Fully Sequential Injection
Turbo Mini Forums

Any value will do as the cam signal is not timed into anywhere specific, or is it?

I would go for 90/270 degrees.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Very interesting......

90/270 but still all your other variables (I assume different VE tables) and it took 6k RPM and a bit more boost but just at the point of merging...


The alternate short/long has gone.....

And obviously I now have the external trigger working so you see both ports.

The cam signal is at a fixed RPM from the JimStim Jean said earlier and the earlier plots seem to confirm this.

Schrödinger's cat - so which one am I ???

Good to see you lost the short/long pulses, but the two drivers are not in synch. They were on the first plot.

Is that the scope or is there a timing issue?

Also, the time scale suggests lower than 6k, or is something else going on?

Also, the outer cylinder seems to be getting the short pulse at times.

Edited by Paul S on 15th May, 2009.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Well spotted.....

It may be the scope as I'm freezing a moving display (whilst trying to operate syringe and PC at the same time)

They are in step when seperate (6K, no boost)


As for the timescale, not sure - it only seems to have happened since I changed to the external trigger, maybe a glitch in the firmware of the scope or the way I set the external trigger up - I shall investigate. Megatune still says 6K.

Schrödinger's cat - so which one am I ???

They now seem to be operating under an entirely diiferent mode. Both ports operating the same.

You're starting to alarm me Rod.

EDIT: Maybe my fault. When I said 90/270, I should have said single value at 90 and then the other would have defaulted to 270. If you have set it to dual value 90/270, then they will coincide...... I think.

Edited by Paul S on 15th May, 2009.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Now I'm getting really confused..... I swapped to single value (no difference) and I even went back to channel A trigger (in case of a USB scope glitch) but the timing seems to change with RPM....

Let me try to do a few more tests, more methodically.

Schrödinger's cat - so which one am I ???

The last scope picture looks fine for a 90/270 setup. This is equivalent to semi-sequential except for the different pulse width visible on the graph (and alternating top and bottom as expected).

What is not as expected is the previous graph. However, the code is not made to handle overlapping pulses in an "intelligent" manner and timing will definitely be affected as well as the overall pulse width. So you can't rely on either being correct and actually can rely on both being wrong. This is not how the code should be used, period.

If I look at the graph above, there is already a problem with the duty cycle being much more than 50% which should not happen. This will definitely inject fuel outside the injection window and result in over rich inner cylinders. If that is indeed what you need to run at low RPM then the injectors are too small.

Jean

http://www.jbperf.com/

Rod,

The problem is the duty cycle is too high. As I mentioned in my previous post, you can see it with the 90/270 timing graph. If you use the same timing then you have overlap and everything gets screwed up (timing and pulse width).

If you want to see the correct behaviour, lower the req fuel value. Also have a look at the duty cycle.

Jean

http://www.jbperf.com/

Jean,

I think what you are saying is that a high duty cycle must be avoided at all costs as the code doesn't handle overlapping pulses the way we expected.... (no offence to Paul or yourself) - these are all run with Paul's msq, changed where he suggested so we could see what happens with overlapping pulses.

I'll try my own msq later - not yet run in anger on an engine but set with much larger injectors as I am trying to keep below 25% duty.

If all this follows through, then staged injection becomes essential.

Rod.

Schrödinger's cat - so which one am I ???

I have always said that overlapping pulses are to be avoided and that is why I have been saying that the safe way to start is to use 180 degrees between timing (such as 90/270) because that places the pulses as far apart as possible (similar to semi-sequential but with separate pulse width settings).

If you can't get the outer cylinders at the correct AFR with such a timing split, no matter what you do to timing and pulse width will ever correct the situation. You either need bigger injectors or staged injection, the latter being my recommendation for better overall behaviour.

Jean

http://www.jbperf.com/

We have a big problem with the sequential siamese mode then.

My tests have shown that a maximum of 15-20% duty cycle will give the correct size of injector.

That means on my 1030cc, I need 4 injectors at 1000cc/min.

No way will that idle even with staged injectors.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Paul,

I'm not sure what you mean. Do you mean 15-20% for one pulse (108-144 degrees)? If so, then the actual duty cycle the injector will see is double that so about 40% which should be ok.

If you mean overall duty cycle then I don't understand because the injection window is certainly larger than 7.5-10% or 54-72 degrees.

Jean

http://www.jbperf.com/

First point,

I've now watched the scope in real time (so I can't do prints) and as I wind the RPM (or boost) up it clearly all goes wrong just as the pulses merge - timing, pulse width, all goes wrong.

