Variable displacement V8 - thoughts?
 

I got the bright idea today to see if I could retrofit a V8 engine (specifically, the 4.7L engine in my Dakota) with variable displacement. It would be beneficial to be able to recover between 8% and 20% of fuel economy in this manner.

Electronics is not a concern for me, since I have a BSEE and some experience with building and programming automotive engine electronics. It should be a simple matter to cut power from the necessary ignition coils and fuel injectors. Granted, the stock engine computer very likely would throw codes for multiple cylinder misfires and loss of continuity of said coils and fuel injectors, but this could be overcome in time with a standalone engine controller (like the Megasquirt).

It seems the big obstacle to this project would be shutting off airflow to the deactivated cylinders. The proposed solution came to me as I was reading ways on how this was done with engines that already had this capability.

I remember that there is a company out there who came up with an adapter kit to fit a Ford 4.6L intake manifold to the Chrysler 4.7L engine, and it was basically a pair of adapter plates that mated the intake manifold flanges of the Ford intake manifold to the intake flanges of the 4.7L heads.

It should be possible, then, to fabricate adapter plates to go between the stock 4.7L intake manifold flanges and the 4.7L head intake flanges. These plates would have butterfly valves to shut off airflow to the desired cylinders. This should work inasmuch as it would allow the O2 sensor to continue to correctly read the lambda of the exhaust gases, instead of trying to read exhaust gases excessively diluted with fresh air.

Thoughts?

Would it be possible to move your oxygen sensor upstream on the exhaust such that it only measures the exhaust on one bank of cylinders? This would allow you to run the air through the the non sparking bank of cylinders as normal.

I had thought of doing just that, in order to completely shut off one bank. Ideally, in order to have a more-or-less evenly running engine, you'd want to have -X-X-X-X, where the "-"s correspond to deactivated cylinders, and "X"s correspond to active cylinders. This is to maintain, as much as possible, the existing engine harmonic arrangement.

However, the problem with this approach, of shutting off a whole bank, is that my engine has a firing order of 1-8-4-3-6-5-7-2. That translates to a LRRLRLLR pattern. If I were to shut off one bank (e.g., the left bank), so as to facilitate using the O2 sensor on the other bank, I would have -RR-R--R, which looks very uneven, and could shortly lead to destructive bearing wear or even a snapped crank.

Seems like peeing into the wind to me.

Variable displacement systems shut down intake and exhaust valves on the disabled cylinders. If it was as easy as adding a butterfly valve to the intake, that's the way it would have been done OEM. If the valves are still working, there will still be pumping losses from the intake tracts and from sucking/pushing into the exhaust tracts.

So the trick will be to disable valves. I believe it can be done at home, but it would be quite involved.

The firing order for the 4 cylinder 2009 yamaha R1 motorcycle has these intervals; 270° –180° – 90° – 180° which is actually the same as you would end up with. They did this for better traction.

Don't let the perfect become the enemy of the good.

OEMs don't use butterfly valves to deactivate individual cylinders because they cannot allow the exhaust to be contaminated at all with inlet air before the pre-cat O2 sensor(s) can read it. There will be more leakage with butterfly valves than with the poppet valves in the heads. This is the reason why OEMs tend to use the poppet valves - the OEMs can comply with EPA regulations and the Clean Air Act. Granted, some vehicles inject air into their cats, but that's a different matter.

That aside, pumping losses will offset much, if not all, of the gain provided by causing the remaining cylinders to work more closely to their full potential. Therefore, minimize these pumping losses, and you minimize the drag induced on the engine itself.

I would have to design some of the hydraulic lash adjusters to be able to collapse more completely than normal, on command. After that, I'd have to figure out a way of keeping the rocker arms in place, since they are only secured onto the valvetrain by the lash adjuster at one end, the cam at the center, and the valve itself at the other end. Sounds like a lot of trouble.

On the other hand, it's a compromise to put butterfly valves on the intake side. I should still see some meaningful gain.

...contact the SAE for papers on (A) GM's variable (V8-V6-V4) engine (fiasco) and (B) Chryslers's Multiple Displacement System (MDS) Hemi engine(s).

...and, I'll bet, FoMoCo has dabbled in this arena too.

Sounds like a cool idea, but also quite involved.

My wife's '09 Impala SS has GM's 5.3L with cylinder deactivation and it does seem to work nicely on the highway. We averaged 29 mpg on a recent 1100-mile round trip. That was with two adults (and I weigh 240), 150 lbs of kids and the trunk loaded full-to-bursting with coolers and luggage going 75-80 mph the whole way (speed limit is 75).

That said, I'd be willing to bet that you'd be better off just changing out the rear gears to something taller. It would have a similar effect...reducing the number of cylinder fires per mile.

the best, most reliable, version of variable displacement is a smaller engine with a turbo. I wish it were different.

