Transbox

[moroza]

the irony is that the two main benefits of the trans-box layout aren’t reflected in stats. transverse drivetrains are significantly less lossy than longitudinal ones, by 5-10% depending on whom you ask, but Automation gives little to no benefit to this. boxers have a lower center of gravity, which AFAIK doesn’t result in less body roll in Automation even though it should.

i extensively studied this layout ages ago when i first got started working as an auto mechanic, c.2011, much facilitated by working for a shop that got a lot of subarus and aircooled VW/porsche. if you’ve ever done serious work on VW/Audi in particular, i wouldn’t need to tell you that packaging disadvantages to a sideways boxer utterly pale in comparison to the shoehorn nightmares that are higher-spec VAG products.

the difficulties are two: clearance with the firewall, and clearance with the axleshafts. the former is directly related to stroke, rod length, and cylinder head type. the latter is related to how low the engine sits relative to the axle. a flat-four is wider than it is long, as is a typical DOHC flat-six. in a roughly square enginebay, it shouldn’t be hard to see that squeezing one in sideways is the same packaging hassle as front to back. sure, subaru sparkplugs and headgasket jobs are cramped affairs, but they’re manageable. transverse clearance would be helped by firewalls typically curving in their lower areas. as for axle clearance, consider pre-GM saab 900, where the engine sat on top of the transmission. also consider this photo. surely you can see there’s ample room to pass a driveshaft under the cylinder bank?

there’s more to efficiency than drivetrain orientation, and more to handling than CoG, but those are solid bonuses that other engineering can build on. as a relevant tangent, check out a Carrara GT tranmission, where the engine sits so damn low that they had to use a teeny little silicon carbide clutch, and the axleshaft outputs are above the transmission input.

also consider mid-crankshaft jackshafts, rope driveshafts, F-heads, shared engine/trans sumps, and countless more weird shit that actually was made.

crude sketches, based on enginebay dimensions measured on a BMW E34 shell. transmission measurements are likely based on a VW 02A as fitted in a 96 passat tdi.

[Hshan]

Okay, but… What would be the point, or rather, intended use for that config? Lower CoG is not that important in a typical “regular” car, so that’s one plus of the boxer wasted. The other is being short… largely irrelevant in trnsverse packaging below 4 cylinders in length. And the last I can think of is better smoothness, which is also not that important above the typical I4 level in a regular car (and if it is, just use balance shafts, simpler than a boxer with its two heads).

And what about performance cars, you may say? There the lower CoG is important. Yes, but you’re trading more even mass distribution for that I’d guess - you can put an inline leaned backwards in a transverse layout, meaning that even if it’s technically a bit in front of the axle, its CoG is likely right above it, or close to. You can’t do that with a boxer - most of the engine will stick out in front of the axle. Unless you mount it behind, which means moving the axle further forward from the firewall, which means longer wheelbase for the same interior, which means potentially lower agility and higher turning radius (the first bad mostly for performance cars, the second for regular ones). But if it’s in front, then you also have to consider that your engine is closer to the front bumper, meaning less space for the crumple zone - and basically all the space disadvantages of a longitudinal FWD, just to a lesser degree I guess. And most importantly - why would you even want FWD in a performance car? It would make sense if it’s based on a regular one, but in those the transverse boxer makes little sense too, and switching inline to boxer for performance derivatives makes even less, and honestly sound like an engineering & cost nightmare.

In short - yes, it’s possible, interesting, and has some advantages, but seems to be the answer to a question never asked.

[moroza]

One use might be in a high-performance mid-engined car, where the efficiency recovered could be 3 figures of hp/kW. Also potentially where width is more available than length, allowing a flat-12 that is no longer than an equivalent flat-4.

