A 4-stroke 4 cylinder engine which, as well as operating normally, can also operate with two active cylinders feeding their exhaust into the other two cylinders which are ganged together as the large expansion chamber of a compound engine<link>.
In a typical I-4 ¹ (which has pistons 1&4 and 2&3 opposite on the crankshaft<link>), we can transfer the exhaust from one or the other active cylinder into both expansion cylinders, once per revolution.
Cyl 1) Intake.. Compress Combust Transfer.. C2+3) Exhaust. Expand... Exhaust. Expand... Cyl 4) Combust Transfer.. Intake.. Compress
Even if there's just enough extra energy to make up for the friction and reciprocation losses from the two otherwise non-working cylinders, it's an improvement.
The difference between this and a bog-standard engine is... - extra plumbing and insulation for the transfer pipes/passages - a handful of valves or gates, for above - a slightly more complex cooling system ²
---
¹ while the familiar four-banger makes for easy visualization, a narrow-angle V4 or an I/V compromise would probably be best in practice, in order to minimize the distance between active and expansion cylinders.
² the expansion cylinders should be heated by the active cylinders.-- FlyingToaster, May 18 2014 I-4 Inline 4-cylinder engine animation http://xorl.files.w...m/2011/03/img11.gifanimated .gif [FlyingToaster, May 18 2014, last modified May 19 2014] Compound Engine http://en.wikipedia...iki/Compound_engineThe expansion cylinder is always larger than the original of course. For the post we expand into 2 cylinders at the same time to achieve the same effect. [FlyingToaster, May 18 2014] Variable Displacement Engines http://en.wikipedia...riable_displacementSounds a bit grand for "turn a couple cylinders off" but there you go. [FlyingToaster, May 18 2014] Compound IC Engines http://www.douglas-...ndIC/compoundIC.htmMostly just designs, all from 1900'ish. [FlyingToaster, May 18 2014] Opposed piston engine http://en.wikipedia...posed-piston_engine [spidermother, May 24 2014] Free-piston linear alternator engine http://en.wikipedia...n_engine#Generators [spidermother, May 24 2014] If one cylinder is going down while the other is going up, then the net volume of both plus interconnecting tubes remains constant therefore no expansion can happen therefore no work done by the combustion gases.
Perhaps if you made the cylinders offset a little? Would work on a 6 with three pairs 120º apart?-- pocmloc, May 18 2014 When 1 cylinder is pumping (going up), 2 are expanding (going down). While the two "hot" cylinders are going through a complete 4-stroke cycle, the expansion cylinders are splitting their time between servicing the hot cylinders in 2x 2-stroke cycles: expansion from cyl 1, then an exhaust, then expansion from cyl 4, then an exhaust, rinse and repeat.
The volume of the connecting tube(s) is an outright loss, as is the bit of volume left over at TDC (the piston head doesn't go all the way to the cylinder head of course), so the expansion isn't a perfect 1:2 .
I'm not gonna think about a 6cyl right this minute: the problem is that while one pair of cylinders are indeed at TDC or BDC at the same time, the other two pairs are at 120deg and 240deg respectively, so you have to change horses 1/3 or 2/3 of the way through or something like that. With a 4 or 8 it's easy, at least in inline configuration... hmm maybe it wouldn't work for V's at all.-- FlyingToaster, May 18 2014 OK, yes, I was imagining the cylinders in pairs. But you are right, if 2 expansion cylinders match one exhaust cylinder you are sorted.
If this worked why aren't there dedicated double-expansion IC engines like there are double- and triple-expansion steam engines? With really big cylinders for the expansion stages?-- pocmloc, May 18 2014 <link> There were many designs in the late 1800's and early 1900's: typically they were for two cylinders on either side of a huge expansion cylinder. This included designs by Diesel and Deutz, and were mostly stationary engines.
They don't seem to have been too efficient, but the article doesn't state what the criteria for comparison was (overall weight ?, displacement ?).
Bear in mind that this design isn't for a full-time expansion engine. The purpose is to give a small engine another sweet-spot for efficiency, at the low end of power requirements, where it would otherwise be quite wasteful.
[edit: I've been pitching the inner cylinders as the expansion cylinders, but less vibration would be had by using the outer ones. Noted in the post, and added a cooling system note]-- FlyingToaster, May 18 2014 I'd have thought that the valving and pipework for exhaust gasses would be tricky, complex and expensive. It would make heat management a little more interesting too.
Instead of using lots of complex galleries and valves, we could just route all the exhaust gasses to one place, i.e. a turbo. What's the advantage of using a piston?-- bs0u0155, May 19 2014 I'm pretty sure that I would find the design tricky and complex once it got to the stage of making a blueprint, but I'm also pretty sure that people who know what they're doing wouldn't feel too challenged.
As far as expense is concerned, for a factory standard engine then we'd be talking in terms of tens of dollars for manufacturing cost increase over a comparable engine without the feature.
