Home | Wiki | Live Chat | Dev Stream | YouTube | Archived Forums | Contact

Direct Injection May Be Getting Special Treatment


I searched and found no threads that had contested or suggested this and so I figured it might be a potential design oversight.

It has been the policy of Automation that design options which would require drastic changes in cylinder head geometry are chosen at the family level rather than at the variant level. These features include camshaft configuration, number of valves, and whether or not the engine supports variable valve lift i.e. having two different cam profiles. It occurred to me today that direct injection is not receiving this same treatment though it possibly should.

My interjection is this:

Direct injection should be a feature explicity chosen for an engine family to support just like VVL and should NOT be chosen as a feature of a variant.


Because just like camshaft configuration, valve number, and VVL, direct injection requires extensive design consideration for the head itself unlike other fuel system options which only affect the design of the intake plenum. Heck when swapping between carburetion and single point EFI, the intake plenum may not even change hence why so many late '70s and early '80s cars have it. Direct injection is significantly more complicated however.

Direct injection first of all requires consideration of fuel injector placement and associated port geometry. This is especially problematic for multivalve SOHC engines which are already make spark plug placement difficult. There is also high pressure fuel rail routing to consider since direct injection systems run at extremely high fuel pressures on the order of 100 atmospheres.

Also, because the air-fuel mixture is no longer helped in homogenization by the intake plenum, special considerations need to be made for piston and combustion chamber designs to ensure proper mixing and even burn.

Because of these factors, adding a direct injection system to an engine already in production that was never designed for it would require some heavy modifications if not a clean sheet head design. Thus it seems to be receiving special treatment compared to especially VVL.

So a question:

Has this discrepancy been considered? If it has, then out of curiosity what was the reason that direct injection is allowed the leeway that it has? If it has not, then what are the devs’ thoughts regardless?

If this is indeed an unnoticed discrepancy, my suggestions for addressing it would then be one of the following:

  1. Direct injection, like VVL, should be chosen as the feature of a family and only then may it be used.

  2. Direct injection could remain a variant option but if direct injection is added in a facelift / is not originally included in the initial production variants, adding direct injection should incur an engineering time penalty proportional to the number of variants in production. (This option would be as if the engineers chose to use the same cylinder head for all variants. And as any engineer knows, modifications in one thing often have unexpected effects on things that you thought would be inconsequential.)

  3. Direct injection could remain a variant option but incurs a flat higher engineering time if not one of the original production variants, and a flat higher production units cost if produced together with non-DI variants since it would need a different cylinder head. (This option would be as if the engineers chose to use a different cylinder head just for the DI variant)

Families, facelifts, and limitations to variance

Good post, nicely written and indeed a “problem” with the game.

You are correct in that DI should not really be an option unless the Family is engineered for that specifically. However, as you point out yourself to some extent, that would mean that you’d have to have a separate option in the Family choices to accommodate for DI, like in the case for VVL.

That is not really acceptable from a UI point of view, so other solutions need to be explored. Our solution is to make the design “overlap” between MPFI and DI small, i.e. only parts of the software can be used, but everything else is “new”. That makes a switch to DI basically the same as designing a whole new engine with the coming engineering time calculation changes in the Lite Campaign. DI becomes the big ticket item.

That definitely balances the option vs. other fuel systems and instead of barring you from changing to it, we make it as ridiculous a prospect as it would be IRL. I think that is a good compromise.


Edit: btw, MFI is not the direct injection variant, so there is no overlap to DI from that.


While we are at this, I would like to double down on another thing DI is seemingly missing VS it’s real life counterpart.

A substantial bump to maintenance costs. DI engines have a very real problem with valve coking, BMW and VAG have been dealing with it since the 2000’s by having customers pay for an expensive (600USD-1000USD) walnut blasting process which cleans the fresh air side of the intake valves.

Toyota, and some manufactures are now running dual injection, and have both DI, and MPFI on the same engine. (See GT-86/BRZ) and this design policy is gaining traction and will be commonplace on ICE engines in a few years, as Port injections main advantage is that fuel is an excellent solvent and is keeping the fresh air side of the intake valves clean.