[Hshan]

Okay, in that use it would make more sense… But still, why boxer specifically? A “V” is still more compact (so potentially a smaller and lighter car), with a typical angle lighter, and with a wider one, while less compact, it may actually be better than the boxer for keeping the mass low. I’ve seen some explanation for that, but I can’t find it now - but basically there were reasons why boxer couldn’t feasibly be mounted as low as a wide angle V. I’m not sure if that wasn’t some example specific thing though But at that point it’s the purpose-designed engines territory anyway, so I guess any weird layout might appear, as long as it works well.

[Oreology]

There’s a lot of reasons this layout isn’t used. A lot of it is due to tactical allocation of investment in research and development and supply chain and market availability.

Packaging:

• it’s no better than any current layout. in fact, it’s worse.

Efficiency:

• you’re just being efficient if it’s a FWD and saving the energy cost of drivetrain and power orientation conversions. ‘high’ performance means people want AWD or RWD.

Power

• literally no advantage

Lower center of gravity

• ok, by a couple inches? dry sump lubrication could also help with that and mounting an engine lower. a hybrid powertrain with batteries could also help. not worth the multitude of tradeoffs for 1-2 inches.

---

Let’s say you’re Toyota/Subaru/Honda, or Ferrari.

Toyota/Subaru/Honda - mass consumer cars. Focus: efficiency & packaging

• Wanna save 5-10% on efficiency? Well damn, guess we should try to create an entirely new engine architecture, make a massive block, have two camshafts. Quite complicated… but fuck it! we ball! We can even add a CVT since we can’t fit a long manual or automatic transmission to it, and we can’t really squash it like a pancake since no one on the market sells custom-made transmissions??? Hmm… Why don’t we re-use our existing inline engine architecture and just downsize it and add a turbo? We can even look great in governments around the world for taxation and emissions purposes by going for a tiny engine that is extremely fuel efficient but still powerful enough!!! Or… maybe we could save 20-40% efficiency by just saying fuck it we ball and go hybrid??? Since there’s a dozen suppliers of power inverters and control electronics and batteries and motors and we literally can just buy it off the shelf without trying to develop an entirely new engine for 3-5 years… and what if governments just ban high displacement engines or ban combustion engines??? damn, i think my CEO will fire me if i told him i spent his entire R&D budget for this year on a new engine project

• Wow! This transbox layout sure can be packaged well. We can fit 6 cylinders here!!! But wait, it’s so wide? Also what is torque steer??? pretty sure having the transmission next to the left wheel is fine. the right one we’ll just use some super lightweight shaft since its all the way over there??? idk Empty space on the sides of the engine bay? that’s space we can use!!! Let’s just mount it directly to the sides of the engine bay. its called space efficiency baby. Oh we need all these extra auxiliary stuff, yeah we can mount it above the engine. Oh, there are oil issues in the gaskets from the cylinders being constantly in a horizontal position??? Damn, we gotta check it out then and remove it. Oh fuck, it’s so big and wide, we can’t get it out of the engine bay without removing all this extra auxiliary stuff on top of the engine. Ah damn, it’s so deep in there mate, do we just remove it from below? But then we gotta remove the driveshaft??? Wait so we can’t mount it that low anyway without excessive driveshaft angles between the CV joints and the wheel uprights??? but this is a FWD economy car, we will have to angle the engine so we can get it out properly. Can you angle it with a single crane to squeeze it out of the hood opening??? fuck, remind me again why we cant just use a simple, tiny inline 3 engine??? it’s literally an economy car, nobody cares. i dont wanna work on this no more it so fucking complicated.