Variable displacement schemes are often talked about on EcoModder, but I don't think we've seen a DIY success yet. Probably because of the enormous engineering effort required to extract fairly small gains.

Using intake parts from an engine that leaves the factory with variable displacement is a great idea. A better one, that involves less fabrication and doesn't require a custom controller, is to transplant a variable displacement engine + ECU into your target chassis.

The biggest problem with MegaSquirt is it's not OBD-II compliant. In my state, I would be unable to install it in a road-going vehicle '96 or newer.

Best answer so far. Reduce revs per mile and increase load.

The add some aero to improve even further.

regards
Mech

A lil mind game: figure at best it won't match the claimed +8 to 20% and at worst it won't be any worse than stock. 19 x 1.08 = 20.52; x 1.2 = 22.8. My WAG is, if it works at all, don't expect more than 1 to 2 mpg.

23,000m/yr @ 19mpg = 1210 gal/yr; @ +1 you might save 57 gal/yr; +2, 115 gal/yr. 1210g @ 2.80 = $3388.

As an example: $500 30 mpg car yields 767 gal/yr @ 2.80 = $2148.

You could find an econo beater that pays for itself in less than six months then it's all gravy after that.

The Cadillac V864 was actually a beautiful design that was plagued by a woefully inadequate engine computer. This idea wasn't tried again until Mercedes experimented on their V12 engine at the turn of this century. GM recently reintroduced their AFM system, while Chrysler has MDS, and Ford has VDE. Honda and Mitsubishi also have variable displacement systems.

The systems appear to fall into two categories, which are fairly easy to implement. The first system is suited for pushrod engines. Allow selected lifters to completely collapse, and the corresponding valves just won't open. The second system is suited for overhead cam engines where the rocker arms are fixed in the center by a pivot bar. Each valve to be controlled gets two rocker arms for that valve. The first rocker arm follows the cam, and the second rocker arm operates the valve. Normally, the two arms are mechanically linked by a pin. However, remove the pin, and the two arms are no longer mechanically linked, and the valve again won't open.

However, neither system is suited for my engine. This overhead valve engine has the rocker arms underneath the cam The rocker arm pivots at one end on a hydraulic lash adjuster, is actuated by the cam at the center, and actuates the valve at the other end. It's held in place solely by placement on the valve stem and lash adjuster and cam. Remove the cam, and the rocker arm lifts right out.

Fine and dandy, but where would I find a decent diff gear that will fit my truck without me having to do a lot of machining work to make fit, and without having to spend a total of much more than $1000? I've got the tallest available 3.55 gear already.

Um... Your post above highlights the widespread fallacy of using MPG as opposed to L/100km. I actually happened to do this mind game, and I got substantially different results.

235.2 / 19 MPG = 12.4 L/100km

Now, assuming an 8% improvement, that means I used 8% less gas to go the same distance. In other words, I went the same distance using 92% of the fuel I would have normally used. So, 12.4 L/100km * 0.92 = 11.4 L/100km. And, 235.2 / 11.4 = 20.7 MPG.

Doing this with an wildly optimistic 20% improvement: 12.4 * 0.8 = 9.9 L/100km. And 235.2 / 9.9 = 23.75 MPG.

Therefore, I could save anywhere from 1.7 MPG to 4.75 MPG with this mod. And that's assuming 19 MPG as a starting point. With 21 MPG, which is what I'm shooting for now with my aerocap, the gains are a little more.

Finally, keep in mind that my goal is to improve the fuel economy of my existing truck, while retaining its existing power output capability.I don't want an econobeater, and spending thousands of dollars transplanting a different engine and transmission is similarly not an option. An econobeater can't haul a ton of gravel, nor can it haul 4 or 5 cans of recyclables and mulch at a time.

I realize that variable displacement has had a checkered past here on this board, and that you've seen your share of snake oil here. I do happen to believe, though, that I can get variable displacement to work on my truck. As an aside, I do appreciate the thoughtful arguments against this idea.

One extreme measure would be to run the oxygen sensor on one cylinder only and spark cylinders 1-4-6-7. Alternatively, if you had a 4-2-1 manifold you could run the oxygen sensor of two cylinders such as 6 & 7.

Instead of sucking air in or blocking the intake as that would introduce a restriction making the other working cylinders work harder you could try one of these options.

1. Reroute/route exhaust to the deactivated cylinders controlled by a butterfly valve, the exhaust would be ran from just a little after the O2 sensor before the cat. This would allow the O2 sensor to be happy as no new O2 in being ingested. The issue that might happen is leakage of exhaust gas when running on all 8 that could be reduced if the new line is added as a T with an additional valve to block the exhaust.

