DIY displacement on demand system project
 

Thoughts and feedback welcome on the following scheme:
Engine: 1994 Toyota Camry 3VZ-FE CAT (3L V6 non-VVT-I)
Transmission: 4-speed automatic FWD transaxle with lock-up Torque Converter
Odometer: 340,000Km
Fuel injection system: Sequential multi-point fuel injection EFI
Oxygen sensors: 3 total: 1 on each engine exhaust manifold and 1 sub-oxygen sensor mounted downstream of the Catalytic converter (CAT)
Firing order: 1-2-3-4-5-6
Current average fuel economy: 10L/100Km
The plan:
Through testing of the engine I have found that it still runs reasonably well on 3 or 4 cylinders for cruising.
I intend to have three modes of operation:
Normal (all 6) for start-up, warm-up and acceleration
V4 (4 cylinders) for high-speed cruising
I3 (3 cylinders) for low-speed cruising.
For 3 cylinder mode I will deactivate cylinders 2, 4 and 6 (the front bank) and for 4 cylinder mode I will deactivate cylinders 2 and 6. I have found these configurations the smoothest.
I also intend to deactivate the oxygen sensor for the disabled bank and the sub-oxygen sensor. The sensor for the running bank (cylinders 1, 3 and 5) will remain untouched.
The reason for this is it will either:
A: trick the control computer (ECU) into thinking everything?s normal and not just dumping more fuel into compensate, negating any economy gains; or
B: put the ECU into ?Limp-home mode?, which will cause the ECM to substitute pre-programmed values for the turned-off sensors, which since my engine is well maintained should run like normal when at operating temperature.
The method by which I will deactivate the cylinders/oxygen sensors is by rigging up a bank of centre-console switches to turn off power to the appropriate fuel injectors and Oxygen sensors.
The intake and exhaust valves will not be deactivated so the deactivated cylinders will act like air pumps.
The idea is to rig everything up and then do fuel economy tests in various modes to find out what effect it has on fuel economy. But before I do I want to make sure the engine will handle everything (i.e. by not blowing up), both short and long term, so would like opinions/constructive abuse on whether it?s a good idea.
Potential worries to consider:
-uneven wear
-uneven cooling
-deactivated cylinders not being lubricated properly
-burning intakes valves due to no cooling fuel passing over them
-extra oil consumption due to lack of sealing pressure on deactivated piston rings
-shorting or arching when flicking the switches, at best blown fuse, at worst blown ECU
-mechanical failure of connecting rods, pistons, bearings, engine mounts
-Fuel economy gains offset by engine overcoming pumping losses

Limp mode unfortunately defaults to very rich mixture and very conservative timing. You are confusing limp mode with open loop. Open loop is when the ECU ignores the O2 sensor and runs straight off the fuel and ignition tables. This happens any time you mash the gas to the floor and wait for the RPMs to climb past a certain point. It also happens any time you are below operating temperature or before your O2 sensor has had enough time to warm up.

Whether you're burning fuel or not, the cylinders will receive the same lubrication. All your injectors stay closed every time you let off the pedal in gear above 1000 RPM. They turn back on when you drop close to your idle RPM. So lubrication on deactivated cylinders is not a problem. Oil circulation is mechanical. You could turn the crank by hand and it would pump oil.

When a cylinder compresses air without burning fuel it decelerates. I'm not sure but it seemed like you were assuming that since there's just air flowing in and out of the combustion chamber there's no resistance besides drivetrain losses and the friction of the wheels.

You can't deactivate an O2 sensor with a simple switch. The ECU will sense something wrong and throw a CEL. Instead what you want to do is emulate an O2 sensor with normal readings. This is not easy but there may be products on the market that do this. The signal from a narrow band O2 sensor is constantly bouncing from one voltage to another, approximately .4 volts to .9 volts. That is what your ECU wants to see. If it sees nothing at all, or one constant voltage, it's going to throw a CEL and go into limp mode.

If you don't want electric spikes across your switch you simply bridge the terminals with a diode facing backwards. You could also switch relays and put diodes on the relay terminals facing backwards.

