Engine technology thread.

[Dimman]

Originally Posted by madfast

would it make sense if they artificially detuned the engine? i mean we're comparing a GT3 and a Scion here... from a marketing standpoint, maybe the subaru does have a bit more power and costs a bit more?

The BEAMS vs FA20 vs GT3 just above, aren't the numbers I am basing my assumptions on. They are the high-output BEAMS and the fantasy GT5 version of the FA20 and they are not as highly tuned as the GT3. Basically to show that even the GT5 output numbers are quite feasible (ie an STI version or something...)

The automatic-tuned BEAMS and what I think the FA20 will start with are as follows:

BEAMS 3S for automatic (11.1:1 CR, port injection)
197 bhp @ 7000 rpm BMEP ~183 psi ~93% of max
159 lb-ft @ 4800 rpm BMEP ~197 psi

2200 rpm separation

FA20 'base' (12.5:1 CR, port/direct injection)
197 bhp @ 7000 rpm BMEP ~183 psi ~88% of max
168 lb-ft @ 4000 rpm BMEP ~208 psi

3000 rpm separation

Hardly a far-fetched prediction, compared with the 13ish year-old BEAMS motor.

[serialk11r]

Okay you're saying this engine is tuned for a bit more grunt at a very low rpm to make people like SUB FT-86 "happy" (but really, do you think they will be? lol. turbo motors have ridiculous torque peaks that are more than 20% greater than torque at the power peak), since it has only one cam profile it won't be very optimized for high rpm operation. Got it.

[Dimman]

Originally Posted by arghx7

I have actual manufacturer BMEP and torque curves for the 2GR-FSE, which is the IS350 engine using the D-4S system. I'd like to hear what your predictions are before I reveal them.

All I can estimate is ~196 psi at 4800 for peak. Figure of ~178 psi at the 6400 power peak will be less accurate.

(JDM engine figures will be ~198/~182)

[Dimman]

Originally Posted by serialk11r

Okay you're saying this engine is tuned for a bit more grunt at a very low rpm to make people like SUB FT-86 "happy" (but really, do you think they will be? lol. turbo motors have ridiculous torque peaks that are more than 20% greater than torque at the power peak), since it has only one cam profile it won't be very optimized for high rpm operation. Got it.

Turbo motors mess around with BMEP figures because the intake density is variable. Ie, they are not always pushing the same boost levels at their torque peak and power peaks.

As for making SUB happy, I think it would make a whole lot more people than just him happy. 17 more lb-ft 2600 rpm sooner, without affecting the hp peak would be awesome. The lighter the car, the more significant things like this are.

The other thing is that they are promoting this as an entry level RWD driver's car. Getting a bunch of drift-wannabe ex-fwd Scion noobs into a peaky RWD car may not be ideal. Figures I'm predicting would be a more predictable motor.

[Ryephile]

Dimman, you're getting awfully bent over this tight torque and power peak. ....and yet, we still don't know the torque curve beyond those two points in the leaked data sheet.

151 LbFt @ 6600 RPM and 197 HP @ 7k RPM are logical figures. You can go and wax lyrically for years about the BEAMS engine, it doesn't change that this Subie engine, from the two torque data points we have right now, is more like a low-revving Honda S2000 engine more than anything else.

[serialk11r]

17 more is a tiny bit over 10%, okay. More powerful N/A engines in cars designed to run ~5 second 0-60 have probably like 30-40% greater acceleration at any time. Turbos peaking at low rpm can have like 30% greater acceleration at one point in the rev range (after spooling) compared to another. If you look at how the torque gets up from idle the difference is even bigger, the torque can double from idle to 3000 rpm in some cases. People complaining about power want the "push back in your seat" feel, and somehow I doubt 10% is enough for that.

[Dimman]

Must... continue... pointless... internet... argument... despite... auto-show... web-traffic...

Originally Posted by Ryephile

Dimman, you're getting awfully bent over this tight torque and power peak. ....and yet, we still don't know the torque curve beyond those two points in the leaked data sheet.

