Thanks, that will be a big help. I will do the same.

 

OK. After researching several sources, I have come up with the same ol' thing, Lots of opinions.

The general feeling seems to be that the injector should fire early in the intake window. I am using the term "intake window" to represent the point in time that the intake valve is open. I made it up myself, so if there is a better term, just let me know.

Some say (Toyota) that the entire injector charge should be injected into the port just PRIOR to the intake valve opening. SAE, however says that wetting the port walls is bad. So there are others that say to begin injection just as the intake valve opens, to prevent wetting the port walls. They all agree, however, that the pulse should be complete by the time the intake valve closes.

Another agreement was that the injector should spray the fuel on the back of the intake valve if possible. This flashes the fuel to vapor, increasing atomization.

As with anything, the truth is probably a compromise. I would guess that as long as the injector is pulsing when the intake valve opens, and is done before it closes, everything would be fine.

The problem with this level of fine tuning is consistency. How fine is the valve timing with hydraulic lifters? How much does valve timing change between the times you lash your valves on a solid lifter engine?

I wonder if it would be wise to estimate your full throttle pulse width, and put your intake valve opening right in the middle of it, assuming that you can fit it in the intake window that way.

The other thing that pops to mind is just how much intake charge is introduced to the cylinder after the valve is back below .050 lift on it's way to closing. Maybe it would be better to keep the injector closing event prior to the intake valve falling below .050 lift.

Along those same lines, would it be better to have the IPW finished by max lift?

I realize that I have created more questions than answers, but I guess that's part of coming to a conclusion.

If you have found anything interesting let us know.

-Scouder

 

Scouder,

Wetting the port walls will give you bad emissions, as well as a not-so-complete burn, I would imagine. Anything that drops fuel out of atomization is going to hurt in power or emissions. But, if you're racing, it might not be such a bad thing at high RPM's, but then who cares about emissions in that case?

Not sure if vaporization of the fuel mix is that good an idea, I thought it was bad to heat the mix?

As to timing the fuel charge with the intake valve, I'd think you'd be trying for that imaginary point where the air is flowing into the cylinder fast enough that all the fuel you spray goes into the cylinder instead of hitting the bottom of the port - with the longest duration possible so as to get 100% of the mix mixed, and I can see the point with varying intake valve timings based on hydraulic or sloppy solid lifters, but I think you can agree that you wouldn't have more than .100" of difference in valve openings? If so, fire the injector at that .100" lift - BTW, that's where the @.050" comes from for cams, the imaginary point where it actually starts to flow something worthwhile (not taking into account radical cams that fire exhaust back up at the fuel injector

I think you COULD estimate the pulse width required at full throttle/max power by taking volumetric efficiency, the stoichometric thingy (optimum fuel-to-air ratio), and the size of one cylinder. What that works out to for, say, a stock-bore 390 would be interesting.

Another thing to think about - you can compute the amount of fuel it takes to charge one cylinder, and then ELONGATE the fuel pulse as far as possible (lower fuel pressure, smaller injectors maybe) to make sure as even a mix as possible. If you have a pulse width of say only 20 degrees of crank rotation and a big injector/high pressure, you'll dump a lot of fuel into only a small portion of the air going into that one cylinder. So, running too much pressure or too big an injector can get you in more trouble than it's worth. (I'm theorizing, totally...) Unless you have a LOT of swirl going into the cylinder (almost non-existant in an FE) and even then, you might wind up with a less-than-optimal mix.

art k.

 

Art,

I am trying to remember exactly what the wet walls theory was, but it seemed to have an effect on flow if I am getting it right. They didn't mention emissions.

It was also mentioned that spraying the fuel on the back of the intake valve actually increased charge density. So apparently there is a point that the vaporization of fuel reduces chamber temps such that an increase in density occurs.

I see what you are saying about timing the pulse with the incoming air flow, but I don't have enough knowlege to know if that is possible. I have seen of several websites that IPWs in the 2 to 10ms range are typical. Well, if the intake window is 4ms (6000rpm w/pretty hot cam) and our IPW is 10ms, it is obvious that we have to "get a run at it", by starting the IPW before the intake valve opens. This may be why the Toyota website that I found said to finish the IPW just as the intake valve opens. In fact, the MegaSquirt appears to have an IPW maximum of 25ms.

