Cam shafts as a means of controlling valve motion are simple, cheap and pretty reliable (although in the 70's there were quite a few failures due to the use of finger followers). Most cars now use either direct acting bucket tappets, or roller followers. Roller follower valve train systems also give low friction, but come at higher cost and mean a higher engine.
An electrically actuated valve train system would require something like 16 different solenoids (assuming there are 16 valves) which would all have to work perfectly for in excess of 10 years, and even assuming all this worked, you would probably need a 42 volt electrical system on the car, and I believe you would find that the power losses needed to drive the system would exceed those for the simple cam shaft system we have now.
SOME hybrid engines operate at constant speeds. I believe most do not. There's many different types of hybrid control strategies out there, most of which are subsets of "parallel" or "series" strategies. Some operate at the most efficient RPM, some at the most efficient torque, some vary over a range based on predicted driving and how it is coupled with the motor, etc.
SOME hybrid engines operate at constant speeds. I believe most do not. There's many different types of hybrid control strategies out there, most of which are subsets of "parallel" or "series" strategies. Some operate at the most efficient RPM, some at the most efficient torque, some vary over a range based on predicted driving and how it is coupled with the motor, etc.
Just to throw this out there, F1 cars use pneumatic cylinders for closing the valves only, they still are required to use cams for valve opening. So the air cylinder replaces the valve springs for them but not the cams.
The reason I have heard for not having electric solenoid valves to this point is that electric solenoids are hard to precisely control when the valve nears the valve seat and closes the valve too hard putting too much stress on the valve for long term operation. Whereas the gentle ramp of a camshaft gently closes the valve. Maybe someone else has some insight into this.
Once they figure out how to implement electronically actuated valves, the throttle plate will go the way of the dodo, why do you need a throttle is each cylinder valve is effectively a throttle?
----- Original Message ----- From: "BretCahill" Newsgroups: sci.engr.mech Sent: Tuesday, August 10, 2004 2:24 AM Subject: How Did Cam Shafts Survive Computer Controls?
Let's throw a few numbers at this....
a camshaft will control a valves movement to less than 0.001" over a 0.5" travel, so lets say for our computer to read valve position we go for a 10 bit position " byte", and we check it once every degree, that's 3,600 bits a revolution, Lets say the engine is turning over at 7,000 rpm, the data will be coming in at 420 kHz from each valve....
with a modern 4 valve layout on a V8 using 32 valves, the data stream is incoming at 13 MHz, this has to be processed, in real time and outputs produced...
Cam profiles are not simple, their generation includes consideration of position, velocity,acceleration and jerk.It may be that in view of the speed of response required a look up table will be used, with a secondary computation system for adjusting the table.
the outputs have to drive power electronics, even a light valve and bucket follower require a spring with a compressed load of 500 N to keep it following the cam nose... the actuating solenoid will probably add to the moving mass significantly and will thus need to be able to produce even greater force levels.
Given the likely size of the actuators, could they be packed round the cylinders or even fitted under the hood ?
How much will the package of sensors, computer(s), drivers and actuators cost ? a lot more than a cam mechanism I would expect.
A single failure of any part of the system could lead to a valve being in the wrong place at the wrong time........ I would not like to drive a car where a single wire coming adrift could lunch the engine.
My bet is that camshafts will still be in use at the end of this century.
Jonathan
Barnes's theorem; for every foolproof device there is a fool greater than the proof.
----- Original Message ----- From: "BretCahill" Newsgroups: sci.engr.mech Sent: Tuesday, August 10, 2004 2:24 AM Subject: How Did Cam Shafts Survive Computer Controls?
Let's throw a few numbers at this....
a camshaft will control a valves movement to less than 0.001" over a 0.5" travel, so lets say for our computer to read valve position we go for a 10 bit position " byte", and we check it once every degree, that's 3,600 bits a revolution, Lets say the engine is turning over at 7,000 rpm, the data will be coming in at 420 kHz from each valve....
with a modern 4 valve layout on a V8 using 32 valves, the data stream is incoming at 13 MHz, this has to be processed, in real time and outputs produced...
Cam profiles are not simple, their generation includes consideration of position, velocity,acceleration and jerk.It may be that in view of the speed of response required a look up table will be used, with a secondary computation system for adjusting the table.
the outputs have to drive power electronics, even a light valve and bucket follower require a spring with a compressed load of 500 N to keep it following the cam nose... the actuating solenoid will probably add to the moving mass significantly and will thus need to be able to produce even greater force levels.
Given the likely size of the actuators, could they be packed round the cylinders or even fitted under the hood ?
How much will the package of sensors, computer(s), drivers and actuators cost ? a lot more than a cam mechanism I would expect.
A single failure of any part of the system could lead to a valve being in the wrong place at the wrong time........ I would not like to drive a car where a single wire coming adrift could lunch the engine.