I propose that while we are re-calibrating DI, perhaps we should increase it’s service costs as well, and perhaps even consider adding DI/PI combo to late game (Not needed if we aren’t going to run the simulation past 2020 of course)

IN short: DI should see a service cost increase to reflect how it actually works IRL, as it requires a considerabal re-design of the PCV system to slow the process, but so long as an EGR, or VVT system is used, it will be impossible to negate the problem with out a dedicated port injection system

Example of problem below: I can attest as a Direct Injection engine owner, this IS a problem, I have already cleaned my valves with only a paltry 20,000 miles on the vehicle. I have been considering running water injection to help prevent this in the future.

(Picture is not mine, just a good example found online of the problem DI is having in normal useage)


It could be designed perfectly. You maintain the block, and you only change the head.


Yes. That’s an issue on DI engines, not on indirect injection systems.


double post m8.

plus that’s exactly what they’re talking about.

in automation we “treat” a major head change as a new engine family. or basically, you can’t. there’s a lot of engine families irl that have wildly different head design. we can’t do that in game.

and yes that is an issue on DI Engine BUT THAT’S WHAT WE’RE TALKING ABOUT HERE.


I agree DI needs an additional nerf beyond its engineering time, service costs would be a reasonable place to do so. That would make MPFI the de facto standard for anything but engine focused (including City Eco) and premium demographics, which is how it should be. I put it up on the ToDo list.


Aren’t they still being increasingly used across the board IRL?

(Unless you mean earlier direct injection engines being either economy-focused or premium?)


GM has been using it in their trucks, beyond the premium lines for a few years now. Ford now has a hybrid version (combination direct and port fuel injection) for the 2018 Mustang. I am sure there are others, those are just the ones I am aware of.


only post 2015 afaik. now they are on cars that costs $10k USD or less


Yeah, the Coyote 5.0 is dual injection (On the Mustang and F150), and the ecoBoosts are all Direct Injection too.

Plus when diesels come in to play down the road, direct injection will have a very clear market!


That must be a US-Thing (the walnut thing) though, i have not met anyone in Germany who did this. Here its a service offered by professional engine rebuilders to improve engine power and economy (or rather, restoring it.)

While coking is a issue on DI-Petrolengines, its not a thing they have to clean, it wont ruin your reliability, just economy and power in the very long run. Most people and specialist point to low fuel quality, sloppy oil changes and bad oil quality or issues with the whole engine if coking harms your engine at mileage as low as 40kkm. BTW Mazda has nearly no coking issue on their SkyActiv.

I personally would not give DI a nerf via service costs, feels unjustified… If you need to redesign stuff in your engine to have it work and less coke, it should be via engineering time.

Make them unsmoother, they inject fuel in a handful of sprays causing inharmonics (unlike MPFI), especially in the first minutes of running when they regulate out their ECU (Search Golf GTI FSI Nageln)
Lower reliability: Instead of a small fuel pump they need a high pressure pump and different kind of injectors. Especially in the early days of the tech they were not too reliable. Small impurity in the fuel can cause clog ups and malfunctions too, MPFI just shrugs and works.

And the obvious more engineering time.


Definitely not just a US thing - just to backup Lordred here I’ve seen a good number of VW FSI engines with issues that were totally blocked up, along with the Mitsubishi GDI motors in both the Mitsus and Volvos.

It’s so bad on even well maintained cars that Mitsubishi began offering that walnut blasting treatment as a warranty thing in the first 3 years as cars that only did shorter runs were seeing it at stupidly low mileages. Doesn’t help when they start adding EGR systems into the mix.

As an example, friend of mine’s Golf GTI was so bad the ports on the head were closing up with the amount of carbon buildup. It had the cleaning treatment and then it seemed surprised so much it ran on 3 cylinders intermittently for a few hundred miles until it sorted itself out.