I am sure that there is some tuning that can be done with fuel pressure and injector size, but like I said, I'm new to this, so I don't know how much.

-Scouder

 

Art, you want cold air mixed with warm gas - you don't heat the mix if you can avoid it. Cold air because it is more dense and you can pack more of it into the cylinder, warm gas because it atomizes better, giving you a more even mixture. I can not believe that the two would be in contact long enough to transfer much heat from the gas to the air.

And yes, you set your base pulse width by finding the max volume of the cylinder and using your desired air/fuel mixture to solve for the amount of fuel needed. Knowing how much fuel your injector sprays per unit of time you know how long it needs to be turned on to deliver that much fuel. I will do the math tonight and post the answer.

But we disagree on the "stoichometric thingy" - that is a compromise rule from our government which yields the best NOX emissions. For economy you want to run leaner, and I would expect that under full race conditions you will want to run richer. I hope to modify the program to offer different modes for different driving conditions - an economy setting, an emissions setting, and a race mode.

I like the term "intake window", but you should recognize that it is a sliding window. If we decide that our injectors will be on for say around 2 milliseconds we will want them on at the best time - perhaps when the valve opens .100 inch when you are turning 6K rpm, which will cover most of the time the valve is open. But when the engine is running only 2000 RPM that time sequence will not even get us to full lift on the cam.

We will need to slide that 2 ms window later in time, perhaps bracket the time that the cam reaches full lift.

I understand about opinions - there is even one apparently knowledgeable programmer on the MegaSquirt board who thinks the best time to squirt is during the exhaust cycle so that the fuel is already there and waiting for the incoming air when the intake cycle occurs. Indeed, the MegaSquirt batch EFI works and has made a lot of people happy.

I have ordered two textbooks on this subject (one written by a Ford Engineer) and will post anything helpful when I receive them.

Clint

 

"I like the term "intake window", but you should recognize that it is a sliding window."

Absolutely correct. I am going to expound just a little, for my own sake, to make sure I have it right in my head.

I would say that the intake window is dynamic. Although it is linked to rpm, and therefore static, it does get shorter as RPM increases, so I have taken to thinking of it as dynamic. At 750rpm, with a 288 adv duration cam, we get around 32ms of total intake window. At 1500 it's 16ms, at 3000/8ms, and 6000/4ms. So obviously the window itself gets much more forgiving at lower RPM.

Now to what you said, there is a second component, and that is the "sliding" injector pulse (IP). It has to both change in duration (width), and change in relation to the intake valve opening (timing). It only makes sense that the timing would slide in relation to the intake window. If we are at 750rpm, and need an IPW of 2ms, it seems logical that we would retard the IP to about 15ms (roughly) into the intake window (IW from now on, to save whats left of my carpal tunnels), to take advantage of the peak air flow into the cylinder. But at 6000rpm where the IW is only 4ms, and the IPW may be as much as 10ms, we would actually have to advance the IP out of the IW by as much as 8-10ms just to get the charge in before the IW closes.

Is this what you meant?

I also have noticed that the batch firing systems do indeed seem to work fine, and they are firing the injector all over the map depending on which cylinder you look at. It makes me wonder if sliding the IP around is worth the trouble or not. Can it even be done? I guess if your trigger marks the earliest possible IP, then it's just a matter of mapping a delay value depending on rpm and the required IPW. A software thing.

Let us know what you find when your materials arrive. I think this topic is worthy of more discussion, as it appears to be an emerging trend.

-Scouder

 

You understand just what I was saying - except that I think we can juggle the injector size and fuel pressure to deliver enough fuel fast enough to get it all in during the intake window. If need be, we can add a second injector for race mills like yours.

But you have the idea - just pick a time either at or slightly before the earliest time you want the injector to fire, and tell the computer to fire the injector at that time plus a delay - said delay to be controlled by the RPM of the engine.

 

Sounds like you guys are 90% of the way there just from this one thread

If I am repeating anything you guys said, just chalk it up to one more opinion. A theoretical opinion, though...

