EFI straight eight Buick

[panic]

That casting does look like it's got a dimple separating the pairs, but...?
I couldn't find one of the bare gasket surface.

[panic]

The firing order is 1-6-2-5-8-3-7-4, yes? So, the separation between any pair of intake cycles is only 180°.
If the exhausts are joined in 2 groups (front & rear), as common the separation looks like:
Front 1 - 180° - 2 - 270° - 3 - 180° - 4 - 90° - 1  (total 720°)
Rear 5 - 90° - 8 - 180° - 7 - 270° - 6 - 180° - 5  (total 720°)
Both as bad as a bank-separated V8 (in fact, may help with a turbo by increasing the pulse effect).
You could join them for equal intervals:
#1 = 1 -  2 -  8 - 7, and #2 = 6 -  5 -  3 -  4 (all 180°)
but I don't see it as worth the trouble considering the primary length is huge.

Again, I would turn to Vizard's work on th Mini for how cam timing can compensate for irregular and overlapping intake strokes sharing a port.

[4sfed]

You must inject only on the open valve. 

Sorry, but that's not true at all.  There's no such thing as any port EFI system injecting only when the valve is open.  Think about it, an intake valve is open *at most* during 25 % of the combustion cycle, meanwhile at WOT most EFI injectors are running 80-100% duty cycle at WOT.  3/4 of the time an injector is open, the intake valve is closed and on OEM applications, it's even worse than that (conservative cams for torque and emissions).

Emissions at idle and warmup are really the only reason sequential EFI exists...  if what you are saying is true, then how would a batch fire system work on *any* port injected engine?  or TBI injection?  that's how EFI cars were run for a decade until emissions laws began dictating "warmup" emissions be improved, and that is where sequential EFI shines - at lowest duty cycles, during idle warmup.

-scott

You're right . . . there are no production port EFI systems injecting only on an open valve . . . but there are no siamese port EFI engines.  My point is that injecting on an open valve is the only port injection that's possible on this straight 8 engine . . . even more restricting is injecting only when there is no overlap between the two intake valves in one port.  

Jim  

[dieselgeek]

You're right . . . there are no production port EFI systems injecting only on an open valve . . . but there are no siamese port EFI engines.  My point is that injecting on an open valve is the only port injection that's possible on this straight 8 engine . . . even more restricting is injecting only when there is no overlap between the two intake valves in one port.  

Jim  

I see your point.  You're concerned that all the fuel for one "squirt" intended for 2 cylinders' quantity of intake air, would get sucked into one cylinder.  Wouldn't running "4-squirts alternating" or "2-squirts simultaneous" cover this?  (megasquirt gives you the option of configuring the number of injection events per engine cycle) and on top of that, only half of his fuel will be supplied by the port injectors.  THe other half of the fuel is supplied by upstream injectors, squirting through the blower...    add to that, it depends how far upstream his injectors are from the intake valve too.

[4sfed]

I see your point.  You're concerned that all the fuel for one "squirt" intended for 2 cylinders' quantity of intake air, would get sucked into one cylinder.  Wouldn't running "4-squirts alternating" or "2-squirts simultaneous" cover this?  (megasquirt gives you the option of configuring the number of injection events per engine cycle) and on top of that, only half of his fuel will be supplied by the port injectors.  THe other half of the fuel is supplied by upstream injectors, squirting through the blower...    add to that, it depends how far upstream his injectors are from the intake valve too.

Ignoring for the moment, the upstream injectors, the squirts still have to happen on an open valve or they will accumulate and all go into one cylinder.  I did a little searching and found an article that explains the problem fairly well . . . citing one production   example . . . British of course. 

http://www.starchak.ca/efi/siamese.htm

Jim

[dieselgeek]

I read the link - twice - and I am confused.  What was his result?  the guy discusses what he theorizes will be a problem, in a different application...  the post is 8 years old - are there updates?

I can see some unbalance at idle and low RPM.  But at high throttle openings, high duty cycle on the injector, wouldn't the fuel suplpy in the incoming air be the same as with a carburetor? 

Pardon me for not understanding, but so far I am of the opinion that this is an "over-engineering" discussion.   My money is still on the engine running happily, especially since 50% of the fuel is coming from injectors upstream of the port injector...

Guess we'll have to see, eh?

[RichFox]

A lot of Siamese intake port engines have run in race cars both injected and carbureted. V4 engines like mine. Stock head GMC engines like mine. And such. Doesn't really make much difference if they are carbureted or injected except for the larger plenum. The lead cylinder 2 and 5 will get a better charge then 1 and 6. But people have been living with it for 100 years now.