151 LbFt @ 6600 RPM and 197 HP @ 7k RPM are logical figures. You can go and wax lyrically for years about the BEAMS engine, it doesn't change that this Subie engine, from the two torque data points we have right now, is more like a low-revving Honda S2000 engine more than anything else.

No.

The differences in torque and hp peak have to do with intake air speeds, the drop comes from losses from friction at higher speeds. And the speed is based on the relationship of the valves/intake ports the bore/stroke and rpm. More lift/duration on the cam will make the valve/port appear bigger. This is why you see the power band difference between the two BEAMS motors.

BMEP (roughly...) is the net how much air gets stuffed into the combustion chamber, and how well it burns, minus any frictional losses.

CR bumps are a very standard way of increasing BMEP. As is a well designed combustion chamber, fuel distribution, and well mapped ignition. (this is for more pressure)

Moving it around involves playing around with wave tuning from exhaust and intake effects.

Reducing inertia and friction are the other ways (same pressure, less loss).

This is all from the 60's.

So this motor has more CR and from what Toyota claims in a D4-S article much better combustion characteristics. This also allows for reduced emissions. Yamaha was a co-developer of D4-S.

The FB20 was made to be very fuel efficient, and the approach Subaru took was through aggressive friction reduction.

Yamaha designed the BEAMS. It is likely not a coincidence that they chose the 86mm x 86mm configuration. They have data on the port and combustion chamber designs that produced the BEAMS output. Those are transferable, and a good baseline to start at.

The 2GRFSE manages 196 psi and holds almost 91% of it over 1600 rpm. And this is done without the benefit of ACIS or variable geometry intake runners (TVIS-type inertial system or TGVs), both of which can be used to fatten up torque curves. (Although we are a bit unsure as to if the FA20 will use either, but TGVs are a possibility.)



197 bhp @ 7000 is also not a number that is going to need retarded camming (200 degrees @ .050" with lift matched to valve sizing should be plenty), and due to dual VVT-i, overlap is variable anyways, and they have flexibility to play with their opening and closing points. So why they have to spin to 6600 for max torque just doesn't make sense to me.

So I'm not just pulling numbers out of my ass, and as a Toyota fanboi I feel I have every right to be choked if the 151 @ 6600 PEAK is correct, and not my wish of 'it maintains 90% of peak @ 6600'.

As for the F20C, part of the reason for its characteristics is that it has massive duration on its high lift cams. This car won't. It also doesn't have continuously variable cam phasing. This car will.

And since this is the internet, where people are never wrong (just their correct-ness is misunderstood) I'm going to be using the 'I guess Subaru DID do it all themselves' excuse and claim 'I would be right if it was a completely Yamaha/Toyota project'.

[Dave-ROR]

I hope the peak 151@6600 is correct (as I fully expect it to be) just so we can stop talking about BMEP numbers.

[Ryephile]

Dimman, you are putting entirely too much stock in BMEP comparisons. BMEP is a theoretical abstract for comparing dissimilar IC engines, it is not representative of actual cylinder pressures. I get where you're coming from. Yamaha should be able to make an engine with a torque peak BMEP of >200PSI, but for a variety of reasons [mainly emissions, durability, politics, market placement, and cost] they simply did not with the FA20.

HP peak is simply a function of RPM. Remember that horsepower is also an abstract figure to showcase torque through spacetime. In ICE's it also approximates MAF when VE is considered.

Here are the two torque data points we have in-writing at the moment:
151 LbFt @ 6600 RPM
148 LbFt @ 7000 RPM

Like I keep saying, and you keep ignoring; we do not know how the torque curve looks. The torque could be a shelf of 150 LbFt between 3500 and 6500 RPM....we simply don't know....and that includes you. You don't know the cam specs, the ignition timing, the AFR targets, or the VE map.

[Dave-ROR]

Originally Posted by Ryephile

Like I keep saying, and you keep ignoring; we do not know how the torque curve looks. The torque could be a shelf of 150 LbFt between 3500 and 6500 RPM....we simply don't know....and that includes you. You don't know the cam specs, the ignition timing, the AFR targets, or the VE map.