Ferrari - supercars. Focus: performance

• mamma mia che cavolo è sto boxer!!! sembra una schifezza!!! facciamo i v12 e v8 e basta!!! dai, qualcuno vuole davvero un boxer?! ma che stai dicendo, transbox?! montato di traverso?! ma sei scemo?! noi vogliamo solo awd e rwd, fratello!!! come diavolo trasferisci la potenza in modo decente?! e come lo metti pure il cambio?! ah, devi farmi fare uno su misura solo per questo motore?! ma dai, ne abbiamo già un paio pronti!!! non ci penso proprio!!! non voglio perdere tempo, io voglio fare soldi!!! quindi dammi la soluzione più semplice!!!
• oh un motore boxer ha un baricentro migliore??? ok allora gli diamo un corpo in fibra di carbonio e magari pure il tetto in carbonio??? facilissimo lo posso fare oggi stesso!!! non voglio spendere 5 anni per un motore nuovo solo per migliorare il baricentro di 1-3 pollici lololol!!! e poi possiamo semplicemente montargli un pavimento a batteria e ritoccare le sospensioni comunque!!! il motore non deve letteralmente stare sopra l’albero di trasmissione o l’asse e visto che ha il carter secco possiamo montarlo bassissimo bassissimo!!!

[moroza]

@Oreology

The efficiency gain applies regardless of engine location, whether it drives the front or rear wheels. Power at the crankshaft is unaffected. Power at the wheels is improved modestly.

No claim was ever made that the gains are worth the cost of reengineering existing systems, rather than a ground-up design. This straw man of an argument is dumb, and the rest… any of those makers complaining about two whole camshafts? Seriously? No, no serious argument here worth responding to.

Spend a long weekend of afternoons replacing the transmission selector shaft seals on a 308, which pass through the engine sump, and then come talk to me about Ferrari and their maintainability ideals.

[moroza]

Boxer is usually the lowest possible engine. A wide-angle V can sometimes be a hair lower, depending on how much further from cylinder centerline the headers are versus the bottom end. Wide-angle Vs however are far more restrictive about cylinder count and firing order - only certain combinations work - while a boxer layout is a lot more flexible about being scaled up.

[Hshan]

He has a point though about torque steer and unequal ddiveshafts. I’m not sure how that would affect the rear axle, but I guess not in a positive way.

As for scalability, I think it’s worth assessing the boxer option vs the respective V or inline:

• 4: usually inline, simpler to build, much more convenient to service, likely lighter, uses mostly transverse space, which wouldn’t have much use otherwise (while the longitudinal space used by boxer’s width definitely could), although you can’t put as much above the engine (which is not the most practical place to put stuff anyway - heat and servicing)
• 6: inlines rarely used transversely, and for good reasons; Vs are more compact than boxers, lighter, space reasoning similar to I4, except it uses the same transverse space as a boxer, and again, a bit more convenient to service
• 8: at this point, why would you go transverse and deal with the gearbox issue? Because you have to put it somewhere - in line with the engine? So you have severely unequal driveshafts, not good. Next to the engine? Won’t work with a boxer, may with a V if you make the gearbox compact enough. Below? Now you’re nullifying boxer’s lower CoG, because you have to mount it higher. Also, as before with the 6, lighter, more compact and a bit more convenient to service
• 10, 12: basically like 8, but the gearbox placement issue is even worse. Also the only reason for putting those engines sideways I can see would be compactness, and in that regard Vs are better (unless you wanna seat ON the engine… but that nullifies the lower CoG)

So, to sum up, I can’t really see boxers being better in those scenarios, and even if they, maybe, wouldn’t be worse, why bother?

[donutsnail]

Here are some of the issues I see with this concept. None of these issues are to state that such a concept is impossible, but rather, that the sketches provided in the initial post exaggerate the layout’s compactness, and thus, its COG advantage and its packaging viability.

---

1st: The Transmission

First, the transmission. This transmission is depicted as too compact and once the transmission is actually in a housing, this location would greatly hinder the uptravel of the lower control arm, as they are being asked to share effectively the same space. The 2nd image demonstrates that as well; even scaled as-is and with no housing, the transmission, lower control arm, and the wheel itself, once turned and in its range of motion, are all fighting for the same space.

.

Let’s look at the vehicle that I think is the nearest example of a transverse boxer 12: the Lamborghini Miura.