No problem with that, no offence to Jean, if it was only ever intended to have less than 25% duty per injector, or even less once you start moving the timing.... but obviously needs a think about injector sizing with that limitation.

Second, Ive just run my msq and with TWO 680cc injectors (what I'll start with) and 1360cc and the pulses don't merge until 0.6 bar of boost at 6K RPM.

Now obviously that is only with MS's initial self generated maps so could be a fair way out but it sounds promising for staged injection.

EDIT - that quick test was obviously with equal pulses and symettrical timing so it will get worse as I move them to get the equal AFRs.

Edited by Rod S on 15th May, 2009.

Schrödinger's cat - so which one am I ???

Jean,

The problem is that the injection window for each cylinder is no bigger than the intake stroke, so maximum of 25% for each pulse.

Then you have to take away the minimum rest period between pulses which leaves you with 15-20% depending on how much your injector needs.

I know that, in theory, you can shift the first pulse to inject earlier, but I can't get that to work any better.

If you try to work with two VE tables and two timing maps it's a nightmare to setup.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Just to be clear, when I now talk about duty cycle (since I was educated by you two a while ago) I take it to mean percentage of the engine cycle (two revolutions) so I'm taking 25% to be what each port's injector(s) does twice and hence where the pulses merge on a single port's injector(s) with symettrical timing and hence you may need less than 25% if the timings converge.

But overall, each port's injector(s) will run up to 50%..... that's what I based my sums on for doing my injector sizing and msq - I hope I haven't still got it wrong....

Schrödinger's cat - so which one am I ???

Paul, last question before I resort to bathroom tiles to clear my head..... do you think you reached the point of pulses merging ???

The msq of yours I've been trialing doesn't merge until 7K RPM or boost, but you had the rev limiter set at 6K (albeit on spark retard, not fuel) so have you changed much since you sent me that one ???

Schrödinger's cat - so which one am I ???

Rod, I don't think the pulses have merged electronically.

But I do think I have reached the point where the injector does not flow much fuel in the second pulse due to other factors.

What we see electronically on the scope is one thing. The dynamics of fuel surge within the injector itself is another problem entirely and something that no else seems to be able to get to grips with.

Being a hydrodynamicist by profession leads me to believe that you cannot get the pulses any closer than say 3-4 mS before you start to affect the flow rate of the injector.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

Paul,

I can't argue with you about the tuning aspect because I haven't been through it. And I'm sure that having to tune 4 tables at the same time is a nightmare.

But could it be done in a more progressive manner by first setting things to a known slightly rich safe state and playing with the timing first with the only concern on the outer cylinders timing and always adjusting the inner cylinders timing 180 degrees more.

Then once you know the timing is good, you start leaning things up with the 2 VE tables. I'm sure you'd need to go back to the timing after that but after a few iterations at this 180 degrees split, you should have something reasonable that would maximize the duty cycle. And then you could see how much time there is between pulses and adjust the inner cylinders timing to keep a good 2ms (or more) between the pulses.

I'm aware that writing this doesn't make it as easy as I make it sound but the principle has some logic behind it. Unless I'm wrong you haven't tried to keep the pulse widely apart in your tuning which means you have hit artificial limits due to the code not behaving as you expected. I'm not trying to dictate how you should do things but it may be worth another try.

In any case, I will be looking at implementing the hybrid mode with the possibility of going from 2 pulses to a single one using 3 tables. This way, anyone can chose to use a single pulse, 2 pulses, or a mix of the 2 however they see fit. The only thing this will not allow is the use of the hybrid mode with a single pulse at low load/RPM and 2 pulses at high load /RPM but this should not be a limitation because that doesn't make any sense.

Jean

http://www.jbperf.com/

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

http://www.jbperf.com/

Schrödinger's cat - so which one am I ???

The hydraulic problems are due to pressure surges.

If you rapidly close any valve on a hydraulic system it will cause a reverse pulse, the higher the velocity and pressure, the worse the pulse can be.

That reverse pulse will prevent the flow in the injector at the next pulse from reaching it's design flow if it is too close.

Edited by Paul S on 15th May, 2009.

Saul Bellow - "A great deal of intelligence can be invested in ignorance when the need for illusion is deep."
Stephen Hawking - "The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge."

http://www.jbperf.com/

Schrödinger's cat - so which one am I ???

create a greater reservoir of fuel ,before the inj , so crreating a damping field ,bit like a large plenum .. try changing the diameter of the fuel rail to a much larger one ....if theres room of course.

Medusa + injection = too much torque for the dyno ..https://youtu.be/qg5o0_tJxYM