In terms of extra or different parts, there's a few extra valves or gates for exhaust transfers, and the cooling system needs to change its flow pattern: it'd be advantageous for the hot cylinders to keep the expansion cylinders warm.
//advantage of using a piston// The cylinders are already there and the pistons are already moving around anyways. The added power occurs during the active-cylinders' exhaust/compression strokes, thus lessening vibration depth. The added power is direct.
A comparison to a "turbo" is very ambiguous: the only case where it might be a valid question is ... "If there already exists on the engine a turbo, with all its associated plumbing and hardware (which is much more complicated than this), that is actually useful in some way at low power requirements, would it be worth it to ditch that at low-power usage (since the exhaust heat and pressure is now going into the cylinders) in favour of this system ?" No idea.
A full-time compound system (which is what all the designs from the 1900's were) might have merit though would best be realized with one big dedicated expansion cylinder, however this post is to increase the usefulness of a deactivated-cylinder system.
<ramble> In fact, one could envision (and perhaps post if one hadn't run out of steam and audience interest) an engine specifically designed and built for both compound and exclusive modes (rather than an add-on). Such an engine (still using a four-banger as a base) would have all the cylinder heads as close together as possible, perhaps a V4, or a combination of I and V designs where cyls 2 and 3 are V and 1 and 4 are in the same plane (between the V) or form their own V interleaved with the other one. These cylinder arrangements are to minimize the distance, ie:heat wastage, during compound transfers. </r>-- FlyingToaster, May 20 2014 //cyls 2 and 3 are V and 1 and 4 are in the same plane (between the V) // I already considered this and thought of including it in an anno above but I assumed 2 & 3 would add to vibration too much.-- pocmloc, May 20 2014 There's not much power, comparatively, coming out of the expansion (total guess 10% on a good day ?), so adding or subtracting from an existing vibration or creating its own isn't going to do much.
There's a few different configurations for IC pairs and expansion pairs (cylinders are numbered from front to back):
Inline4: 1,4 vs 2,3 IVI(2&3 share a crankpin/splitshaft): 1,4 vs 2,3 V4(4 crankpins): 1,4 vs 2,3 V4(2crankpins): 1,2 vs 3,4
The first 3 are basically the same thing with the same plumbing routing, just wedging the cylinders together in different ways. It's a tossup as to whether the IVI or V4(4crankpins) provides the shortest transfer paths. The V4(2 crankpins), which is the only one where it's back cylinder pair vs front cylinder pair, has the shortest crankshaft, so could have the least vibration properties.-- FlyingToaster, May 20 2014 //coming out of the expansion (total guess 10% on a good day ?)//
If it's only going to be 10%, use 10% longer stroke, Shirley?-- bs0u0155, May 20 2014 That's an optimistic 10% extra power after taxes, not 10% extra expansion. You might have to double the stroke length to squeeze that out of one cylinder by itself, all things considered, which is the point of having compound expansion; in this case one cylinder's volume of high-pressure exhaust becomes two cylinders' volume of low-pressure exhaust, doing work along the way.-- FlyingToaster, May 20 2014 For a few seconds I hoped [rotary] was back here-- sninctown, May 21 2014 Gee sorry to disappoint.
Wasn't he in the process of baking something last time 'round ?-- FlyingToaster, May 21 2014 I still think a turbo is a continuous, single moving part, solution to extracting energy from a heat/pressure differential that doesn't need any real plumbing and scales beautifully to all sorts of engines. Call me crazy.-- bs0u0155, May 21 2014 I've no clue how well a turbocharger works at low-power, but this method is certainly less complicated.
If you mean use a turbocharger to overcome the airflow problem in engines, the fact that Audi, BMW, etc., offer cylinder-deactivation engines, some equipped with turbochargers, belies that somewhat.-- FlyingToaster, May 21 2014 Interesting idea [+]
Regarding using a turbo to extract more energy during low power operation, it seems like it might be possible to put a brushless induction generator on the turbo (permanent magnets might overheat). Maybe you'd get enough power so you can ditch the alternator. And if you have your turbo spinning all the time, there won't be much lag when you stomp on the gas. Of course you loose a lot of energy having the compressor side of the turbo spinning all the time, so probably not worth it...-- scad mientist, May 21 2014 Would this suit a boxer engine? It might simplify phase and balance issues.-- spidermother, May 24 2014 So, no Danish pastries at all?-- not_morrison_rm, May 24 2014 For this idea a boxer is the worst case from a viewpoint of heat and displacement losses, due to the length of pipe required to join either of the cylinder heads on one crankpin with the opposite cylinder on the other pin. (The adjacent cylinder on the other pin is fine of course)
The I4 given as an example in the post is pretty bad as well. But either an IVI, or a V4 (2 or 4 crankpin) puts the heads right next to each other.