My bet is that camshafts will still be in use at the end of this century.
Jonathan
Barnes's theorem; for every foolproof device there is a fool greater than the proof.
Processing the data would not be as hard as acquiring it. The .5" of travel would be a severe problem. If the motion were a simple sinusoid then calculate the speed and accelerations. The peak speed would be:
2*pi*(7000 rpm/60) * .25 = 183 inches / second.
The acceleration would be :
(2*pi*(7000 rpm/60))^2 * .25 = 134336 inches / sec^2 which is 348 Gs. Wow.
A sinusoidal motion would be a best case. I bet the actual motion would have some dwells and faster accelerations to make up for the dwells. I doubt the motion needs to be .5 inches.
How many of these valves are doing exactly the same thing? All of them. They are just doing it out of phase. The PID and feed forward parts can be done in much less than a micro second if each valve has its own PID with feedforward controller. One can put many controllers into one gate array.
The math to generate cam profiles is a no brainer. A table would be required for the position. The velocities, accelerations and jerk would need to use a table but be modified by the RPM. The velocities, accelerations and even the jerk woud be crucial for feedforward calculations.
Much more I think. Think of what motion controllers cost now. A dedicated motion controller could be made more economical in volume.
Stuff happens even with mechanical valve lifters.
Me too. The electronics to generate the motion profiles and the controls is not a problem. The motion controller is only shuffling electrons and it can do it at the speed of light. The real problem is the force required to move the mass. That sounds like a mechanical problem to me. There are a lot of applications where electronic gearing and camming make sense. This discussion has taken a tunnel vision view that cams are only used on engines. There are applications, like lumber processing, where the camming tables or formulas are changed every 4 seconds because each piece of lumber is different. I would like to see a mechanical cam adapt that like that. Ok, I know that is an extreme example.
Pleease re calculate based on a sinusoydal opening for one cycle, say 200 crank degrees, makes it even worse :-(
Valves are restricting the flow and need to be opened as rapidly as possible, they are generaly accelerated rapidly to at least 1/4 diameter opening, after which the have to be slowed , returned and lowered " gently " onto their seat, this means a total movement of about 1/2 diameter, my 0.5" is probably conservative for a large V 8 even if it has 4 valves per cylinder.
No, but two sets, inlet and outlet would be O.K. if you don't care about using the possibility of running some cylinders at maximum economy and the rest just keeping warm... One of the tricks an electronic system might use to get payback...
Even knowing what you want the valves to do is a real problem... chalk up a few thousands of hours on test rigs.... then try to work the test data backwards to a formula that can be calculated.
Try finding any low ( ?? 48 V dc ) acctuators capable of producing 500 + N and operating over 100 times a second.... will water cooling be needed for the coils... Huge heavy and very expensive, It might help to have a much higher supply voltage for the coils, but that creates problems for the drive electronics...
The poster looked at the maximum acceleration, would he care to work out the peak power required by an actuator....
Yes... but at least a glitch in a computer... driver.... actuator... You have a lot of extra components to go wrong, and connection faults are fairly common with car electrics. you would need to build the electric parts to a very high standard... read hugely exvpensive.... even formula 1 cars engineered almost regardless of expense seem to have lots of problems with electrical gremlins.
Pleease re calculate based on a sinusoydal opening for one cycle, say 200 crank degrees, makes it even worse :-(
Valves are restricting the flow and need to be opened as rapidly as possible, they are generaly accelerated rapidly to at least 1/4 diameter opening, after which the have to be slowed , returned and lowered " gently " onto their seat, this means a total movement of about 1/2 diameter, my 0.5" is probably conservative for a large V 8 even if it has 4 valves per cylinder.
No, but two sets, inlet and outlet would be O.K. if you don't care about using the possibility of running some cylinders at maximum economy and the rest just keeping warm... One of the tricks an electronic system might use to get payback...
Even knowing what you want the valves to do is a real problem... chalk up a few thousands of hours on test rigs.... then try to work the test data backwards to a formula that can be calculated.
Try finding any low ( ?? 48 V dc ) acctuators capable of producing 500 + N and operating over 100 times a second.... will water cooling be needed for the coils... Huge heavy and very expensive, It might help to have a much higher supply voltage for the coils, but that creates problems for the drive electronics...
The poster looked at the maximum acceleration, would he care to work out the peak power required by an actuator....
Yes... but at least a glitch in a computer... driver.... actuator... You have a lot of extra components to go wrong, and connection faults are fairly common with car electrics. you would need to build the electric parts to a very high standard... read hugely exvpensive.... even formula 1 cars engineered almost regardless of expense seem to have lots of problems with electrical gremlins.
Opening the valves too quickly reduces the flow velocity resulting from the pressure gradient through the opening. depending on engine speed, inlet tuning, etc, more nett flow can result if the initial opening is gradual.
