anesthes

iTrader: (12)

I was reading another thread, and it got me thinking about this again.

Those of you with the ability to control your injector firing on port injected V8s, what is your preferred strategy?

How many times do you fire per cycle? (two revolutions)

Do you fire all 8 at once, or alternate?


Stock TPI I believe fires once per revolution, and all 8.

-- Joe

RBob

iTrader: (1)

If I had a choice it would be SFI. Although batch fire does work rather well.

The '86 - '92 TPI ECMs use two strategy's. Not sure about the '85 ECM, so leaving it out.

There is the normal double-fire mode. This is where the injector PW is above say, 1.2 mSec. The ECM fires all 8 injectors once every engine revolution.

When the injector PW gets small, the ECM reverts to single-fire mode. It doubles the injector PW, applies the compensations and fires the injectors once every two engine revolutions.

RBob.

anesthes

iTrader: (12)

Hello,

So why wouldn't you always fire it once every TWO revolutions? You would get back the dead time for opening and closing the injector one of the revolutions, so you could run a higher DC.. ??

Is it a matter of fuel distribution ?

-- Joe

efiguy

iTrader: (2)

Firing it once every two revolutions sucks. That's the way the Holley Commander 950 did it with the "PRO" software.
Why? Simple, think about it. You're now firing at least a few cylinders at exactly the wrong time, at or near both valves being open at the same time. Half the fuel will go right out the exhaust, with no second pulse to even slightly correct that.
And when you fire all eight at once, either once or twice per power cycle, it's harder to keep even fuel rail pressure.

Univ of North Carolina Charlotte did a test years ago on a mild 300hp 350 Chevy. They used a Holley C950 first and dialed it in as good as possible. The instructor told me that he always had 2 cylinders that ran richer than the others and 2 that ran leaner. The other 4 stayed pretty close. They then did dyno pulls.
Then they did nothing other than change to another aftermarket system that ran in sequential. Peak hp was better with the sequential system but only by about 5 or so. However the midrange was much better, almost 30hp/tq better. And it used less fuel to do it.

Remember this as well, even the F1 stuff that winds to 18K runs in sequential, even at that rpm.

anesthes

iTrader: (12)

I don't have enough injector drivers to run sequential without busting out the soldering iron, which isn't happening.

What's your preferred firing strategy on batch fire?

I have two drivers, so I can do bank to bank, both banks at once, I can fire anywhere from 1-10 times per CYCLE (two revolutions).

I'd have to solder 6 more transistors in to do sequential (and run new wires since the OE harness is bank/bank.

-- Joe

Tuned Performance

iTrader: (92)

I will sell you my dfi 7 joe

anesthes

iTrader: (12)

Hah, No thanks. I don't want to lose functionality

-- Joe

Saar

no V8, but i've played with this quite a bit on my wife's car.

1990 3.1MPFI: as RBob described, normally double-fire, PW drops too low, single-fire kicks in, when above another PW threshold(there is a hysteresis to keep it from switching too often), back to double-fire. i don't have enough injector to get the PW down to the range the stock calibration requires to switch in and out. so, i did the next best thing, raised the thresholds significantly, enough to where anything less than about 20% throttle was single-fire. i could never get the transition to work consistently though, roughly half of the time i would get a single-fire (double-width)pulse of fuel on one revolution, then on the revolution, a double-fire pulse would happen, resulting in a super rich AFR for a small amount of time. with the TCC locked, huge stutter/bog. i seem to remember the double to single transition being correct, but it has been a few years.

i could never find the perfect place in the code to insert some the fix required for the single-to-double transition before we got rid of the car, so i backed the thresholds down to where they originally were. i tried running single-fire at all times, i don't remember having any issues with that either, but i could be wrong.

i would assume that if you have any incorrect values in the voltage compensation and minimum pulsewidth tables, they would be more obvious while in double-fire since they are accounting for a larger percentage of time that the injector driver is commanded on.

anesthes

iTrader: (12)

I guess my question is, why would you ever want to run anything other than single fire? Pulsewidths down low on double fire could be too short, then you have the extra open/close time issue with upper RPM duty cycle.

But then I was thinking, did the engineers choose double fire to satisfy some uneven fuel puddling problem where it's batch fire ?

-- Joe

Saar

from what i understand/have been told, having the fuel charge sit on the back of the intake valve(or the port walls close to it) isn't a bad thing for fuel atomization.... hot fuel evaporates/mixes easier with the available oxygen, so there is the potential for a better burn. in double-fire mode, at least one of the injection events will be against a closed valve on every cylinder. if i'm thinking about it correctly, each cylinder's fuel charge will also have roughly equal amounts of time in this condition as well, or at least close to the same amount of time.

in single-fire, that isn't the case. one or two valves(i need to think about it more/chart it to be certain which) will be open every time the injectors fire, so it won't get that benefit. the other 6 or 7 cylinders will all have different amounts of time that the fuel sits in that state before they open for the intake cycle.

switching to single-fire at low pulsewidth feels like it was done due to the mechanical limits of the injectors, rather than it being some kind of advantage.



of course, i could be wrong.

