Torque Dip
 

There was no response to this post in the Perrin thread and I think it's too important to let slip by unnoticed?
I wondered at what revs the port and direct injection come into play and this chart shows exactly that. Interestingly the direct inject is a constant, at least from 2,300rpm when this chart starts but the port injection shuts down from 3,000 - 5,000rpm.

Perhaps this was done for emissions reasons in that rpm range?

The question is, does the direct injection struggle to supply enough fuel on it's own and is that the cause of the torque dip?

If that's the case then why can't the port injection be turned on for the entire rpm range to fill the dip?

I realise that this is far too simple a solution and would have been implemented if was that easy but just interested now that i've seen this chart?

I think the general consensus is that the torque dip is not due to injectors. I think the going idea is that it is caused by poor cylinder filing in that rpm range due to intake manifold harmonics. However, as far as I know there is no conclusive proof of this. Many engines have variable length/volume intake manifold designs (TVIS) in an attempt to develop a flatter torque curve. To my knowledge this engine does not have such a feature but could benefit from it.

So why do you think they shut down the port injection at 3,000rpm and increase the direct injection to compensate?
Notice that the opening of the port injection again at 5,000rpm also results in a net gain which is exactly what I feel at a constant throttle when it hits 5,000rpm.

has nobody messed with the ECU to keep the port injectors on? it would be interesting to see what the torque curve would look like if they were just kept on and tuned to keep proper air:fuel.

but yeah, a major objective of the direct/port injection system is for emissions.

Port and direct injection is so complicated that Subaru tried to implement it before this joint venture with Toyota, but literally gave up. It is far more complicated than it seems. One thing that is interesting to note is that the injection volume reaches somewhat of a low point at 4K rpm. I think this may relate to the fact that there is less air in the cylinders due to poor filing.

Edit: The reason as to the dip has been disused to great length on this and every other FR-S forum, FYI.

here's an in depth look at the evolution of the 4-DS system if you want to go read the intricacies of the setup and how it works. i set it aside for future reading. it's a lot...

http://www.ft86club.com/forums/showthread.php?t=3172

I was thinking the same thing as the OP when I first saw this, but after reading the comments above, I think it's more complicated.

Im a noob on this topic, so I could be way off, but if you look at the AFR ratio, it's fairly steady and doesn't show a dip, so regardless of where you inject the fuel from, it's not really going to make too much difference to that dip. Or will it?

Who's to say the throttle isn't limited in the 3-5,000rpm range to match the limited fuel available? It's electronic so we wouldn't have a clue.

It seems a big coincidence that the torque dip is where the port injection is turned off and the direct injection total time open is no match for the combined port and direct.

the torque curve dips around 4000~rpms if i recall correctly, but the port injectors open up at 5000, which almost doubles the amount of fuel being dumped into the engine. my guess is that the engine is using the DI mainly for the purpose of emissions in the early rpms.

It pulls well until 3,000rpm when the port injection shuts down and it's like a switch being thrown at 5,000rpm when it comes back on again in an instant.
So if this period from 3-5,000rpm is for emissions using the more efficient direct injection, why can't ecutek or one of the others turn the port injection back on?

it seems like it would be a simple process of turning it back on and correcting the air:fuel, but then again, i'm no programmer... or engineer, lol.

I think some of you are missing the importance of the green (INJECTION VOLUME) curve. The fuel rate isn't dropping when the type of injection changes.

I have an engineering background but the ECU side of engines is new ground and I can only apply old fashioned logic to this.
In all the talk about the torque dip in the past I don't recollect any mention of the port injection just happening to be turned off for the exact range of that dip? Lots of 'is this the point where the port injection stops and the direct injection takes over?'

Truf! but most DI engines use forced induction as well. i think the method and spread of the fuel within the cylinder chamber has an effect on the torque that the engine can produce.

Blue injection volume curve?

The problem with that is if it's showing the total fuel being injected at full throttle? Then why is it only showing another 10% of fuel being used at 5,600rpm than is being used at 3,500rpm? Or almost the same fuel at 7,600rpm as at 4,900rpm?

they should run both much earlier because that mix of port and di give the car its best performance. giving the benefit of the doubt i assume the tuners have tried this already

They do both run much earlier but for some reason the port injection shuts down in the torque dip range.

You would assume they have but have you ever seen this mentioned before?

I was running on 100% port injectors for a little while. Doesn't help your argument because I wasn't anywhere near stock though.

