VVL is OP?

[Leedar]

Technology in Automation is meant to reflect real world behaviour within a reasonable margin of error (IIRC the ‘official’ number is 10-15%). Ultimately my question is about whether VVL in Automation reflects real world behaviour within the margin of error–I’m not an engineer, mechanic or even really a car guy, I have no idea whether complicating factors mean the ideal advantages of VVL are heavily negated in the real world. The best I can do is look to the markets, and they don’t seem to think VVL is all that important compared to turbos, VVT or direct injection, which makes me wonder if VVL is over-performing in Automation.

This all said, I gather that the new intake system may nerf VVL performance somewhat by making builders have to choose an optimal RPM range to have good airflow (or maybe have to pay extra for variable intakes).

[KA24DE]

Very good point about engineering, Strop.
I should have went in that direction when I mentioned that different companies have different schools of thought.

[strop]

@Leedar

Ultimately my question is about whether VVL in Automation reflects real world behaviour within the margin of error

I’m no engineer either but there’s a general answer to the math of it: I do believe that it’s ‘approximately correct’ in that considering the range of VVL technologies out there, the values given here ought to give out the ballpark values. There was a time when the math for various other things did need adjusting, and there may be in future as you point out, so the devs do remain open to that kind of question.

However it’s only reasonable to expect Automation to reflect real world behaviour assuming that real world behaviour is entirely rational and forward thinking. My earlier post didn’t explicitly state this, but we shouldn’t assume that at all, because as I did say earlier, rational behaviour when it comes to engineering for a consumer market is still a retrospective judgement. If anything, you’ll note that Automation assumes that demographic attributes themselves remain static and the lite campaign very crudely models major landmarks in Automotive development by banning leaded and having different fuel availabilities for different regions and also having an increasing safety requirement, but none of those things will adequately cover the kind of real world reflection you’re looking for when it comes to the impact of a recently implemented technology.

[Killrob]

What you guys should try is to make a company, start when VVL is available and save after you’ve designed your car. Then you play it out with having VVL in your design (say, 15 years just that car) and note your capital and score. Then you reload (i.e. you get the exact same economy) and choose to not have VVL in your design, play 15 years again and then compare. That gives a better picture of how OP or not the tech is.

[Deskyx]

Yes so it means VVL is not that OP when some guy can scrape 40 MPG out of a pushrod in modern times. VVL is just so you can get both performance and economy.

The Assoluto Crinale has a VVL V12 and gets worse milage than a KHT Eau Rouge which doesn’t. VVL is not OP.

[Leedar]

I don’t mean OP as in artificially prohibitive to other gameplay choices, but OP as in out-performing reality by more than the margin of error–aside from contextualizing costs to produce (production units, material and engineering time), this is independent of how good an option it would be to use in game. That is, if VVL is ‘OP’ in reality then of course there is no problem with it being ‘OP’ in the game, like how PFI is better than carburation in almost every way (in quantitative terms).

Judging whether VVL is OP by testing in campaigns pre-supposes that it is behaving accurately and that the concern is only the costs to use it. By the way you are speaking you sound confident that VVL is behaving similar to how it does in reality, so I’ll take that as VVL is not over-performing in terms of mechanical performance.

Edit: I would add that I guess it is wrong to say it is purely about mechanical performance. The bonus to drivability from the torque curve is significant and not necessarily based on reality–would buyers be so significantly preferring a flat or otherwise linear torque curve? Most modern engines don’t have such a torque curve and would likely get a bad drivability rating either due to a mountain shape with NA engines or due to turbo spooling.

[strop]

I think we’ve refuted your question of OP as in out-performing reality a number of times now, so you’re going to have to bring up some hard numbers to clarify your assertions so we can really look at it.

@Leedar

would buyers be so significantly preferring a flat or otherwise linear torque curve?

YES A THOUSAND TIMES YES

Or, to really rephrase your question in a manner that’s real-world relevant, buyers who drive their car in traffic would love having torque down low and up high. Just ask any Australian car reviewer. What they tend to say very closely resembles what I’m about to say.