If one or two oxygen sensors ?stop working?, would the ECM use the remaining one to control the engine normally, or go into open loop mode or ?limp home? mode?
For example mine has three?two on each engine exhaust manifold and one after the CAT (the sub-oxygen sensor). I would switch off the sensor on the deactivated bank plus the sub-oxygen sensor and keep the sensor on the running bank connected.
I ask because I have disconnected the OS on the front exhaust manifold, with the engine running, without the CEL coming on.
Also I knew there would be drag from the deactivated cylinders, but there may still be fuel economy gains to be had which I may or may not find during test drives.

Another idea I've brainstormed is to disconnect permanently 2 of the oxygen sensors (OS) and leave the one on the running bank connected.
I would then split the siginal from that sensorl into three to feed all the OS inputs into the ECU.
Would this work?

Well, that's probably the way I would go about it. I'm plotting DOD on my V6 voyager, and wondering if I would add 2 O2s to do the same thing.

Now THAT's a good idea...if you couldn't split the signal from one, just install a second one next to it and use that one to read the active side and feed the ECU that data for the inactive side. Neat idea!

I think I'll just wire the one Oxygen sensor into the three inputs. May do this in the weekend and see how it affects performance/economy before I go any further.
Another thing I've been concerned about is uneven loads on the crank with one bank shut-down. I've read that my engine model started off Toyota's trend of overbuilding thier engine blocks, cranks and rods but all the same I don't want to do this if it's going to destroy my engine. Also, when cruising, the Honda VCM system shuts down an entire bank of cylinders so obviously it can be done...
Does anyone think this will be a problem?

"The intake and exhaust valves will not be deactivated so the deactivated cylinders will act like air pumps."

This will be a difficult problem to solve. Perhaps add a valve that blocks off the intake tract to one bank of cylinders? This way the cylinders will be working in a vacuum.

Another thing I've been concerned about is uneven loads on the crank with one bank shut-down. I've read that my engine model started off Toyota's trend of overbuilding thier engine blocks, cranks and rods but all the same I don't want to do this if it's going to destroy my engine. Also, when cruising, the Honda VCM system shuts down an entire bank of cylinders so obviously it can be done...
What is everyones' thoughts on this?

Look at a picture of how the pistons/rods line up on a crankshaft in a V-type engine, you'll get a better feel for the answer to your question.

My gut feeling: In I-3 mode the imbalance shouldn't be too bad. In V-4 mode it might be a bit much.

My gut is also that you'd be okay as long as you keep the loads evenly distributed, run in I-3, don't try V-2 with cylinder 1 and 2, do 3-4 or 1-6, 2-5 instead (though that would probably be real rough) V4 could be 1-2, 5-6 or 1, 3-4, 6 or 2, 3-4, 5. You probably wanna just stick with a simple I-3 though.

Roadwarrior...what are your plans for the voyager...you mentioned earlier that you were planning a DOD system.
What arrangement would you have and how would you set up the switches?

I'm just going for I3 on either bank, so I'd have a switches for cutting out either side. I am just going to play around with it as I go really... one advantage I have is that the O2 sensor is right at the Y junction of front and rear exhaust manifolds, this means that the higher pressure and temperature exhaust flow from the active bank will tend to expand around the O2 sensor, and shove aside the "fresh air" coming from the inactive bank. If it was any further down the tube the flows would mix better.

Hopefully I'd work towards a mostly automatic system that triggers on load, temperature, vacuum etc... so then I'd have switches for V6, I3 either bank and auto. Whether I'd set the auto up to alternate banks or not, depends on what I see when I play around, whether one side "wraps round" the O2 sensor better, whether one side will go cold too quick...

I'm also thinking that ideally I want to put on a dual throttle body setup, with ability to close the TB on the inactive side, this motor seems to give little drag when engine braking with the throttle closed at lower revs, I'd equate it with turning the air con compressor turning on, maybe ~2HP for the whole six, so that would be about 1HP drag with one bank in vacuum. Whereas sometimes shifting down with light throttle applied (i.e. not giving it enough pedal to rev match) seems to give twice as much engine braking as closing the throttle all the way.