2. Is the opposite idea of the first. Let the engine take in all the fresh air it wants but modify the exhaust on the deactivated cylinders to vent to air. Once again use a butterfly valve so you can change the exit of the exhaust this way when deactivated the valve will move blocking off the original exhaust and direct the air out of a short pipe. This option should take care of the O2 sensor problem also as it will not see the unburnt air since it never makes it to the sensor. Doing this you might need to have the firing order even so the O2 does not have a large gap with out a reading. The other issue that there might be is fresh air being pulled into the exhaust or worse exhaust being vented to air. Air being pulled in would be easy to fix with a flapper at the end of the pipe but I doubt that will be an issue so you could add a second butterfly valve to the exit of the pipe to help ensure that is is sealed.

^ :confused:

Perhaps a slightly taller tire would help?

I wouldn't call 20.52 vs 20.7 substantially different. Besides, if the 8-20% figures assumed benefit are given based on the mpg system, then that is where they're "accurate" yes?

I didn't use the wildly optimistic 20% because it's... wildly optimistic.

If you are seriously hauling that much, the vast majority of the time, then OK. Color me amazed though! :eek:

Fair enough, and that's the end of that.

Sure it can. Get a wagon with a 2000lb tow rating, and a braked 4'x8' trailer. Unhook the trailer when your only payload is a body or two.

Now if you're hauling heavy cargo the vast majority of the time, then I agree with Frank's assessment.

Or I might go the route of a standalone engine controller running open loop, and cut the truck's engine computer out altogether during the time variable displacement mode is activated.

The idea is to make the running cylinders work harder. This will raise average pressure inside those cylinders, allowing them to have higher output, and making them more efficient in the process. Keep in mind that I am blocking the intake ports of individual cylinders, not the whole manifold.

Sounds like a fancy way of adding EGR to the engine. Not sure it'll work, if I understand you correctly. You're trying to find a way to not use butterfly valves in the intake ports of the deactivated cylinders?

This option would still cause the engine to do needless work, as it would simply become an air pump as far as the deactivated cylinders are concerned.

You don't know how many freekin times I've said that already.

http://i146.photobucket.com/albums/r...eadonwall2.gif

Well, like my students, this has gotten unreasonable. Good luck, I will wait to see the patents.

The way I look at it even if you get past the valving and ECU O2 sensor issues you have a couple hurtals to over come.

The balance of the engine is engineered for all 8 jugs functioning, it may/will run rough. (this may or may not be an issue mechanical and would be up to you if it is a comfort/annoyance issue)

The biggest one efficient wise is that you are still accelerating and stopping and accelerating a bunch of mass in the pistons and the of scrapping the rings along the cylinders of the deactivated jugs several thousand times a second.

Got anything against pulse and glide?

I expect to feel some difference from the engine if 4 cylinders are deactivated. However, I don't think this is such a big issue from a mechanical longevity standpoint, as the Chrysler 5.7L MDS system shuts off the exact same cylinders as I had considered, and the only thing I have read about is that some drivers complain of a low vibration between 1100 and 1800 RPMs when MDS is engaged.

It's called friction losses. Can't get around that.

However, it's not much of a concern. If the Big 3 can make variable displacement work on their production engines, and can get significant improvements in fuel economy, even with friction losses of the deactivated cylinders, then it's not much of a problem.

I think you meant to say "minutes" instead of "seconds," though. If my V8 were somehow able to make its pistons reciprocate thousands of times per second, I don't think it'd live for much longer than a few seconds.

It's unrealistic for a daily commute.

Try the following:
Chrysler / Dodge / Plymouth Dakota Rear Differentials Chrysler 9.25 Ring & Pinion Sets Auto Parts: YUKON GEAR,YUKON AXLE,YUKON,RING PINION,RING,PINION,GEAR,DIFFERENTIAL,CROWN WHEEL,CROWN PINION,CROWN,CHRYSLER 9.25,CHRYSLER 9 1/4,CHRYSLER,RICHMOND GEA

They list a 3.21 gearset for the 9 1/4 axle for less than $300, and I seriously doubt a shop would charge more than $300 to do an install. I have that ratio in my 1988 Ramcharger with 235/75r15 (and 275/50r15) tires, but it's in an 8 1/4 axle. The 9 1/4 in my Ram has the 3.55 like yours, but with 265/65r17 tires.

I honestly think you're barking up the wrong tree if you want cost-effective efficiency improvements that will be reliable. As you state the valve arrangement of the 4.7L isn't suited to tried-and-true methods of variable displacement. Any home-brewed solution for the existing valve arrangement is likely to be insanely expensive or unreliable enough to cause major engine damage, or likely both. You may be-LIE-ve you can do it, but I wish you luck while recommending you divert attention elsewhere. Maybe modifying the heads to accept the double-rocker style actuators would have potential for reliable success, but again it would be quite expensive.

