Engine technology thread.

[serialk11r]

Could it be the D4-S calibration? I forgot who posted the chart that shows the proportion of DI and port injection at all RPM and loads, but port injection stops at around 3000. My guess is the cam is optimized for higher rpm operation. The port injection creates better combustion efficiency in the low rpm range, and "fills in" the torque. By the time port injection is significantly reduced and the DI supplies most of the fuel, the engine speed is not optimal for combustion efficiency with just the DI. This was my original guess for why they didn't bother using anything beyond phasing.

EDIT: So if mixed fuel injection can "cure" poor combustion efficiency at low speed, then it could be good news for people who want to stick bigger cams on right? Just a thought.

[Dimman]

Originally Posted by serialk11r

Could it be the D4-S calibration? I forgot who posted the chart that shows the proportion of DI and port injection at all RPM and loads, but port injection stops at around 3000. My guess is the cam is optimized for higher rpm operation. The port injection creates better combustion efficiency in the low rpm range, and "fills in" the torque. By the time port injection is significantly reduced and the DI supplies most of the fuel, the engine speed is not optimal for combustion efficiency with just the DI. This was my original guess for why they didn't bother using anything beyond phasing.

EDIT: So if mixed fuel injection can "cure" poor combustion efficiency at low speed, then it could be good news for people who want to stick bigger cams on right? Just a thought.

Partly, but there is also the need for velocity/inertia to keep the charge from getting pushed back out as the piston starts back up for compression which is more significant with longer duration cams.

[serialk11r]

Hmmm so from my limited knowledge, I am guessing that typically long duration cams don't help fuel economy much at lower speeds since the combustion efficiency is severely compromised. But if they can fix that, then the low rpm torque wouldn't be too bad, and the efficiency would be way up... The stock cam seems to be optimized for high rpm operation, aka long duration. At high rpm the losses are slightly higher, so the charge being pushed back out at lower speeds would agree with the unimpressive max. torque and the fact that low speed torque doesn't quite match the peak torque even with D4-S... In the past few days I've been starting to realize this car could really get AMAZING mpgs.

[arghx7]

Notice that the 2GR-FSE does about 65 kw per liter in specific horsepower output but the FA20 does about 75 kw per liter if you do the math. So this engine has already improved on the previous D-4S system significantly in that area.

Long duration with high lift (they tend to go together) is bad for fuel economy during normal cruising conditions in part due to fuel atomization issues. Look at the lift & duration profile range on the Nissan VVEL (VQ37) system:



You can see that the VVEL system is capable of being very mild or very aggressive depending on engine conditions. This allows good specific output and good specific fuel consumption. I'm sure Toyota could have put their system on the FA20 but it just came down to costs. Maybe in a mid-cycle refresh we will see it on the BRZ/FR-S.

[serialk11r]

arghx7, the thing is Honda uses a high duration, high lift cam lobe in the Civic and now all their newer engines for part load operation...so they must be achieving good fuel atomization somehow right? The Honda R18 and stuff is why I think it could work.

[arghx7]

Originally Posted by serialk11r

arghx7, the thing is Honda uses a high duration, high lift cam lobe in the Civic and now all their newer engines for part load operation...so they must be achieving good fuel atomization somehow right? The Honda R18 and stuff is why I think it could work.

Could you elaborate further or present any source of information on this? My understanding is that Honda is still using port injection with cam phasers on the intake and exhaust side and then a fixed high lift lobe for most of the engines. This is as opposed to continuously variable valve lift like Nissan, Toyota, BMW, and Fiat have put into production. Then again, it's no secret that Honda has fallen behind the times.

Some variable lift systems on 4 valve engines have different levels of lift on only one intake valve in order to induce swirl charge motion. Honda has done that on their lean burn engines and Subaru does it with their AVLS system. Basically one valve is essentially deactivated by having very low lift (Honda 90s lean-burn VTEC engine),




or one valve is active but has significantly less lift than the other (Subaru AVLS).



the whole point of that is to induce charge motion, specifically swirl effects. Those are port injected engines.

Older 1st generation direct injected engines (old Toyota and VW for example) relied on swirl effects. Most newer GDI engines use tumble flow instead.

[Dimman]

arghx7:

Any reason engineers went back to swirl for DI, when tumble seems to have been the dominant charge motion technique for port-injected pent-roof 4 valve heads?

[serialk11r]

http://world.honda.com/HDTV/news/2005-4050705a/ (kinda laggy and stream)

(screwed up)

Start at about 2:22. You can see the comparison between the low output and high output cams, the high output cam has slightly higher lift and longer duration.

[arghx7]

Oh. For that Honda engine (current Civic motor?) That's just a late intake valve closing Miller/Atkinson cycle process under low loads. You keep the intake valve open all the way into the compression stroke so that it blows back some of the air as a form of load control. That's pretty normal stuff these days--that's what a Prius engine has been using for years. Mazda was one of the first ones to use it on the engine used in the Millenia.