Agreed 100%

[Dimman]

Originally Posted by Ryephile

Dimman, you are putting entirely too much stock in BMEP comparisons. BMEP is a theoretical abstract for comparing dissimilar IC engines, it is not representative of actual cylinder pressures. I get where you're coming from. Yamaha should be able to make an engine with a torque peak BMEP of >200PSI, but for a variety of reasons [mainly emissions, durability, politics, market placement, and cost] they simply did not with the FA20.

HP peak is simply a function of RPM. Remember that horsepower is also an abstract figure to showcase torque through spacetime. In ICE's it also approximates MAF when VE is considered.

Here are the two torque data points we have in-writing at the moment:
151 LbFt @ 6600 RPM
148 LbFt @ 7000 RPM

Like I keep saying, and you keep ignoring; we do not know how the torque curve looks. The torque could be a shelf of 150 LbFt between 3500 and 6500 RPM....we simply don't know....and that includes you. You don't know the cam specs, the ignition timing, the AFR targets, or the VE map.

I'm trying to take those into account as well. As well as design philosophy. The Altezza was promoted as the AE86 spiritual successor of its time. The BEAMS example is used because it's a motor that was placed in that spiritual successor, by the same company, with the same bore/stroke. I purposely selected the low-cost variant for my example. If all they did was throw an increase of compression plus all their improved tech at making this thing hit cost and emissions requirements it should at LEAST match the previous numbers. Cost-wise with the rev limit where it is, this thing isn't going to have massive loads on the rods/bearings, and that saves money on durability targets. And by using the same injectors on the Yamaha co-developed 2GRFSE that support the same amount of hp/injector that this car has they can reduce costs there.

As for the cams, I have no problem making the statement that they are going to have a ton less duration than the FC20's high-lift profile.

The brochure numbers don't reflect kaizen...

[Ryephile]

Kaizen...I had to look that up That'll be my international vocabulary lesson for today.


Let's look at this engine topic from another angle; what other current production naturally aspirated engines that are sold in the USA [i.e. level emissions playing field] have noteworthy BMEP? I'll omit cars over $100k, as the budget for applying technology is hardly a fair comparison.

Car___________________Engine___Torque____BMEP
Acura TSX_____________144 cuin, 172 LbFt = 180 PSI
Honda Civic SI__________144 cuin, 170 LbFt = 178 PSI
BMW ///M3_____________244 cuin, 295 LbFt = 182 PSI
Corvette Z06___________427 cuin, 470 LbFt = 166 PSI
Ford Mustang GT________302 cuin, 390 LbFt = 195 PSI
Hyundai Genesis R-spec__307 cuin, 376 LbFt = 185 PSI
Infiniti G37 IPL__________225 cuin, 276 LbFt = 185 PSI
Lexus IS350____________210 cuin, 277 LbFt = 199 PSI
Mazda Miata____________122 cuin, 140 LbFt = 173 PSI
MINI Cooper_____________98 cuin, 118 LbFt = 182 PSI
Porsche Cayman R_______207 cuin, 273 LbFt = 199 PSI

FT-86_________________122 cuin, 151 LbFt = 187 PSI

It's right there in 4th place....not crap, just not the best. Just for giggles, it's better than the Lexus LFA's 182 PSI.

[serialk11r]

I wish people would stop using BMEP, it doesn't give any indication to any relevant cylinder pressure levels. Can we switch to specific torque? lol. Nm/L is easier to digest. 100Nm/L or worse is typical for street cars, a nice round number. Ferrari gets 118 in the 458, the highest number for a street car, thanks to their very expensive continuous lift control and disregard for fuel economy.

[Ryephile]

No, it's not literal cylinder pressures, it's approximate mean cylinder pressures. Nevertheless, its the same kind of abstract statistic; the answer however takes units out of the equation [no confusion between metric and SAE or whatever]. Feel free to convert the list I just did and you'll get the same relative picture.