Viewed from above (various round things are supposed to be cylinders):
(1,4 vs 3,2) 1 2 3 4 o o o o - I4 o º o º - 4pin V • 8• - IVI
(1,2 vs 3,4) 88 - 2pin V oo oo - boxer
A 2pin tight V is the same as a boxer [edit: in terms of expansion flow topography]. There's a bit of displacement loss from the cylinders not being completely opened up or closed in synch with each other, but it isn't that much.
Remember that each of the 2 hot cylinders takes turns dumping into both of the expansion cylinders at the same time. for a V4(2pin) or a boxer, cyl 1 dumps its exhaust into 3+4, then, on the next crank revolution, cyl 2 dumps into 3+4. The closer the heads are the less heat/displacement waste from the connecting plumbing.-- FlyingToaster, May 24 2014 So it's a trade-off between efficiency and balance.
What about opposed pistons? Or, for the greatest versatility in cylinder placement and orientation, a free-piston linear alternator engine?-- spidermother, May 24 2014 [spidermother] I can read your links but I don't see how they relate.-- FlyingToaster, May 24 2014 I think [sm] might have been thinking about having 2 crankshafts? You could have 4 cylinders arranged side by side in a square, with 2 pistons attached to each of 2 side-by-side crankshafts. That is similar to your tight V4 but without the timing issues.
Or perhaps make it more like the V4 but instead of angling the cylinders, put the cylinders vertical, connected together by a Y shaped piston rod which connects to the normal roundyroundy piston rod. Might also need guide rails like on a steam engine piston rod.-- pocmloc, May 24 2014 Those types could allow the balance of a boxer, but the port proximity (and therefore shorter plumbing) of a V or inline. That's all.
[edit] Yes, just like [pocmloc]'s first paragraph. (Unfortunately, you would also have 8 pistons.)
And in the case of the free-piston engine, there is no crankshaft at all, so the cylinders could be arranged in any convenient configuration - such as a tight bundle with a :: cross section.-- spidermother, May 24 2014 Ah, I see what y'all mean... I thought you were funning me.
For a V there isn't really that much of an issue. A 20deg V (10 each way from the centerline), the actual cylinder displacement mismatch is only 1.5% (cosine 80deg = .9848...). But there's still a bit of extra pipe-length than there has to be.
In the IVI arrangement the relevant cylinder-heads butt almost right up next to each other.
//8 pistons// okay, I just saw it as 2 I-2's.
//::// mmm, actually ....
compared to O O O O
O O .O O is better-- FlyingToaster, May 24 2014 BWM tried this years ago, and moved on to the current deactivation method of locking the valves down to deactivate the cylinder, it then functioning as a spring & balancer.
The problem with the proposal is that it would require a more complicated crank and bearings that would produce more friction, burning more fuel than it would save.
But, more variable displacement on more vehicles would be great. Could we have first two cylinders Atkinson cycle, second two Otto Cycle, and replacement of the throttle valve with control of the duration of valve opening please?? - see FIAT's FIRE engines in the 500.-- sstvp, May 25 2014 //BWM tried this years ago// cite, please : I can find no reference to that. To reiterate the idea, in case you read it wrong...
A 4-banger has two pairs of pistons 180deg apart on the crankshaft. When two are up the other two are down. This allows all 4 strokes to be happening at the same time, one in each cylinder, giving two power strokes per crank revolution. For cylinder deactivation, one of those pairs is deactivated, leaving one power stroke per revolution.
The idea is, instead of deactivating 2 cylinders when in low-power mode, use them as expansion cylinders for the two that are still active.
For each turn of the crank, one or the other of the live cylinders will have an exhaust upstroke as usual (while the other live cylinder is on its compression upstroke). But...
Instead of dumping the exhaust directly into the manifold, a path is opened into both non-combusting cylinders which are both on the downstroke. This is the "compounding" bit, similar to steam-engines where the steam travels from smaller through larger cylinders, shedding pressure (and doing work) along the way.
My contention is that the extra power will overcome frictional losses of the two otherwise dead cylinders, actually _adding_ to the power.-- FlyingToaster, May 25 2014 Water could be sprayed into a live cylinder during its transfer stroke for some more pretty cheap and easy power.-- FlyingToaster, May 09 2015 i believe what the other posters are failing to take into account is the expansion of water to fill whatever universe is containing it-- 8o8, Sep 02 2015 The water spray option was an afterthought (though it'd probably work): the power comes from the expansion of exhaust gases, from one cylinder into two.
A bit non-intuitive since the sending cylinder is headed up to TDC when the transfer takes place, but the two receiving cylinders are going to BDC at the same time.-- FlyingToaster, Sep 02 2015 Hmm, actually the water spray would work out quite well. Unlike a regular engine where the spray has to happen during the combustion stroke or you'd just be blowing steam out the exhaust, in this one (while in compound mode) the water spray can start during the latter portion of the combustion stroke, after combustion is finished, or any time during the exhaust(transfer) stroke.-- FlyingToaster, Sep 05 2015 random, halfbakery