Saar

a became a strange combination of curious and bored and generated these.




double-fire followed by single-fire. i tried to keep everything as centered as possible in both, but some cropping issues came up so it isn't perfect.

anyways, for the fuel pulse, i chose an on-angle of roughly 90* per revolution(with the double-fire obviously split into two pulses). 90/720 = 12.5% injector DC. injection events in this scenario happen as soon as a cylinder goes TDC, real-world..... it may match, i haven't touched an ECM controlled dizzy in quite some time, so i'm a bit rusty there.

for double-fire, injection events are happening at the pulses for cylinders 1 and 6, where for single-fire, only happening on 1. the actual cylinder where the EST module and ECM sync up at will vary, probably a different cylinder every time you start the engine.

in any case, selected(for less typing) notes my "analysis" of what is happening:

double-fire:
1 - fuel injected at beginning of power stroke, charge has all of power and exhaust strokes to sit on the valve. second injection happens right as intake valve is opening up.
4 - fuel injected at beginning of power stroke, right after intake valve closes. first charge has 3 entire strokes on valve. second charge delivered at beginning of exhaust stroke, has 1 entire cycle.
3 - first charge injected(and likely ingested) during last half of intake stroke, no real time sitting on the valve, just moving past it. second charge delivered halfway through power stroke, 1.5 cycles of sit time.
6 - first charge injected at beginning of intake stroke, probably a small amount of time sitting on the valve before airmass brings the fuel with it into the chamber. second charge at beginning of power stroke, 2 cycles of sitting time.
7 - first charge injected at beginning of exhaust stroke, 1 full cycle of sitting. second charge injected right as intake valve closes, 3 full cycles of sitting.

so, the way it plays out in terms of how long the fuel sits on the valve(in cycles), it comes out to:
first charge + second charge
2+0 (2)
2.5+.5 (3)
3+1 (4)
0+1.5 (1.5)
0+2 (2)
.5+2.5 (3)
1+3 (4)
1.5+0 (1.5)

so, a range of 1.5 to 4 cycles.

no commentary on the single-fire chart, seems kind of self-explanitory after the double-fire chart. just the time the fuel charge sits on the valve.

2
2.5
3
0
0
.5
1
1.5

and here, we have the range of 0-3 cycles.




i guess, draw what information you will from this. all of this is done under the assumption that time on valve = good, if that assumption is wrong, all of this is pretty flawed.

in any case, the larger the pulsewidths get, the less difference it makes as to which scheme is used.

RBob

iTrader: (1)

In the '86 - '92 TPI ECMs single fire mode (S/F) is done via the firmware. As such it is limited to a maximum of a 50% duty cycle. So it can't be used all of the time.

The other issue with always using S/F is tip in response. With the injectors firing once per two revolutions, it will take some time for the larger volume of required fuel to actually be injected.

This can cause driveability issues. Can cover some of this issue with AE. But that can cause other problems.

RBob.

anesthes

iTrader: (12)

That is how the MS deals with it:

http://msextra.com/doc/ms2extra/MS2-...anual.html#EAE

Complex and crazy system of adding injection events during AE called "lag compensation".

So I could in theory run single fire and use lag compensation to try and cure tip in problems, but I'm not convinced I'll always be in AE on throttle transitions.

The other problem is, there doesn't seem to be a safety catch to prevent it from adding too many injection events (and as such, creating such a small BPW that it becomes too small).


How does the other aftermarket stuff deal with this?

-- Joe

-- Joe

junkcltr

iTrader: (1)

Many GM ECMs only had one injector driver causing the problems associated with single/double fire.

GM ECMs with two drivers like the 427 ECM can be reprogrammed to do alternating bank fire, and have hardware to do a quick AE. This ECM can also do P&H injectors. With alternating bank larger injectors can be used for high HP (boost) and still have idle pulsewidths greater than the min speced for the injector.

anesthes

iTrader: (12)

What would the benefit of bank firing be on an application like $8D ? (if it was available).

I can turn on bank to bank on my MS, but since the MS has absolutely no idea which cylinder I'm on I don't see how that would be a benefit.

Right now I'm running 2 squirts per cycle, simultaneous. (Double fire, batch).

-- Joe

kurtis72

I played with this a fair amount on my 5.7TPI running MS3X, 19lb injectors at ~60lbs pressure. I found 4 squirts/cycle alternating to be the best.
It seemed to have a small amount more HP at lower RPM up to 3000 or so. After that seemed to make no difference.
Idle seemed unaffected. Duty cycle seemed unaffected.
I always wanted to test what affect this setting had on MPG, but never got to it.

anesthes

iTrader: (12)

With really small injectors, that makes sense to a degree I guess. The 4 squirts / cycle is surely better for fuel distribution, and being only 19lbs you can get away with a larger pulse width without over fueling.

What is the dead time on the injectors and what was the pulse width at idle?

If you're capped at a max PW of 5ms at 6,000 RPM though if you are squirting 4 times per cycle. (each cycle takes 20ms @ 6k RPM). Realistically, the dead time is probably between .700-1msec, so you can theoretically only be open for around 4ms before being static. I bet if you commanded the injector open for longer than 5ms you will suddenly be way richer from losing the dead time.

I'm sure you have read this:

http://www.useasydocs.com/details/numsquirts.htm


I'm still torn about the whole thing myself. I'm leaning towards 2 squirts being best, because I think you can have a stable quality idle, and decent throttle transitions. The 'lag compensation' can add additional injection events as needed, although tuning that is complex.

Edit: I just re-read your post, I missed the "alternating" the first time around. 4 squirts alternating is almost the same as 2 squirts simultaneous. Instead of firing two at 0* and at 360* in a 10msec window each, you are firing 0* on bank 1, 360* on bank 2,
180* on bank 1, and 540* on bank 2.


-- Joe