Bingo

Did it do anything to the torque dip?

nope i havent seen anything about the details of their tunes. id really love to cut the bs and put a AEM in it and bring it to my tuner over here in newark.

Just googled AEM tuning and found a whole new world!

I was turbocharged and had an extremely custom tune. Can't really attest to anything.

No, turbocharging always fills the torque dip too.
Do you know if Ecutek can turn the port injection back on?

Interesting question - but wouldn't doing so likely hammer the fuel economy?

unlikely. the fuel economy would take a tiny hit at best. keep in mind that the air:fuel remains relatively constant throughout the powerband. the biggest issue is usually getting the air into the cylinder chambers since the camshaft on most cars is usually designed for low end torque. this is why Honda's VTEC is a thing; although, most people have no idea what it does because they have never sat in an S2000 or TypeR.

what would take the biggest hit is the emissions, but if you are messing with the ecu, emissions is probably not your concern.

As Ziggz has said, the mixture can be kept constant. If restricting the injection to direct, for emissions at that rpm, is restricting the the available power then keeping the port injection operating will restore that power. The fuel consumption might suffer slightly because direct injection is more efficient and the emissions would be slightly higher but let's face it, they look to have done this to pass specific rpm emission requirements which is bending the rules anyway.

The torque dip is overrated. I don't even feel it..

Blue-green on my screen.

I think it's the volume per cylinder, not volume per second. That would make it much less dependent on RPM.

You can be more aggressive with the tune when you're only running direct injection as it's fundamentally more detonation resistant. My read is that the combination of the intake and the cams have been designed to give a big hump in torque low down for daily driving and a hump up top for when you're driving hard. This leaves a hole in the middle which they have been really aggressive in the tune in an attempt to minimise as much as possible.

you didnt know about them? i been running one in my civic for years. best ecu ever. closes think to a motec and a sub normal price

@Visconti perhaps he could comment on adjusting the direct/port injection overlap and its affects on the torque dip in his tunes?

At the time we were using ProEFI.

The Mass Airflow is green but it's the only one without a trace on the chart.
I've been picked up many times by women for the blue/green border, they have a better sensitivity to colour so who knows?

Good point with the volume. If that's the case then the chart should indicate volumetric efficiency. The only problem with that is that the lines climbs from just below 3,000rpm to 3,600rpm when it's losing power?

It still makes no sense that they would have left the gap. The only thing that makes sense is emissions at say 50% of the max rpm which is 3,700rpm and they turned off the port injection to keep it super clean at the expense of some torque.

That would be very interesting. I'm wondering if Ecutek has the ability to turn it back on?

No, never heard of them before. The warranty is only 3 years here so i'll save that for later. What do they sell for?

I think you're underestimating how hard it is to get a flat torque curve! The engine has variable valve timing but lift is fixed and intake length is fixed, you're going to end up with a hole somewhere as you just don't have the adjustability in the intake to maximise VE everywhere. It's a sportscar so it needs to work up top but it is intended as a road car so it needs bottom end to cope with DD. That leaves the lower/midrange which you don't really need for either duties as the one which is the least important. It's not a matter of choosing to leave a hole, it's a matter of not being able to fit the more expensive intake/cam technologies in the price target and then selecting the other items (intake length, cam lift and duration etc) so that they are their most efficient where it matters and least efficient where it doesn't.

You also have to remember that direct injection only gives better torque, not worse. As it's able to deliver the fuel charge very close to the plug the flamefront propogates through the fuel charge quicker minimising the risk of detonation, this lets you run more timing to maximise torque. A number of tuners have remarked that the tune is very aggressive through this area in the maps which re-inforce that the issue is intake harmonics, not tune. The fact that no-one has been able to tune this out even though they have contol over when the port injectors are used also tells you that the problem is harmonics, not tune.

Everything in an engine is about compromise and to me it screams that they've compromised on cost by not using variable lift or variable length intakes (also a reliability issue for that one) which will result in losing VE somewhere which they've placed at the least important spot in the rev range. To minimise this they've gone aggressive in the tune however you can only make up for a lack of air so much in the tune so we're still left with a hole.

Very interesting thread...

Hoping one of the guys who has been able to mess with the tuning on this car can shed some light...

Are you are suggesting the port injection was turned off in that rpm range so that purely direct injection could maximise the torque to compensate for a lack of air? That it's nothing to do with an emission target?