except max power isn't always just short of the point of detonation and a lot of big cam'd engines make too much valve train noise for a knock sensor to work.

To sense load the program would have to know the actual load on the vehicle, which is difficult to really measure and predict the exact fuel and spark requirements at each load/rpm point. Close works, and makes a fine running engine and that's how 90% of the self tune systems work, but they are always leaving something on the table because the specifics of the vehicle and engine combination. OEM's do a lot of dyno loading of each engine for each model offered, even if the engine is offered in a different vehicle, to simulate dozens of load conditions and then figure out how much fuel and timing are required at each condition.

Anyways, I figure if it was easy it would of already been done and have the market cornered.

[TheSilverBuick]

Not to stir the pot but I wonder who's going to be the first to try out the msd tbi setup and let us know how well it's programmed.
http://www.atomicefi.com/home.aspx

I'll never consider tbi to be a worthwhile upgrade over a well running carb, but the programming could be harvested and used in a multipoint setup. At the very least it would swap to a 4 cylinder multipoint directly (assuming 4 injectors in the tbi unit) or could be used just for the tables and charts and plugged into the programming for the system of your choosing.

I know one guy looking to try this one. I am waiting to hear how simple the installation goes. That's the advantage to that particular system, it's very integrated to keep wiring and such very simplified and for people who want the carb look but better fuel control. That's the market niche MSD is going for.

[pmuller9]

Is this discussion being limited to just gasoline and no power adders?

Paul

[TheSilverBuick]

I'm game to hear what's on your mind about power adders :beer

[Loyd]

Since I do not like carburetors, and timing adjusted by vacuum modules and/or springs and weights, I put a FAST XFI system providing sequential fuel injection control on a 455 engine in a 1970 GS equipped with a 200 4R transmission with lock up torque converter. The sensors include a one bar Manifold Air Pressure sensor, an MSD crank sensor, the distributor was modified to also be a cam sensor giving one pulse per two engine turns, the carburetor replaced with a Edelbrock 1000 cfm throttle body, and the manifold modified to incorporate eight 43 lbs fuel injectors located over each intake port with an Aeromotive fuel regulator controlling the fuel rail pressure. The fuel tank was replaced with Spectra EFI type internal pump tank made for A bodies with the GS filler neck, and the rubber sections in the stock fuel line replaced with high pressure fuel hose.

I love driving this car and would not even consider going back to the 1970 technology the car came with. Yes the incremental difference in fuel economy will never pay a fraction of what these components cost, but the satisfaction of getting this project to a point where there are few tweaks that need to be done are reward enough. Being able to go out and just turn the key to start to car is priceless, and not stalling when pulling out of the garage on a cool morning is worth it's weight in gold.

I'm at the point where I would trust this car anywhere. This conversion takes time as there is not a kit so it took time to research the parts needed and how to hook them up to make a working system. The FAST supplied engine wiring harness needed slight modification and many sets of wires and relays were added for fuel pump control, air-conditioning control, electric fan control and of course the MSD 6 box. There was a separate fuse block added to keep the loads from the new devices off the 40 year old wiring harness.

The only regret in this whole process is that it could have not been done, with results this good years ago. The advances in aftermarket Electronic Injection Technology has come a long way, and this offers a very flexible route to make the fueling and timing what you want and it needs, forgetting the compromises the automotive manufactures had to make. Having traction control through the XFI box is worth it by itself.

Therefore I make the case it is worth it for those who want to venture into this realm. It does make the car more complicated, but again I really hate carburetors.