If a "positive" pressure can be established ahead of the valve cracking open, then opening the valve quickly is advantageous as long as the volume of air "waiting" to enter the cylinder is sufficient to satisfy and not excessive for the cylinder's need; else you'll get a back-pulse as the inertia of the column of air entering slows against a reducing pressure gradient.
Precise control of the valve position to optimise the pressure gradient during cylinder fill is one of the motivations for direct digital control of valve position.
The feedback to make it effective under variable ambient conditions is demanding. Manifold and valve port pressures need to be monitored closely before, during and after valve opening not only to maximise the flow during one valve opening, but to prepare a pressure pulse for the next fill. Such pressures have to sampled many times during the valve opening so as to "catch" the presence of pressure pulses that may be utilised for passive supercharge.
All of that has to work within the physical limitations of the valvetrain and actuators.
The solenoid itself could be constructed to produce the proper valve position vs time curve. This might be done with a circuit and/or several coils in each solenoid. Then you only need to keep track of crankshaft position, at least for a less performance oriented system.
That's the only problem but it may be tricky or impossible to build such a solenoid that works well over all rpms. But if it could be built, it could be mass produced very cheaply, maybe cheaper than a cam shaft.
The other concerns such as force necessary to accelerate valves (are solenoids orders of magnitude more inefficient than cams?) and mangled valves from electronic failure (like other fail safe systems set the default of every valve closed) shouldn't be problems.
Conditions are highly variable. Attempting to use a solenoid as a placebo camshaft cannot be justified on a cost-benefit basis.
All engines are "performance oriented". Increases in volumetric efficiency don't just benefit nett power; they benefit fuel consumption and emissions. And there's more to it than volumetric efficiency as well. Flow velocity is an important factor in delivering and mixing air and fuel, both in the inlet tract and in-cylinder.
If there is no substantial benefit from replacing a "single" moving part with dozens of others, then there is no justification to do it in terms of Engineering.
I doubt *very* much that it could be cheaper than a camshaft. Even the whole valvetrain would cost less than solenoids that do a half-arsed job.
"Failsafe" sort of involves springs to close the valves; which requires greater forces to open the valves; especially when the valves have to be closed very rapidly ahead of the piston at high engine speeds. To be able to exert such a great force, you must add a lot of iron to the valve stem, which adds inertia and increases the forces required to close and open the valve rapidly. Catch-22.
Using compressed air to close the valves is an option that's been tried. It requires solenoid air valves to relieve the closing pressure so that only a small actuator force is required to open the valve. Such a system adds to the complexity (read co$t) for which it provides a small degree of assurance that there won't be a great gnashing of metals when an actuator fails. The main benefit is that actuator loads can be relatively low most of the time.
What ever happened to the KISS* principle? The less complication, the more reliable the machine. A Cam is geared to the drive. It fails only if the gear train (or belt, or chain) fails, or if the item (Valve, whatever) being activated by the cam fails. A solinoid activated pseudo-cam has a number of ways to fail. Murphy is NOT a lurker! Roger
The performance requirements of engines have become much more demanding; increased efficiency, reduced emissions, less space, less mass, more power, greater flexibility.
What about variable lift and timing mechanisms? They add dozens of components. And they have only limited scope for control of valve motion in response to variation of operating conditions.
Would it be possible to scale down those two stroke diesels they use in ships? They must be efficient because fuel is about 60% of vessel shipping costs and owners are cheap cheap cheap.
Since a hybrid has a lot of battery power, you might be able to melt the 6 oil with an electric heater.
Depends on what you mean by performance. It's usually not in car manufacturers' best interested to maximize something like engine power. If they did, it would generally cost the company more in the long run. Doing something like maximizing power output induces higher stresses on engine, transmission, and other components. This results in more maintainence costs for cars under warentee. Not only that, but their car looks bad in the public's eyes cause it's in the shop more.
Bingo.
Has anyone here even quoted what sort of efficiency increase you could expect by going to solenoids?
back to the simplicity issue. Cams do a (one(1)) thing very well. Any compromise in a machine, or attempts to exceed its limits leaves some performance criteria at a less-than-optimum level
Performance in terms of "specific" power output over the entire operating range. :-)
Maximum power output is largely irrelevant in day to day motoring. What is important is maximum energy from each milligram of fuel. i.e. the most-efficient burn environment for the fuel; which also tends to increase the potential for power and the available torque under all load conditions.
One doesn't have to push safety factors in order to _reduce_losses_ within the engine. Reducing losses is how you gain power and improve efficiency.
Something like 15 to 20 percent increases in fuel efficiency can be realized by electro-mechanical variable valve lift and timing mechanisms, compared to "static" valve timing.
Solenoids increase the ability to control the opening profile even more finely... What's still left to gain in comparison to "conventional" electro-mechanical variable valve actuation is probably close to "not much at all".
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.