Since I'll be messing with the wiring anyway, I'm also pondering putting in forward biased diodes that can be switched in, in series with the injectors... the point of this would be that silicon diodes drop 0.6V, and lower voltage to the injectors increases the "dead time" slightly, meaning that the actual open time will be reduced, i.e. changes the pulse width without the ECU knowing about it. This might be useful to lean back the injectors if the ECU sees too much O2 and puts in too much fuel. However, some O2 sensor trickery would probably be preferable, hence why I was thinking I might need to stick an O2 sensor in each manifold and switch between them according to shich bank was running, and average, or switch back to original in 6 cylinder mode.

Here's a quote from P.70 Popular Science, January 1977 about Ford's early DOD efforts on an inline 6, that gives us some clues...

I'm planning to use that as a guideline for the automatic version and read the TPS for throttle angle, (or rig a cam and a microswitch) read the TC lockup status instead of checking it's in high gear and up to temperature and up to speed, and read the MAP sensor, or add another for checking vacuum.

Like your plan.
I'm thinking of just keeping mine purely manual...literally displacement on demand (DOD)...though I may make up my own acronym hehe.
Maybe Variable Cylinder Control (VCC) or Variable Displacement System (VDS).
Also, at this stage I'm not planning on sealing off the inactive cylinders...hopefully will still see a FE inprovement. if not I'll look into it.
Also in response to theclencher, it is not just of matter of unplugging the injectors and going for a drive because, and I've tried this, it makes the car veeery slow off the mark, horable to drive untill up to speed and is propably not good for the engine under such loads. If these ideas are going to work the injectors needs they need to be switched off at cruising speed where they will have the greatest effect on economy. The system will also need to deal with the Oxygen Sensors somehow because to not would propably negate any gains due to the ECU compensating by adding more fuel.
I think the debate now is weather any of this will relise any usefull FE gains and this is something that can only really be established through testing and trial and error.

Yeah, really needs to warm up on 6 cylinders before you switch to 3. Although if gas doubles I might just unplug 3 and remove the lifters and knit a woolly sweater for that side of the block.

Gas? Double? Never!!!! LOL.

IIRC on the production DoD systems that GM and Chrysler are using, they have to close valves on the disabled cylinders to prevent the engine from pumping the extra air into the catalytic converter, which combined with the unburned fuel from the operating cylinders would cause a very lean condition in the cat and cause it too overheat. Or am I way off base here?

That's possibly more necessary for new vehicles due to cats getting more delicate and having less palladium and platinum in, spread over more area...

If it's a problem could always remove the CAT...I need a new exhaust soon anyway

Ya know, before you guy's get to carried away with your gut feelings and cylinder deactivation, you might want to research some older threads which have pretty thouroghly hashed all of this out, previously.

In short, unless you can find some way to disable the valves, just shutting off the fuel flow is not going to make any improvement. If anything it is just going to cost you more, to have it run worse. If you are going to do anything, you need to have something like a VTEC, where you can switch to a second cam profile, which doesn't open the intake or exhaust valves. Anything you are trying to do, otherwise, is most likely going to waste your time and money. MO, FWIW

Quote from 1997 Haynes Engine Management book:
?A weak (fuel) mixture with a high level of O2 is good for the efficient oxidation of CO and HC. On the other hand, a relatively rich mixture with some CO aids the reduction of Nox. A compromise is reached by adjusting the air-fuel ratio of the catalyst-equipment engine to the stoichiometric ratio of 14:1. This means that the engine is perhaps adjusted slightly richer than desirable??. ?A catalyst needs to reach a minimum temperature of 300?C before it begins to work efficiently, and a working temperature of 400-800?C is more desirable. As the temperature rises over 800-1000, the precious metals will begin to break down. Above 1000?C, the catalyst will melt.??. ?Excess fuel due to misfires causes overheating. Leaded petrol and excessive oil residue also destroys the catalyst.?
So. more oxygen in the exhaust dosen't lead to CAT destruction.
Bit for the critics:
This my be a bit hair-brained but if anything else it is interesting to study about...I'm learning heaps about engine managment.
At best if it works it works, at worst I've leant something and given it a go in the process.

I've looked at all the threads I can find here about this kind of thing and all I see is the same kind of FUD you get about putting more air in the tires, and dismal failures with 4 cylinder vehicles that were marginally powered in the first place. GM through abject incompetence appears to have "proved" it impractical, much as they "proved" diesel engines aren't for cars, EVs are baaaaad, and as they seem to be "proving" that hybrids are impractical and unsellable due to offering a system in the same class as the conventionally powered Elantra that gets less mpg....