To package this V12 transversely, Lamborghini placed the transmission just behind the engine block. You can see here how much larger this transmission is proportionally to the engine compared to the sketch in the original post, and further, that its V layout gives the option to mount the transmission this way without reserving even more length for the engine bay. Transverse I6s have taken a similar path as well. But this is a luxury the boxer layout does not have. Given the width is finite and we are already beyond the absolute limits with the layout in the diagram, we could follow the route of the Testarossa:

Now we have a flat 12, mounted longitudinally, above the transmission. This does, however, move the entire block much higher in the engine bay, raising the center of gravity, and trivializing an advantage presented in the original post for adopting such a layout in the first place.

However in doing so that longitudinally mounted transmission underneath the engine creates a pair of open spaces along its flanks for the exhaust manifolds:

This leads me to my next big question mark about the sketches in the initial post.

---

2nd: Routing of Intake and Exhaust

The provided image of a Subaru engine in the original post make a strong argument for the flatness of the engine.

However, it tells a different story when fully dressed. And this is even of the more compact NA SOHC variety:

@moroza

surely you can see there’s ample room to pass a driveshaft under the cylinder bank?

Well, not really, because we need space above and below the a boxer engine for intake and exhaust routing. The initial sketches show the cylinders and valvetrain, but do not depict the block itself, the head, nor the routing for intake and exhaust manifolds. And in the original sketch, even generously assuming the block and cylinder walls can be paper thin as depicted and that the head is quite compact, all with no issues, there is still an awful lot in contention for that real estate beneath the block.

Everything is packed so tightly that the exhaust from the front bank has no easy route out, its best option is to just squeeze out of the tiny space between the head and the center differential, then to fully wrap the differential to get beneath it and head towards the rear of the car. The rear driveshaft is contending with space for the center rear cylinder so much so that it is invading the space the rear bank head would occupy, let alone the exhaust manifold. Once again, this can be mitigated with a higher mounting point, but that is once again counterintuitive to the lower COG argument for the layout.

With modern fuel injection we can flip the intake and exhaust to have the exhaust run out the top of the head rather than the bottom, but it still doesn’t cure these woes, as the intake runners too, need to exist. We could make this engine reverse flow with exhaust and intake both utilizing the top of the block, but to do so we’re once again severely limiting the engine’s capability by forcing it to be reverse flow.

Another option would be to further offset the engine rearward from the transmission, giving the exhaust a place to go beneath the engine that isn’t all ate up with driveline, but we’d then be asking for quite a huge amount of space between the front wheels and the firewall.

---

My overall point to make here is not that the idea is impossible, nor even that it is unviable. My argument is only that the diagrams/sketches offer an extremely optimistic representation of how low and how compact such a layout could reasonably be, thus exaggerating the COG advantage of such a layout. The COG advantage would be severely mitigated by necessitating:

1. a more significant portion of the engine bay occupied by the full engine with block and heads, not just the rotating assembly, cylinders, and valves as shown

2. a not-insignificantly higher mounting point than depicted in the diagram due to real packaging needs of the transmission, exhaust, and intake. Ancillaries can be moved around to an extent to assuage packaging issues, but those aren’t the issue here, the intake, the exhaust, the transmission, and engine itself are non-negotiables, requiring space that the diagram as shown doesn’t budget enough space for.

To better drive the points of packaging home, look at this cutaway of a Subaru WRX STI:

Now lets look again at the original sketch:

The amount of vertical space budgeted in the real application absolutely dwarfs the sketch in the first post. The sketch implies the height of the engine is not even as tall as the brake disc. And consider as well where the Subaru’s engine is mounted in the vertical space in the bay. Even without driveline components running underneath it, just making space for the exhaust necessitates a higher engine mounting point, once again mitigating the COG difference between a more conventional layout.

No slander, I’m a huge fan of Subarus, have owned multiple and love them. The Boxer layout is just not as small as the original sketches want you believe.

---

For what it is worth as well, while transverse has less loss than longitudinal, 5%-10% less loss is quite generous. It could be argued that efficiency gain would be lost again and then some with the restrictions on intake and exhaust routing.

