Engine technology thread.

[blur]

Originally Posted by Silverpike

From the mechanical specs I've seen so far, it would take >$10k in parts and a complete rebuild to make this motor do that with any reliability.

I'm thinking it'll be as simple as a re-tune.

[Ice]

Originally Posted by Soravia

What design compromises did Honda engine use to make power on higher RPM like that? I see that it is a very tall motor.

I dont think they were compromises more just a different design of engine. Lets think about it this way, generally an engine is developed for a specific range of cars or model. They are designed with a certain demographic in mind i.e. Corolla, 323, Evolution, WRX/STI.

Each has its flaws and its strengths, would you say using a lower compression piston in a WRX compared to a 2ZR-FE a compromise?

The F20C was developed about the time that Honda was on a roll with its engines (at least IMO). The F20 remains one of the best engineered engines to this day (~123 HP a litre) being only eclipsed by the Farrari 458 Italia (by about a HP).

Considering this engine was developed near on 10 years ago I think speaks for itself in how advanced it was. The engines whole design revolves around high rpm and is tuned very well. (In fact many mofications as far as im aware inflict adverse performance on the engine in terms of figures. They may improve aspects of the engine however (i.e. ITBs))

As a result of high rpm the engine has really light pistons (made from Forged Aluminium IIRC), 12-14 pound fly wheel, coated piston skirts, the block was sleeved in fibre of some sort, it also had quite short piston skirts and a wrist pin mounted fairly high up the piston AFAIK.

It also had a tuned intake plenum and a really long rod (153mm much much longer then K20). (I apologise if ive missed something or its incorrect its coming from memory.)

[Soravia]

Wow, I see how they built it just for the high RPM operation. I'm sure plasma coating of cylinder bore is common on Nissan, Ford, etc now. I think even the FB20 has it (not sure).

So, crank (for additional power, possible vibes), bearings, springs, light valves, light + strong + long rod, light + coated + small pistons, light wrist pin, upgraded lubrication, stronger bearings, high volume breathing tuning, etc are the key to higher RPM operations. (nothing crazy like Formula 1 V8 making 800 HP)

100 HP per liter FA20 makes is pretty good, but somehow I think it can do better since old 7th gen Celica and Acura RS-X also made as much with less CR and no direct injections.

[Dimman]

Originally Posted by serialk11r

Dimman I'm not sure that you can even make strong assumptions based on intake speed with a direct injection system in place; Supposedly they're a game changer. Is it not curious that Toyota uses no lift control on engines with D-4S?

The 2GRFSE correlates well with the intake speeds. Toyota has only used a lift system on a single North American motor that I can think of, the 2ZZGE, and that is their only motor that shows the tight torque/hp gap.

Oddly, I'm thinking that Valvematic could produce the same effect, but without the 'double' torque curve of the 2ZZGE. Instead of the two humps, Valvematic would likely show a line connecting the low/short cam torque peak (which shows up where I predict) and the high/long cam torque peak (close to the hp peak in the 2ZZGE's case), which actually isn't much more than the other one, but maintains to higher rpm.

[bambbrose]

Back to our strange cam cap item discussion, it's definitely not on both sides in this picture.



and in this picture, doesn't it look like the booster vacuum line is running to it?

[Soravia]

Does that part show up on FB engine?

[Ryephile]

Oooh I might've nailed it with the vacuum pump idea. That would be great for track-work indeed, as some vehicles tend to lose brake booster vacuum with extended engine WOT.

[Marrk]

Originally Posted by bambbrose

Man, that is one tight, crowded, complicated engine bay.

[Dimman]

Well I'm conceding defeat on my missing torque conspiracy...

However this possible answer is contingent on AVLS or other staged lift system being used.

[Silverpike]

Wow Dimman that's a very thorough diagram.

I really can't wait until the major tuners get a change to crawl through the car and fill us in on the details. Like someone pointed out earlier, most bolt-ons are going to be useless unless the ECU is tunable out of the box. Based on Tada-san's interview comments, I would expect it to be pretty tunable. My fingers are crossed for OpenPort compatibility!

[serialk11r]

AVLS still has intermediate rockers though, which the diagrams and dimensions suggest don't exist...My current explanation is going to be that they detuned the engine like crazy to get it to super duper ultra mega low partial zero emissions VI (or whatever)

About Valvematic, I don't think it would just "bridge" the 2 "humps". The thing with Valvematic is it would come with a super aggressive profile cam optimized for peak rpm operation, and then it cuts down on the lift and subsequently duration to get it to work at lower rpm. Inevitably what happens is that you cannot really use the extra duration at most rpm ranges, but you have to cut a lot of lift too, so the lift is not really optimal most of the time at full throttle (too high). They could cut it down more, but then duration gets cut, and you have less air to burn. In that situation they'd probably have it programmed to have duration cut more since it would improve thermal efficiency and combustion efficiency a little bit, but it's hard to say how the torque would behave. Since at 3000rpm ish losses tend to be the lowest, a 3d continuously variable duration cam plus continuously variable lift (aka the perfect cam system) would be able to get a torque curve that falls very slowly and gently from 2-3k rpm all the way to redline, but when you control lift and duration together you cannot optimize low rpm power that way. Nissan VQ37VHR seems to have freakishly consistent torque from 3000 all the way to 7000, which supports this theory I think. The max lift profile is designed to work best at 7000 rpm or maybe a little less and then they "throttle" it away for lower speeds.