This is soooo beyond the shade-tree mechanic on so many levels.

If you want displacement on demand, go find a 5.3 or L92 GM engine. Their system works really well. Don't expect more than a 3-5% improvement in MPG.

You want mileage, swap in a Cuymmins 4BT3.9 diesel. There are now commercially available kits for that swap. Expect a 25-50% improvement in MPG.

In order to test whether variable displacement will or will not work on my engine, I will move ahead on my own. I will fabricate spacers out of 16 ga Home Depot sheet steel that will simulate the closure of the intake ports of cylinders 1, 4, 6, and 7. I will then install said spacers onto the engine of my truck, with gaskets. I will then electrically disconnect the fuel injectors for cylinders 1, 4, 6, and 7. I will then perform an A-B-A test with this setup. I will then figure out what fuel economy gain I received, if any, and will report back here.

I do appreciate the advice and suggestions you all have given, though. Thank you. However, I do ask that from now on, that any advice or suggestions given in this thread directly relates to supporting this project.

It will be interesting to see the results! It's a neat project to tinker with, regardless of how much or how little it saves you.

-soD

Try it w/o the spacers first. I have a feeling putting spacers to close off the intake port will simulate the highest pumping losses you could see for those cylinders (throttle is not open). Just disconnect the FIs for whatever cylinders and see if there is any improvement. Make sure you can keep test conditions as consistent as possible. For instance if your Dakota stays in third w/ the TCC unlocked after the FIs are disconnected during a highway run, when you're doing the run w/ the FIs connected do it in third w/ the TCC unlocked.

Just disconnecting the four FIs will not work as the oxygen sensor will pick up on a lean AFR and the engine management will try to adjust by adding extra fuel.

Even if 3.55s are the tallest factory gearing, there may be other gearsets available to fit your axles. Do you know offhand what axles are in that truck?

Remove the O2. Closed loop versus open loops shouldn't result in a huge difference in fuel consumption.

The GM AFM system works very well, but required enormous engineering resources and two decades to develop.

You think you can DIY, you are kidding yourself.

It would be far easier and cheaper to find a 5.3 GM engine and swap it in.

positive thoughts, dave, positive thoughts.

Chances are it won't see the kind of gains that the factory setup did/does. It may not even work. But much will be learned. Hopefully the knowledge will be shared with us.

So it won't be a waste, no matter what, and it could be really very cool. As long as the expectations are reasonable, and it's an enjoyable thing to put together, I fail to see a downside.

Even blowing up a motor usually teaches you something; and it's not exactly as if Dodge V8 motors are that rare. And odds of that happening are pretty small if you pay attention to what the motor is actually doing.

-soD

The OEMs see a 5% improvement because they hardly use the cylinder deactivation. They don't use it at idle, or under moderate load. Mostly just during cruising. An ecomodder would use it much more often.

They deactivate the valves so that they can run the 4 cylinders with the best balance, so that the customer only sees a little vibration at certain rpms. They have to keep the air from coming in or out of the chamber so it doesn't mess with sensors.

On my Mustang, there are separate O2 sensors for each bank, so I could turn off fuel injectors from one bank, fake the signal from the O2 sensors, and the ECU would stay happy. The air from each bank joins after the sensors, so O2 from the one bank shouldn't get to the other bank. Honestly, the only problem I see is how bad the vibration is.

For your project, what you are looking at I think is basically individual throttle bodies, maybe sealed up better. The 3 valve 4.6l has flaps inline (Charge Motion Control Valves - Tumble generators) that could be turned into full fledged valves/throttles.

http://image.mustang50magazine.com/f...r_controls.jpg

I didn't realize you already had the 3.55 gears. I say go for the cylinder deactivation if you feel you are up to it. It sounds like you have the skill set needed, and what a feeling of accomplishment it would give you to finish it and have it work.

SlideWRX has a good point about the OEMs under-utilizing it. They have to be very conservative to prevent the owners from thinking something is wrong with the engine, whereas an ecomodder is much more forgiving.

I sure wish I could see results of "motoring" an engine- reverse dyno'ing it by spinning it with an electric motor and measuring power requirements to spin it. Data that shows how much power it takes to move that piston with valves going and with valves disabled would be most useful.

That would be cool! I bet the OEMs do it, at least every once in a while.

It could be inferred from how much current it takes for a given-sized electric motor to spin the combustion engine to a given RPM. A motor intended for small EVs might be ideal.

You would, ideally, want to have the engine parts at their operating temperatures. You could run the engine for a while until it got to operating temp, then shut down the fuel and ignition and let the electric motor go.

This is sounding like something that a resourceful person could do at home...

-soD

The OEM I used to work at did that for the entire driveline.

Do you mean something like a starter motor with a disabled ignition and injection system? Yeah, its on the motor already.