I’m going to draw abstract illustration to real world experience here, since that’s what the Automation metric attempts to do, take that as you will:

The red line shows the engine with which my car is equipped: the 1.8L i-VTEC. My car is a daily and I live in a central metropolitan area, so traffic is frequently busy. I commute to work in peak hour every day, which means I have a lot of stop and start. This means I spend a lot of time in the 1000-2000rpm range.

This also happens to be, as you can see, the range in which my car is the most anaemic. There’s poor throttle response, poor acceleration, and poor fuel economy. It literally makes my car harder to drive in traffic and this would hopefully be reflected in the amalgamated assessment of the drivability score not being as high as it possibly could be (Automation assesses the impact of the torque curve as well as performance as part of the Drivability amalgamation). Side note: Throttle response is chiefly the domain of sportiness calculations.

But wait! I hear somebody say: the R18A motor has VVTL! Yeah, it does. But the engineers did a half-assed job: let’s just say in simplified Automation terms, the cam profiles would be set to at least 35 and 50, as opposed to where it might have been more useful for daily driving, like 20. On the plus side it’s a hoot on the mountain pass, but my Civic is a fatass hatch with el cheapo suspension, and I don’t do the old spirited mountain pass driving very much. When it comes to that rare moment where I feel like I don’t want to get chopped at the lights and I push the car past 2500rpm, I can feel the valve profiles switching over, precisely at the point where I no longer have to use it. Annoying.

[Leedar]

You seem to misinterpreting what I made this thread for. I’m not asserting that VVL is definitely over-performing [edit: mechanically, e.g. specific output], but rather asking whether VVL as it currently exists in the game is behaving comparable to real world application. Nobody has answered that explicitly.

For instance, in Automation it is favorable to achieve a strictly linear torque curve, which can be done with VVL by pairing a low valve timing with a very high valve timing and setting the redline near peak torque, yet I am under the impression that in the real world such a configuration is not favorable, but that that unfavorability is not modelled/graded in Automation. When you look at dyno charts, the vast majority of engines (VVL included) have decaying torque as you reach redline, because real people aren’t looking for mathematically linear torque, but rather the general impression of linearity (and that lower RPM torque means higher average power). In the real world a turbo engine with extremely low RPM spooling and a wide flat torque band would be considered to have ‘a flat torque curve’, but in Automation the driveability would likely be penalized for spooling and decay towards redline.

Put in other words, I am concerned that VVL was superficially adopted into the game, but that the game does not reflect how it would be used in the real world, and that simplifications in how the game models and grades engines encourage unrealistic engine designs. If people are happy with that, or that what I think are oddly effective design choices are completely realistic, then whatever, but I thought it was something that stood out but hadn’t been discussed.

[rcracer11m]

I think part of the reason behind automation engines slightly outperforming real engines is that real engines dont use such drastically different cam profiles as we do in automation, with as big of a difference as 20 and 100 profiles whereas real engines would be more like 30 and 60. Part of this is because of simplicity in switching over from 1 cam to another. Such a major change in profile that we have in automation would be very difficult to have happen smoothily.

Also this smaller difference in profiles also explains your torque curve idea as well. As real engines are running something like a 60 profile compared to automations 100 cam real engines simply cant make as much torque at high rpm so naturally begin to develop a downward slope in their torque curve

[Leedar]

Right, that is exactly the sort of concern I mean. Because of how Automation quantifies qualitative-tinged measurements (torque linearity, engine smoothness), this isn’t exactly the minor divergence from the real world that simply looking at specific output might imply.

[strop]

You’re right, I’m glad that your specific query there was answered. In my defence your thrust wasn’t even hinted at until at least post eight and then the thread kind of got distracted from there, but anyway

In short the perfectly flat torque curve would be very drivable, because you can imagine that throttle response would be consistent throughout the rev range. But in a separate concern, if it isn’t dropping off at the end you’re wasting potential power and consistency at the top end so other metrics will suffer I.e. performance index. How that relates to real world behaviour obviously depends on real world applications, but suffice to say on the occasions I give my car a good booting, a big dip in power at the wheels when I shift up at the redline would be quite disconcerting as I’d perceive a sudden loss of acceleration.