I'm thinking about the sliding window - I think at max RPM (whatever that is), you want to get as close to the END of the intake event with the injector pulse. At low RPM, you want to get as close to the BEGINNING of the event as possible. Then modulate the length of the pulse according to throttle plate angle and RPM, with a cold-start/run setting (manual?)

This is all depending on whether you can realisitically charge the cylinder with a single pulse from the injector, still waiting on those numbers

As to the stoichometric thingy, yeah I guess you're right, that's an emissions "sweet spot" not a real-world situation. I know there's a range that's a good theorietical starting point, I remember reading that somewhere...

Looks like the only way you're going to find out is to build it

This is getting me started thinking about doing this exact same type of thing but on my snow blower first ... I wonder if I can run my V10 on nine injectors/coils?

 

I hadn't considered two injectors. Interesting idea. I calculated my total demand at around 48lbs/hr, so I can probably pull it off with one pretty easy. This is beginning to interest me more and more. I looked up injectors at less than $500. Throttle body looks pricey, everything else looks pretty reasonable.

Truthfully, at the point that I rebuild this engine the first time I am going to stick a supercharger on it. I am kicking myself for not doing it the first time since I'm at 7000'. That's when I'll probably switch to SEFI.

-Scouder

 

Ford Racing has a 52 lb/hr one good for 106 horse per cylinder - about 850 horses for a V8. My guess is that I will need a 24 lb/hr one!

 

I seem to recall reading where manufacturers were toying with/using multiple injector pulses per "IW" as Scouder calls it. The theory is that to increase economy, the entire cylinder A/F charge didn't need to be equal. They would use a relatively short injector pulse earlier on to provide a lean A/F mixture to the cylinder, then late in the IW would inject a shorter pulse to create a smaller, richer mixture near the spark plug. So when the plug fires, it ignites a small rich mixture that in turn ignites the leaner mixture and promotes a more even flame front and better combustion.
Now this seems to make sense to me, but I can't for the life of me remember where I read it or if it even applied to port injection or if it was in some of the newer direct injection engines (similar to diesel engines).
I would think this might be the way to go if you were going to use 2 injectors per cylinder. 1 smaller charge early on, then another nearer the end to squeeze out the best economy.

 

Sounds good to me, except I do not think we have enough outputs to run two injectors per cylinders independantly. Also, Scouder has a full race mill, his worry is getting enough fuel at high RPMs.

 

Egad. Supercharger and nitrous? That thing'll flip over for sure!

 

That's a lot of great information. In reading your posts, one thing that springs to mind is the ability of the computer to compensate for the variables mentioned in the intake timing events. Given a computer with enough power, running in closed loop mode (using a wide band 02 sensor) you could test these theories by adjusting the pulse width and pulse timing.

Fuel pressure is another tuning variable that the racers are using to great advantange. The base fuel map can be "richened" across the board simply by turning up the fuel pressure. There are boost reference fuel pressure regulators available, it would be a relatively simple (I hope) to add a computer controlled regulator that would increase pressure in realtionship to the demands as rpm increase, thereby helping to compensate for the shorter "intake window". There are many other possibilities for computer control in fuel injection. Some are so obvious, it makes me wonder if they had already been tried and found to be of minimal benefit, or if it is lack of computing power or increased complexity.

Keep up the research guys, your about to push me off the fence to start playing with EFI.

Hey Scouder, how's the recovery coming??

 

"Keep up the research guys, your about to push me off the fence to start playing with EFI.

Hey Scouder, how's the recovery coming??"

I would think that of all of us you would be among the most qualified to "play" with EFI. It is undoubtedly a trend that will continue to grow, so I think these discussions are appropriate.

Today was the first day I have left the house since last Thursday. I jumped in the car with a buddy and went to Salt Lake City to look a some Corbeau seats. Unfortunately I'm too friggin' big for all of their "normal width" seats, and none of them are built for guys over 6'. I want the shoulder belt mounting points to be above my shoulders, but all of their shoulder belt holes are too low in the seat. Whichever seat I pick, I'll have to pass the shoulder belts over the seat instead of through it. I'm pretty sore now, but getting better.

I also bought a $49 pedestal mounted manual bender from Harbor Freight. I'm gonna get some tubing next week and see if it's gonna work. I'll keep you guys posted in case anybody else decides to bend some round stock.

-Scouder