Injection Timing

[Whittey]

A couple questions on fuel injection...

1) How long of a pulse width is 'normal' on an injector?
2) When (compared to valve timing) is optimum to start the injection?
3) When valve open time is less than pulse width, is this very bad?

A quick bit of math shows that at even a modest 5,000rpm on a 280 degree cam the valve is open for 9.3333ms. This seems like a very short time and when i'm seeing some of these smaller n/a motors running over 10,000rpm I kinda wonder how they manage.


-=Whittey=-

 

[patdaly]

Whittey:
At least on the GM Injectors I am familiar with at idle you see somewhere between 1.7-2.9 MS open values. They increase the Injector on hold times to add fuel, but I am not aware of specific time values, however I would say they have to range to 20-30 Ms or more.
As for the start time, I will defer to someone more knowledgeable, but unless you are looking at it from an emissions standpoint, or pushing for the very last ounce of HP, I would say it isnt as important as you think. Most pre-OBD2 ( US only )Injection systems were Batch fire ( Bank to Bank ) systems, so naturally you would be spraying fuel even when you did not have the valve open. Finally I would say that having the valve open should not pose significant problems, with the exception of perhaps emissions.

 

[GregLocock]

Pat's right, at full power the the injectors are open almost continuously. We made a big hoopla a few years back when we went to sequential fire for injectors, it might have given us 1 or 2%, but do I know that at speeds up to 3000 rpm, and up to max torque, it made NO difference at all.

Cheers

Greg Locock

 

[schmidtj86]

Let me clear the injector on time up for you.
What you have to look at is:
Number of injectors / cylinder
Flow rate of injectors
Number of times injector fires per intake stroke
Power / Torque per cylinder
For each intake stroke there is a certain mass of air taken in, and this depends on the VE and displacement.
Fuel injectors flow a certain mass rate of fuel, so to get the right amount of fuel mass, you must turn the injector on for a certain amount of time.
The amount of fuel mass desired is related to the desired air / fuel ratio.
The VE depends on the manifold pressure (MAP) and RPM.
When you match the injectors to the engine, you have to look at the BSFC of an engine, i.e. the amount of fuel per horsepower hour (roughly .5 lb(fuel)/(Hp*hr)) and the maximum duty cycle (on time vs off time) (85% is a good number), and the horsepower.
Let's put some numbers in the spreadsheet I made:
5.7 L, 8 injectors, .45 BSFC, 1 injection/intake stroke (sequential), 22 lb/hr injectors, 240 hp @ 4500 rpm @13:1 A/F, and I want the pulsewidth for 3 different scenarios
Idle (34 kPa) with 14.7:1 A/F = 4.9 ms
Part Load (50 kPA) with 14.7 A/F = 7.2 ms
WOT (100 kPa) with 13 A/F = 16.3 ms
If these were fired in "batch fire," the pulsewidths would be half, but fired twice per intake stroke.
If the same engine had 36 lb/hr injectors (way over sized),
it would look like this:
Idle (34 kPa) with 14.7:1 A/F = 3 ms
Part Load (50 kPA) with 14.7 A/F = 4.4 ms
WOT (100 kPa) with 13 A/F = 10 ms
and if it were batch fire again, the individual pulsewidths would be half.
When the injected pulsewidth gets below 1.5 ms, like idle in batch fire mode, there is a greater chance of it getting into the "non-linear" range, which closed loop operation has a hard time dealing with.
When the injected pulse width starts reaching the time allowed between intake strokes (RPM dependent), the coils in the injector start to not have time to cool, because there is current flowing through them almost all the time, and they need the off time to keep alive.
But, at 7500 RPM there is still 16 ms between intake strokes.
You may be thinking, wait a sec, what about those high rpm scream machines, how do they do it? Well if you look at the units closely, you'll see that if you combine RPM and Hp, you can get Torque. Well that is the actual thing that is of real concern, it's just that most people think of BSFC in relation to Hp, but when you look at each injection individually, you are injecting for torque only. The dependence on RPM suddenly goes away. So, as long as the injectors are properly sized for the Hp and BSFC of the engine, then, there is NO relationship to RPM. Just don't always expect a decent idle without doing some tricks.
If this hasn't totally confused you, then, congrats, you can make your own spreadsheet and make a base fuel map of your own. Or, if you give me all the specs I'll do it for free.

 

[patdaly]

Schmidtj86,
Where were you when I was stumped getting my 83 Lb Injectors to idle......LOL, just kidding, but you had better be careful offering that service, especially with all the aftermarket EFI units popping up, you could stay very busy!

 

[Whittey]

Pat: I was thinking of sequential when I wrote the initial post. As for the valve open, I was thinking that valve-open would be the best time to inject. I can't think of a reason why it would make little/no difference (as Greg pointed out).

Greg: Do you have any idea why there is little/no difference with sequential as opposed to batch?

schmidtj86: Actually, I do understand what you're saying and I actually can't believe I didn't think to just do the math myself before. Thank you.