@Leedar

In the real world a turbo engine with extremely low RPM spooling and a wide flat torque band would be considered to have ‘a flat torque curve’, but in Automation the driveability would likely be penalized for spooling and decay towards redline

Yes this is a separate issue… The turbos behave as they do in the eighties LOL and so there’s no control over how your turbos spool except through the strict mechanics. Having a flat torque curve past a certain point is often accompanied by a massive torque surge the moment the turbos spool, and if you look at the detailed stats, that’s where the drivability penalty comes from. Turbos are due for a massive overhaul, this is fairly close to the top of the queue post UE4 transition.

[Daffyflyer]

If your question is relating to if VVL is behaving as it would in real life, the answer is basically yes. (at least within the limits of realism of the current Automation engine simulation). All it’s allowing you to do is pick the best out of two different cam profiles throughout the rev range. Now you can argue that our simulation of the effects of cam profile is overly simplistic, and that is something we want to improve in future for sure, but it’s pretty close.

Are the engineering and production costs too low? Maybe, I’m sure there are quite a lot of parts/options that will require future balancing once the campaign is more finished, and it’s entirely possible that this is one that will get more expensive.

Why don’t lots of car manufacturers use it? I suspect the main answer is that it’s only really useful if you’re trying to get as much power as possible out of a given engine capacity, and in most cases that’s just not the goal of engine design. If you look at something like the Corolla, they have either the 1ZZ-FE (105kw @ 6400 rpm, no VVL) or the 2ZZ-FE (141kw @ 7600rpm, VVL). This is a pretty typical use of VVL in my mind, and in this case the 2nd cam is only used between 6200 and 7900 (or 8200) rpm. Why wouldn’t you use that on the base model road car? A bunch of extra complexity and expense, and the only advantage you’re gaining is an extra 36kw at sky high RPM. Sure it’s great for a sports car, but I can see why they’re not doing it to every road car engine they have.

Really VVL is there to allow you to use what would other be an entirely unsuitable for a road car cam profile, but still retain good emissions, idle quality, power at low RPM etc. If you’re not trying to make a high revving sporty motor it shouldn’t really be worth your time most of the time. (and if that’s not the case currently, we may need to look at making it more expensive to use)

Now, systems like BMW’s Valvetronic are a different matter, as they are infinitely variable in lift (but not Duration), they let you do awesome stuff like use valve lift adjustment instead of the throttle, which has lots of economy benefits. That’s probably a system we should add eventually.

If you’re talking about problems with the stats (Like drivability, sportiness etc.) and their judgement of what makes a good engine, all I can really say is that they’re based off a combination of our own personal judgement of what makes a pleasant engine to drive, and what characteristics car reviewers rate positively in various types of car. If you can see any glaring flaws in how those work I’d be interested to know them, but I think they’re pretty close to an accurate description of what people like in a car

[Daffyflyer]

I think a fairly flat torque curve is achievable in the real world with VVL as well, here is a dyno sheet from a stock Honda B16, which is a 1.6ltr VVL 4 cylinder used in small high performance hatchbacks and such, and represents a fairly typical application for cam-switching VVL of the type we model in Automation.

it’s making pretty close to 100ft-lbs from the beginning of the run (around 2700rpm) until about 7500rpm.

Our torque curves might end up being a bit flatter than this, but that’s because we don’t take into account the compromises of port sizing and intake manifold design that would in reality still favour a specific RPM range regardless of what cam was used. That’s something a future overhaul of the engine calculations will be aimed at fixing.

With regards to " the vast majority of engines (VVL included) have decaying torque as you reach redline" that’s because, given the limitations of having a limited number of gear ratios, you often get more acceleration by staying in a lower gear after peak power than you would by shifting up to the next gear.