I see where you’re coming from with the argument that it should be an engineering time thing to make it better, but to me that’s where the dual systems have come from like the mentioned Toyota and Ford systems. Indeed VAG have also started fitting this system to the newer TFSI cars. If you have to use better oil, perform more frequent servicing or have the top end decoked, definitely seems a service cost issue to me.


Quite the case. Just to tack on even more, Hyundai-Kia have had issues with accelerated valve wear and injector clogging on their DI engines thanks to high temperature lean burn conditions and sensitivity to fuel impurities. Granted the former of these two is more an issue with the engine cycle itself but one of the most attractive things about direct injection is better fuel atomization which allows for ultra-lean burn.

This is completely reasonable as UI clutter is already a problem with Automation. But to me this one actually has an easy solution. VVL currently occupies a whole section of the engine family tab for a boolean option; this is a collosal waste of space. What if the VVL box was replaced with a “Advanced Technologies” checklist where you could select via checkbox which of the advanced feature sets, such as VVL and DI, you want your engine to support?


That may also be the solution to the long requested “balance shafts” option people have requested. Off topic but it seemed relevant.


I’d like to give this post all the support I can.


We’re deliberately moving away from check-box lists, as you see with the driver assists options, that would be quite breaking the UI flow. Aslo: no feature creep please! :slight_smile:
It would have to be the current listbox but with


There is room for that, but we’re not going to go there, just makes things overly complicated and doesn’t add much to gameplay over the solution now already implemented in the game.

Balance shafts have been discussed a lot here already and their implementation is not going to change, feature creep that too. We’re happy with the current implementation. Maybe in Automation 2.0 or something we’d include them and the DI rules for engine families.


Of course. I don’t mean to suggest that any of this should take precedence over core development work on say the campaign because we could nitpick balance issues to death but it wouldn’t really help the game be any more complete. My only intent with all of this was just to point out a potential balance issue and suggest some ways of eventually fixing it.

My one point of contention here is that the checklist legitimately made no sense for the driver aides because they are rarely if ever standalone technologies. As a guy who works on autonomous vehicles and robots I know for a fact that if you have what it takes to implement stability control, then you have more than enough to implement traction control and ABS and the differences would be almost entirely in software. Thus, accumulation made far more sense then custom selection. Heck I don’t think I can name a single production make that has TC but not ABS or ESC but not TC. Yeah, that made no sense. But VVL and DI are very much standalone technologies; the design requirements of one have barely any overlap with the other, so a checklist does make sense.

BUT! Yes. This all comes with a very big “BUT”!

You’re the dev. And if you say its not happening then its not happening. I can accept that :slightly_smiling_face:.

And in order that we don’t feature creep ourselves to death (since as a software guy myself, I am painfully aware of how easy it is to do and how toxic it is to projects), I find your solution perfectly acceptable. Long term, of course. Campaign before nitpicking!


Here in Brazil he have engines from Ford that are just modified Ecoboost engines without the turbo and the direct injection.

We have an aspirated 3 cil 1.0 85hp that is just the 1 liter Ecoboost without the turbo, without the direct injection and of course with lower quality parts that don’t have to withstand the additional pression and power from the turbo engine.

We also have the new “1.5 dragon” aspirated engine, that people say come from the 1.5 ecoboost.

So the best way to a modern engine family work in game is:

  • If the family DO NOT have direct injection, you can’t add it;
  • If the family DO HAVE direct injection, you CAN take it out to use a multipoint injection in a cheaper lower spec engine. Just like in real life. The DI engineering is done, but you should be able to just make it a multipoint.


I agree that we should have dual injection in the game!

The new V8 5.0 Coyote from Ford also come with both DI and multipoint. This helps increase cruising mileage while also delivering the fuel needed when you go flat out. Better engine, at a cost in engineering and material parts.


The Coyote isn’t the only one from Ford like that, but yeah, it’s a really cool solution to Direct Injection’s problems. Having trouble fighting compression to put in fuel? Put in fuel before compression. Having trouble with intake valves dirtying up? Shoot some fuel at them to clean them up.