Being as the injector pulse width is longer than the intake valve opening time on a high output, how does this effect the in-chamber air/fuel homogenity? It would seem to me that when the injector is spraying and the valve is closed, you've got an extremely rich mass of air built up behind the valve and a terribly lean mass of air upstream of the injector. Does the turbulence in the chamber 'take care' of this? That just seems like a very crude way of doing things.

Greg also pointed out that below peak torque, going from batch to sequential made absolutely no difference. This is where I would have assumed the largest gains would take place. It would seem to me that firing that injector so halfway through its pulse would line up with the halfway point of the valves lift. It would seem to me that the most air would flow past the injector during that time period to give the best air/fuel mix before it enters the chamber.


-=Whittey=-

 

[GregLocock]

I don't know why, I just remember overlaying the torque curves and being unimpressed by the 'advantage' of sequential. Incidentally for reasons that I know NOTHING about, we fire the injector into the closed valve. Maybe this is why we see such small differences?
Cheers

Greg Locock

 

[Chumley]

Whittey: I agree with Greg in that sequential firing vs.
continuous makes very little difference except at peak
torque and best economy power settings.

Bosch, the father of modern fuel injection, after spending
years on electronic sequential port injection went BACK to
CIS, continuous port injection, in about 1975 and used
various forms of partially electronic controled CIS up
until around 1992 when they switched to Motronic systems
which were once again sequentially fired with fully
electronic controls. From a flow rate standpoint, at a
given common pressure, higher flows are possible with CIS
than with sequential injectors. Think about the inertial
mass and turbulence of the fuel as it is injected. Bosch
opted for the CIS systems on most of their higher performance turbocharged applications.

From a fuel vaporization standpoint, a good arguement can
be made that the "time" the mixture remains in the intake
manifold runner helps improve the fuel mixture because it
has more time to vaporize, something sequential injectors
limit by design, however, sequential injectors offer better
fuel mixture control between the individual cylinders than
continuous injectors can hope for. I was "shocked" to see
that Bosch allowed up to a 10% fuel quantity differential
between cylinders on their CIS systems. Not quite good
enough for modern OBDII emission and knock control systems
while eeking out that last little improvement in fuel
economy to meet CAFE standards since the leanest cylinder
has to run just rich enough to keep things happy under the
hood!

Chumley

 

[patprimmer]

My practical experience with an EA Ford Falcon (1988 Australian Model) with nonsequential multipoint injection was that for the size of the motor, it had quite bad very low speed stability under load. Specifically while allowing a boat to settle into the cradle on a steep ramp. I had several options.
1) Go up the ramp at a speed that endangered the crew members guiding the boat, and droping the boat to far back on the trailer.
2) Fry a clutch again.
3) Go and get my old air cooled VW and do it with that.

The Falcon had a 2.9 diff and 3.5 low gear, which also did not help.

The Falcon ran real rough at 500 rpm under load.

A Holden Commodore I had next, while not nearly as good on the road as the Falcon, had much better charicteristics on the ramp. It had sequential injection, and never fried one clutch. Regards
pat

 

[Whittey]

Hrm. Interesting. Thank you for your replies.


-=Whittey=-

 

[kimbo1]

For non-sequential firing (one injection every revolution) most cars I have tuned idle at about 2 ms, although the for different vehicles the range varies from 1-2.5 ms depending on injector size etc, the lower the time the harder it is to control the a/f ratio stability.
Full load pulse width tends to vary from about 8 ms (high rpm cars with large injectors) to 12 ms (standard cars). Now they all vary, but the basic idea is that you try and size the injector for about 80-85% duty at full engine power(ie ratio of actual pulse width to available pulse width).
Remember also that injectors have a "dead time" of about 0.5 ms this is the time they take to open and no apprecialble fuel is delivered during this time.
sequential injection asissts enormously with idel control and a/f stability at light loads but doesn't help much at higher loads.

 

[patprimmer]

My problem was very low speed, with moderate to high loads.

ie 500 rpm, 1/4 ? throttle. It just kicked back and stalled.

A FI expert, fixed? it by tricking it to run quite rich. It then guzzeled fuel, and left some black deposites on the rear bumper near the exhaust, but did however pull the boat at low speed without slipping the clutch to much. I doubt it would have passed a HC emmissions test.

I am also shure that if I had changeed the diff fron 2.9 to about 3.9 it would also have fixed it. What happened to the days when the heavy duty tow pack had a lower diff? or it was at least an option. Regards
pat

 

[kimbo1]

doesn't sound like an expert fix if that was the solution - more like a band aid!

 

[patprimmer]

I couldn't agree more.

I expect that fuel being randomly injected, occasionaly puddled behind a closed intake valve, then failed to evaporate sufficinetly to burn reliably, causing eratic or rough slow run. Note, I needed to labor the engine to less than it's normal kerb idle, partly because of the rediculous final drive ratio. If it had a 3.9 instead of a 2.9 final drive, it would have worked much better in my application, and somewhat better in most towing situations.

One of the main attractions of the "big Aussi Sixes" is their aleged towing capacity.

Regards
pat