At the end, I think it comes out to less than a wash, chasing specific ideals at the cost of the overall package; a layout that cannot see the forest for the trees.

[moroza]

I wish there were a Superlike forum feature. Major appreciation for a thoughtful reply, @donutsnail . Thank you.

Those are totally valid criticisms. Some of them were new to me, others I had addressed in notes taken years ago but that I forgot about when posting this. The original design was a four or a six, and the twelve shown above was meant to be a back-of-the-napkin What If tangent, not a serious proposal as I perhaps implied it was in the OP. Sorry about that.

Even in the 4/6-cylinder version, the packaging was maxed-out for compactness, perhaps unrealistically so. But perhaps not - one thing not shown or mentioned is that the middle axleshaft, and transfer case shaft if AWD, would be integrated with the engine block. As for the manifolds and head design, they were based on a VW TDI engine (1Z or AHU), which is non-crossflow and manages to be great anyway. Both manifolds would point up, the underside of the head containing only a fuel injector and glowplug. The packaging is much different from a Subaru, which makes me realize something:

The argument for trans boxers has a lot more constraints than I’ve been mentioning, and that indeed it’s a flawed idea for most use cases.

In certain cases, I believe the concept still offers some inherent advantages, but you’re right that they’d 1. be less than hoped for, and 2. a lot less than that sketch. Thank you for deconstructing my argument.

More later.

[abg7]

@donutsnail

At the end, I think it comes out to less than a wash, chasing specific ideals at the cost of the overall package; a layout that cannot see the forest for the trees.

You hit the nail on the head right there - it’s too space-inefficient to be of any real use.

[MrChips]

Look, I’ll make this simple - if it hasn’t been done in the real world, it’s absolutely not going to be done in Automation. No exceptions.

[Djadania]

Aisi?

[Hshan]

Umm, but it can already be done in Automation, if I’m understanding the concept right The whole discussion stems from, I think, the view on this solution in challenges, because in at least one it has been used, and criticised. It’s not a suggestion of anything new, if I’m understanding it right (am I, @moroza?).

[moroza]

Very well, but where do you draw the line? A large AWD hatchback, nominally a wagon, with a longitudinal all-iron 60-degree V8 was in the current QFC. That hasn’t been done in the real world to my knowledge, as a production car. Or if it somehow has, suppose it were a V16. Or in the other direction, suppose it were a 90-degree V6, or a DAOHC 60-degree one, or a replica of something like an Audi A3 barring one minute difference that makes it, technically, have never been done in the real world. Which of those that haven’t exactly been done get exceptions for being close enough?

[Hshan]

I think Chips meant introducing new solutions into the game, not combining what we have already in wacky ways. I mean, FWD V10 hasn’t been done IRL and nothing stopped me from making it, and it from working surprisingly ok, but that’s just a combo of a V10 (realistic and already in the game) and longitudinal FWD (realistic and already in the game).

[MrChips]

As I see it from my perspective, it seems you’re kind of hinting that you’re looking for a re-balance of the stats to make this idea more viable from a gameplay perspective, which is where my comment came from.

It’s not going to happen because a transverse boxer engine is, in all honesty, a completely terrible idea that totally negates any benefits that a boxer engine actually has; that being, a low vertical profile, and a very short length from nose to tail, which gets it into some otherwise unusable places.

[moroza]

If that’s what I were after, I’d’ve said so. There are plenty of other things I’ve called to be rebalanced in the software, but this isn’t about the software at all.

[MrChips]

Then again, I’ll state that a transverse boxer engine is a terrible idea. If they were any good, we’d see them all over the place in the real world, and yet in the 140 years that the modern automobile has been around, I can’t think of a single example.

It’s maybe a bit presumptuous to think that you’ve discovered the One Weird Trick to making a car better, faster, etc. (Car Manufacturers HATE THIS!) that somehow literally everyone else has missed out on…

[moroza]

Rather than making assumptions about intent, refuting exaggerations, or repeating what we all agree on by now, would you answer my question?