[maguromic]

Like some of the LMP cars and some of the modern road cars, could you solve this with a direct port injection right into the chamber?  That is if its not going into a vintage body.

[4sfed]

I read the link - twice - and I am confused.  Pardon me for not understanding, but so far I am of the opinion that this is an "over-engineering" discussion.   My money is still on the engine running happily, especially since 50% of the fuel is coming from injectors upstream of the port injector...

Guess we'll have to see, eh?

I made a timing chart and discovered that the straight 8 will only need 4 injectors if they are timed 90° apart . . . but the timing and length of the pulse is critical.  I ignored the upstream injectors in the analysis because they work more like a carburetor, supply an A/F mixture of basically equal strength to each valve.  The port injectors need to deliver equal amounts of fuel to each cylinder for the engine to run properly.

The first two charts show intake air flow and fuel flow for a properly timed engine . . . first at low power, then the second showing a longer pulse width.

The third shows what happens when the pulse width gets too long . . . the #2 valve closes before all the fuel is delivered.  The fuel puddles and enters cylinder #1 along with all the fuel from the next injection pulse.

In the fourth example, the pulse width is okay but starts at the wrong time.  All the fuel from the port injector enters cylinder #1 and none into cylinder #2.

[Milwaukee Midget]

The port injectors need to deliver equal amounts of fuel to each cylinder for the engine to run properly.

Good homework, but it raises a question -

Given the limitations of the design, wouldn't you want your intake charge fully atomized before it actually enters the port?

I don't think you want equal amounts of fuel per cylinder, but equal air-fuel ratio per cylinder.

Even with a scatter pattern cam, it's a tall order getting equal cylinder filling between two cylinders sharing an intake port.  The best you can hope for is maximizing each cylinder individually.  Port injection would just exasperate the problem. 

I think Dieselgeek might have the right idea in moving the injector a little further upstream, but maybe just running TBI might be the simplest solution.

Chris

[4sfed]

Given the limitations of the design, wouldn't you want your intake charge fully atomized before it actually enters the port?

The fuel will atomize sufficiently as it enters the cylinder . . . it has been warmed by sitting on the hot intake valve, and has to pass through a narrow opening as the valve cracks open.

I don't think you want equal amounts of fuel per cylinder, but equal air-fuel ratio per cylinder.

Even with a scatter pattern cam, it's a tall order getting equal cylinder filling between two cylinders sharing an intake port.  The best you can hope for is maximizing each cylinder individually.  Port injection would just exasperate the problem. 

I think Dieselgeek might have the right idea in moving the injector a little further upstream, but maybe just running TBI might be the simplest solution.

Chris

Of course, equal A/F ratios is the goal.  In the case of siamese ports, the normal batch-fire scheme (with no attention to injection timing) has little chance of success.  TBI is a simple solution, but will give only approximately equal A/F ratios because of fuel drop-out due to differences in runner lengths and bends . . . and that will vary with throttle position and RPM. 

Port fuel injection has the potential to provide a better distribution of A/F ratios to the various cylinders, but it will take some serious development.  Fortunately for us, only a narrow RPM range at wide open throttle is really important.  That makes the job a lot easier.

Jim

[GH]

Boy, that's a lot of engineering. Just put it all together, make 8-10 psi boost, WOT to 5500 rpm, I'll bet it will run better than before, 144 mph. We'll see this winter when we run it on the chassis dyno. Cya

[Stainless1]

Boy, that's a lot of engineering. Just put it all together, make 8-10 psi boost, WOT to 5500 rpm, I'll bet it will run better than before, 144 mph. We'll see this winter when we run it on the chassis dyno. Cya

And for sure when you put it on the big white dyno... good luck with your project, have fun,
See ya on the salt 

[GH]

Stan Back, when I bought the 320 ci engine from the salvage yard it didn't have the manifolds on it, so I can't help you on the exhaust manifold design. The engine was out of a 1948 Buick Roadmaster, paid $85.00 for it in 1999.

[dieselgeek]

4s fed, what are your thoughts on multiple -squirts per engine cycle?   (not sure if you've read all the megamanual) but one thing that might help us here is, since this engine only runs to ~5000rpm, we have enough time to get "2 squirts" per cycle. 

or it's possible we could just feed all the fuel upstream of the blower, since the manifold is a wet design anyways.

comments/thoughts?