In Honda's implementation they used it as a cheap way to reduce pumping losses, without having to go with continuously variable valve lift (Valvematic etc) or stratified charge direct injection.

[serialk11r]

Right, but my point is that the late intake valve closing is a long duration, high lift cam, which supposedly affects combustion efficiency right? High lift = bad for combustion at low speed. So what's the difference between this and a racing engine that has crap torque at low speed? What trick is Honda using to maintain good combustion efficiency?

[arghx7]

Originally Posted by serialk11r

Right, but my point is that the late intake valve closing is a long duration, high lift cam, which supposedly affects combustion efficiency right? High lift = bad for combustion at low speed. So what's the difference between this and a racing engine that has crap torque at low speed? What trick is Honda using to maintain good combustion efficiency?

I dug into this engine more (it is the R18A). Technically, yes the Honda R18A engine has higher lift and greater duration but only on one (out of two) of the intake valves. However the timing is such that intake charge blows back out of that intake valve with long duration.


On the bottom look at Cam profile B, the red delayed closure line in comparison to the blue one.

So it's not really like a race engine in that way. What you have here is a combination of a Miller/Atkinson cycle engine, like on a Prius, and an AVLS-equipped Subaru engine. Since the high valve and duration occur on only one of the two intake valves, it stands to reason that you actually get a brief swirl effect just like on AVLS. It's not like a DSM 4G63 with some race cams.

Besides i-VTEC inducing swirl, the whole idea behind the late intake valve closing concept is to make the expansion ratio (during the power stroke) greater than the compression ratio (during the compression stroke). You do that by lowering the compression ratio during a given cycle. The Prius engine does the same thing. And these are port injected engines.

The Nissan VK56VD engine used in the Infiniti M56 uses late intake valve closing with continously variable lift, continously variable valve timing on both intake exhaust, and gasoline direct injection. I can tell you right now that the Honda R18A is meant as a low-cost approach to late intake valve closing. It's on a mass produced economy car SOHC inline 4 engine from over 5 years ago, so that makes sense

[arghx7]

Found some more stuff on the R18A.




The gray line on the left part of the chart is the intake valve closing for the higher output VTEC mode. Then as you look along the x axis of the chart you see increasingly late intake valve closing and its effects on a bunch of different factors. In all 4 plots, a lower value on the Y axis should be considered more desirable. The top is Pumping Mean Effective Pressure--this represents work done inside the cylinder wasted on pumping air in and out. The second is brake specific fuel consumption (grams of fuel per kWh of power). The third is coefficient of variation of indicated mean effective pressure. This is a measure of combustion instability. The last is combustion phasing--the crank angle at which a certain percent of the in-cylinder mass is burnt.

Look at the gray line on the right and you will see why the intake valve closing was set the way it is. Remember that the intake valve OPENING is the same. Think about it for a while and you'll see why this system is clearly optimized for cost efficiency.

[Dimman]

^ Is there a difference in combustion quality between charge that is swirling and charge that is tumbling? You mentioned AVLS is used to generate swirl, and it's by one intake valve having lift 'precedence' (for lack of a better word) and acting more like an older 2 valve head, with the intake port kind of aimed at a tangent to the cylinder wall, correct?

So is there a benefit to going back to swirl, since tumble kind of took over with 4 valve heads, and there didn't seem like a good way to produce swirl in a 4v head until variable lift (profile changing or otherwise) became more common?

[arghx7]

Originally Posted by Dimman

^ Is there a difference in combustion quality between charge that is swirling and charge that is tumbling? You mentioned AVLS is used to generate swirl, and it's by one intake valve having lift 'precedence' (for lack of a better word) and acting more like an older 2 valve head, with the intake port kind of aimed at a tangent to the cylinder wall, correct?

So is there a benefit to going back to swirl, since tumble kind of took over with 4 valve heads, and there didn't seem like a good way to produce swirl in a 4v head until variable lift (profile changing or otherwise) became more common?

Those are million-dollar questions. That's how much money it costs to figure that stuff out for an actual production engine. You need lots of man hours, expensive software, and expensive equipment to understand such minute details, especially when you throw in the complexities of direct injection. That's why they have fluid dynamics software. It's beyond the point of where throwing something on a flow bench and making some prototypes will provide that kind of insight.

All I can say is that practically all the new mass produced GDI engines across manufacturers primarily use tumble flow for charge motion. It's no coincidence that they have at the same time moved away from swirl injectors and to multi-hole type injectors. As exceptions, Toyota and BMW are using fan-type and piezoelectic/cone type respectively. The swirl combustion concepts appear to be based on old port injected lean-burn engines and research engines from the 1970s. The tumble motion in conventional PFI engines is comparatively weak to what's used in GDI systems. I have a nice graphic here from the Audi FSI Lemans race engine development program that compares various types of charge motion in context of both PFI and GDI engines. I'm digging up stuff on the original Toyota D-4 system right now, which relied on swirl flow.