[Dimman]

Originally Posted by serialk11r

AVLS still has intermediate rockers though, which the diagrams and dimensions suggest don't exist...My current explanation is going to be that they detuned the engine like crazy to get it to super duper ultra mega low partial zero emissions VI (or whatever)

About Valvematic, I don't think it would just "bridge" the 2 "humps". The thing with Valvematic is it would come with a super aggressive profile cam optimized for peak rpm operation, and then it cuts down on the lift and subsequently duration to get it to work at lower rpm. Inevitably what happens is that you cannot really use the extra duration at most rpm ranges, but you have to cut a lot of lift too, so the lift is not really optimal most of the time at full throttle (too high). They could cut it down more, but then duration gets cut, and you have less air to burn. In that situation they'd probably have it programmed to have duration cut more since it would improve thermal efficiency and combustion efficiency a little bit, but it's hard to say how the torque would behave. Since at 3000rpm ish losses tend to be the lowest, a 3d continuously variable duration cam plus continuously variable lift (aka the perfect cam system) would be able to get a torque curve that falls very slowly and gently from 2-3k rpm all the way to redline, but when you control lift and duration together you cannot optimize low rpm power that way. Nissan VQ37VHR seems to have freakishly consistent torque from 3000 all the way to 7000, which supports this theory I think. The max lift profile is designed to work best at 7000 rpm or maybe a little less and then they "throttle" it away for lower speeds.

No AVLS? Argh... I was working under that assumption for my previous prediction, but I only remembered that it looked like the FB20 didn't use it, and that the cam carrier was modular, so the FA could have used it. It was also the best I could come up with for the high torque peak.

As for the duration changes on Valvematic, that works with the peak to peak line idea as well. The VVTL-i is not simply a lift change, it switches to an entirely different profile. The lift AND duration are different between the two stages.

[bambbrose]

Dimman, why is it exactly that you think this motor should have 160+ lb ft of torque?

Is it solely based on the beams output?

I was looking at K20A motors, which have a more aggressive cam profile on the high lift, and VTC as well, and they are only 150 lb ft, and usually only 130 wtq on a dynopak.

I believe 150 lb ft is in line with the best of them. Even the F20C has 153 lb ft, and it has a superior factory ported head and amazingly aggressive cam profiles.

In looking at motors across the spectrum, the beams is the only one with high torque numbers. I think they are falsely reported, or inflated numbers from the factory.

EDIT: Found the discussion on pages 20-25.

Compression should add torque, but I think that the 151 is on-par.

I'm thinking this motor will respond well to a nice set of cams however.

[Dimman]

Originally Posted by bambbrose

Dimman, why is it exactly that you think this motor should have 160+ lb ft of torque?

Is it solely based on the beams output?

I was looking at K20A motors, which have a more aggressive cam profile on the high lift, and VTC as well, and they are only 150 lb ft, and usually only 130 wtq on a dynopak.

I believe 150 lb ft is in line with the best of them. Even the F20C has 153 lb ft, and it has a superior factory ported head and amazingly aggressive cam profiles.

In looking at motors across the spectrum, the beams is the only one with high torque numbers. I think they are falsely reported, or inflated numbers from the factory.

EDIT: Found the discussion on pages 20-25.

Compression should add torque, but I think that the 151 is on-par.

I'm thinking this motor will respond well to a nice set of cams however.

The issues I have are that the specific output of torque should be higher (from the BEAMS, but ALSO from the 2GRFSE in the IS350) and it should be making it at lower rpm. The gains go back to the BMEP thing.

BMEP is averaged cylinder pressure after losses (heat and friction). Even if the BEAMS was overstated it would probably still be in the 150 range (11.1:1 compression ratio on it). Looking at the gains from the ES350's 2GRFE (no D4-S) and the IS350's 2GRFSE (with D4-S, both on premium gas), that was 9% (which is same valvetrain, D4-S combustion gains and the compression increase).

This motor has the same Bore and Stroke as the BEAMS. With all the data they've accumulated on the BEAMS with regards to combustion chamber shape, compression ratio, rod:stroke ratio, port shape, valve sizing and valve events, valvetrain and piston friction, etc... it would be the logical starting point. Flattening the motor has little bearing on that.

The gains of going to low friction finger followers, the H4's better balance (no extra harmonic damper mass), probably fancy anti-friction coatings on the pistons, 12.5:1 CR, and D4-S alone should have significant improvements. If the 159 lb-ft of the BEAMS is correct, it would need less than a 6% increase. For the same increase of the 2GR (~9%) the starting point would have to be ~154 lb-ft.

Also for the 2GR, the FSE on premium makes a 12% gain over the FE on regular gas. If this 12% gain is applied to the FB20 in the Impreza, it hits more than 151 lb-ft...