So most likely if your torque curve is NOT dropping off at the end of the graph, you need a higher redline and to leave the rest of the engine alone. The reason you might not want to do that though is if you’re already spinning the thing as fast as it’ll hold together, in which case you could end up having to set the redline in a place that’s not ideal for max performance.

[Leedar]

I see. I don’t recall seeing a dyno chart like that B16’s, but it appears extremely different cam profiles have been used in real engines, so the only concern for me now is how Automation evaluates engine characteristics. Maybe I can mock up examples to show how Automation favours VVL for most applications (IMO).

[TrackpadUser]

The other thing to check would be the effects on desirability.

I’ll have to do some testing on my end as well, but while you do end up with the best of both worlds when using VVL, I wouldn’t be surprised if it didn’t really matter for most demographics, with the pure economy engine scoring just as good or even better than the economy+performance VVL engine.

[Daffyflyer]

Well the ideal thing you’d be trying to acheave with using two different cam profiles (performance wise anyway) is making it make the highest average power across the rev range, thus you’d pick cams that did that, as they appear to have in the B16

I’d expect that a VVL engine WILL always be better than a non-VVL one though, in almost any application. Sure it’ll be much more important improvement for sports cars, but even a pickup truck engine could gain a broader power band, better emissions and economy, and more top end power by using VVL. The only reasons I can think of that it isn’t used like that in real life, is:

It’s cheaper and much simpler engineering wise to just make the damned engine bigger.

Economy/workhorse-ish engines usually have a pretty low redline anyway, so one cam profile will usually give a pretty acceptable powerband anyway. If you don’t care how it goes above 6000rpm you can likely choose one profile that works pretty well from idle - 6000rpm.

If you really do care about more power AND economy, these days you just turbo things. It’s not that much more complex than VVL and has much bigger economy and power gains (when combined with engine downsizing) and can give you a REALLY flat torque curve. (Currently not possible in Automation as our turbo sim is a bit sucky)

I think my main question is if the engineering time and production time for VVL is high enough. as that’s what will determine if it’s worth using in an application or not.

[Leedar]

I’ll say beforehand that:

1. I excluded a turbo intentionally, because turbocharging favours VVL heavily in terms of torque curve.
2. The non-VVL engine isn’t strictly optimised for Automation, but represents a typical real world approach to engine designs of this type (for economy cars), with peak torque around 4000 RPM and redline around 6500 RPM.
3. Both engines are biased towards maximum undersquare-ness, and have the same performance index and reliability. If reliability is relaxed, the VVL engine will likely pull ahead further in terms of stats advantage.
4. Gearing for both is set at 100 km/h @ redline in 3rd gear, 51 spacing, 7 speed gearbox.
5. The cars are identical but for the engine; airflow is just above requirements.
6. Exhaust choked for maximum performance index.
7. Only difference between engines is bore/stroke, compression ratio and valve train–same materials used.

A more desirable engine could probably be made for that car/market without using VVL (I managed it with an anaemic very small I6), but I challenge the idea that you could beat all [edit: many of ] the stats of the VVL engine while spending less resources. In this case the VVL’s only weak points are emissions (which is negligible), and 10% higher engineering time. 0-100km/h is also slightly lower, but isn’t really significant in terms of marketability, and the non-VVL engine is faster around a race track (something not measured in terms of the in-game market).

For this specific case/market, I guess it boils down to ‘is 10% more engineering time worth 3% more desirability?’. I would stress again that with the current turbo system, the desirability disparity is actually quite a bit greater if you are trying to make a maximally competitive car.

[Daffyflyer]

Hmmm, the differences in engine performance, and car desirability seem reasonable, although I would expect the the VVL one to have a few more production units needed, which it doesn’t there, so I guess we may need to see to that.

[TrackpadUser]

Could we see the detail screen for driveability?

[Daffyflyer]

Oh, good spot, looking at the screenshot it appears the VVT engine is 8v, not sure if intentional?