So, all the nay-sayers that say it's been proved not to work, are you gonna point out a particular example of a 6 or 8 cylinder motor that someone has gone to all lengths to get this working on, and failed??? It's a given you've got to fight or fool the engine computer, failures in that regard will not be regarded as proof.

I'd admit I might be skeptical, if I hadn't known of a guy who was getting 40mpg highway in a '89 voyager with an injector kill switch. He tried to tell people how to do it, but got much the same closed minded reaction as seen here.

Actually, I think their latest attempt with "AFM" is working fine. I hang out on a GM forum where many users have AFM and they don't seem to be having any problems with it. There is some FUD but I don't recall any reports of actual problems.

They did do a bad job with diesel cars, but AFAIK the only problem their hybrids is the price. They proved that EVs are good, not bad; just ask anyone who had the GM EV1, all the reports I've read say that people were sorely disappointed to give them up.

I was meaning more the spin they put on the whole thing. It appears to be inconvenient that the people who had them liked them.

Pumping and frictional losses together consume about 20% of rated output at peak output, i.e. high rpm, it's a square law thing so round about 3000 rpm cruise, they'll be a grand total of about 5%..... OMG IT'S LIKE A BRICK WALL!!!!!

Roadwarrior, do you have any more info on the guy getting 40mpg out of his Voyager...I'd be interested in the details of his setup...thats pretty impresive.
Good point on the pumping losses...would be good to get rid of them as much as posible by closing off valves but if it's only 5% or so then why bother with that extra complication...

Check the difference between coast down times in a FI manual trans car w/ the engine off during WOT and closed throttle. More to the point, trapping exhaust gases in the cylinder keeps things pretty hot. Just how cold would the block, pistons, mostly importantly IMO, rings get w/ this system if there was new air at whatever the ambient temp is being cycled in and out? Drop temps enough and we'll see accelerated wear IMO. Granted, manufacturers could somehow cycle exhaust gas back into the deactivated bank (maybe not if they deactivated two cylinders on an I4), but even then, w/ all that energy via hot air, I think that the closed pistons would be better "springs" than those filled w/ air at ambient temps. Just my opinion of course. :thumbup:

omgwtfbyobby: I seem to remember an engine safety feature on my mom's old GMC Yukon. It had the 4.3L V8, and if it overheated, it'd alternate firing cylinders. Maybe this is a better option here? Of course, with a v6, you'd have to figure out a better way of 'alternating' cylinders, maybe cycle is a better word.

It twinning the O2 sensors together doesn't work you could at least trigger open mode my removing or setting the coolant temp sensor low (depends on vehicle). This should be better than limp home mode.

As for the rest of it, definately I3 should be ok because most V engines share opposite banks on the same spot of the crank (I say most because I don't think the VR6 engines do), but definately stay away from I4. Lateral force (that is unbalanced perpendicular to rotation) would be very bad and potentially bend the crank.

For the rest of the idea, my friend thought once of somehow installing a valve in the spark plug hole so you could decide when to remove compression. I think the best way would be to add another valve in the head that wouldn't interfere with the piston that you could open either with cable or electromagnets or whatever. I think best would be to hook this to a central plenum between all 'dead' cylinders and then as one went up another would come down and all pressure forces should cancel and wouldn't pull fresh air or pull on exhaust flow but would eventually cool down and may cause the other issues mentioned by others.

The other thought would be to somehow drop out the compression and power strokes and just always intake-exhaust-intake-exhaust. This provides no compression, but causes issues with fresh air in the exhaust and it would be pretty hard to do, you'd need two cam profiles.

I am really curious if it's even worth it though. It just doesn't make sense to me that compression without ignition would cause much of a loss. As said before it's basically a spring, any time spent compressing is returned on the power stroke. I feel like the biggest loss (at least the biggest drag when turning by hand) is the valve train. Next would be friction and last would be compression.

If I fully seal my valves and install the head I can still spin the crank on my CRX very easily but it's a real pain to spin the camshaft.

I've always wanted to explore a rotary or electric valve system but that's a whole nother discussion.