Evolution of the CVD Drive
I have been curious about the innovation process and how and why it happens. In digging through 12 years of files to write two papers on the Continuously Variable Displacement (CVD) drive, I was struck by how the process of creating the CVD drive system happened. Some of the initial 3D models were somewhat embarrassing in hindsight. Digging through the files, it made more sense and the realization a strategy had been followed that had never been written down or verbalized.
There was the initial concept, which was the foundation of the first patent. But that was just the tip of the iceberg of the invention that would be required to complete the design, make it manufacturable, and function correctly during testing. The invention required to create the initial concept was relatively easy, whereas the invention required to make every part function correctly, be manufacturable, and able to be assembled was far more difficult and of much greater scope—a difficult, grinding, and time-consuming process. Many of the issues that arose required looking at every possible permutation, inversion, and alternative over and over without finding a solution that worked.
Then having to park the problem and let it “ferment” a while, to then take it up again weeks, months, or sometimes, years later. And as the number of parked problems accumulate, it’s a challenge to retain confidence they will all be solvable eventually and continue to put energy and effort into the project.
It started with a 2D sketch. All the key elements are included except anti-rotation. None of these parts have survived the development. How do you model it to have the desired characteristics—variable stroke with constant or controlled compression ratio? A model of some sort was needed to do this. It started with basic high school geometry, trig and algebra. Despite the rust, the model of the geometry and motion was laid out.
Once the hand calculations were done, they were committed to a spreadsheet. A stick figure representation was created using X-Y plots. All of the important attributes were calculated, including the actuation position, stroke, displacement, and compression ratio.
Since the compression ratio also needed the volume in the head, not just the TDC volume of the remaining cylinder, a placeholder assumption was used.
An Excel Table was created where the stroke could be stepped from minimum to maximum with the critical attributes calculated Then the geometry could be juggled until the right characteristics were achieved.
When the 2D layout met the attribute targets, a 3D model was created and combined with multi-body dynamics simulation.
The focus was on what was known, the very kernel of the mechanism. Problems without solutions were noted but left to be solved later.
Details for known technology were passed over to be dealt with later. None of the parts in this 3D model survived the development. But the critical elements to make a variable stroke with constant CR are embedded in the design.
In hindsight, this was the engineering strategy that was followed:
“Details for known technology will be dealt with later”
• For example, the geometry of a piston and its rings is very important but known technology. It can be dealt with when the decision is made to go to hardware.
• The valvetrain is known technology. Barrel-type engines have been built before. There is a lot of know-how and technology in a valvetrain—big job but not real technical risk. Wait until it is needed, funds are available, and then contract it out to one of leading engine design firms.
“Problems without solutions will be noted but solved later”
• It sounds like an extremely dangerous methodology as one could get to the end of the design and not have solutions for all the critical problems. This was a huge risk and turned out to be very real.
• Solving problems with an unknown solution take time, sometimes months or years. “Thinking Outside The box” is easy to say but exceedingly difficult to do. Keep the thinking wheels turning and implement a design when a solution pops out.
• Committing to an immediate but poor solution is a project killer. Inventing on a defined schedule often results in poor designs in expensive hardware, squandering money and time. As a consultant, I have seen this happen to other alternative engine concepts where there was a strong drive to have a piece of hardware to show investors, but the “quick and dirty” design became a huge liability with design flaws that could not be fixed. This can set the project back several years and squander millions of dollars. Patience is key along with resistance to implementing a solution that will have to be replaced later.
“Minimize the work where you are not an expert”
• For example, I am not an expert on engine design and am not familiar with the latest technology for each component. An effort was made to stick to the essence of CVD mechanism and avoid designing subsystems I am not an expert in: valvetrain, combustion, fuel system, auxiliaries, cooling system, lubrication system, etc.
• Sometimes the funding and required schedules are not compatible with this philosophy and it ratchets up the risk in an invisible way. The constraints of the DOE Phase II program were such that I had to take on a lot of design of the cylinder block, crankcase, pistons, rings, etc. where there was little expertise.
This was the next-generation design. The realization that it needed to be an odd number of cylinders for a four-stroke engine finally hit. The engine block, heads, valvetrain were unimportant at this stage, so something simple was modeled just as a framework to hold everything together. Most of the work at this stage was on the mechanism—multi-body dynamics analysis defining the motions, loads, stresses, degrees of freedom, etc.
By this stage, it needed to look somewhat like an engine. A water-cooled cylinder block, a crankcase, and oil sump were created. Many Nutator/Rod designs were done trying to minimize the mass. The balance was introduced. A cam drive at half speed was added. Two types of cams were designed—a drum cam and a platter cam, which also included variable timing features, but those were far from my expertise, so they were put aside.
An aluminum clam shell Nutator lightened the nutating mass. Splined journals were added to distribute the load to the Sliders as well as increase the stiffness for a given diameter. One-piece Sliders were created with stepped ends to distribute the torque. The space for an actuation piston was increased, enabling an actuation piston system that worked. An articulated lower balance with cam system and an upper balance weight were added. How to anti-rotate the Nutator had not been solved.
With the DOE SBIR Phase II grant, the deliverable was a tested unit, so everything had to work—no placeholders! Fits, tolerances, stack-ups, for all parts. Parts had to be manufacturable, the unit had to be able to be assembled. A gear system was devised to take the torque to ground (anti-rotation), lots of balance work, actuation hydraulics and lubrication system added. Real pistons, rods and rod bearings were designed. It’s interesting to look at the original sketch and see that none of the parts survived, but 100% of the original design intent was retained.
Since the DOE SBIR Phase II deliverable was a compressor, not an engine, heads, intake manifold, exhaust manifold, and passive valves had to be designed.
The realization of the idea into a working model took a decade and continuous invention was required to eliminate the manufacturing and assembly problems and bring the idea to a functioning design. Without the spark of the idea that started the project, or without the grinding effort of fleshing-out the design, recognizing problems and creating solutions, the CVD would not have reached its current state. However, the calendar time this took may result in the CVD missing its window of opportunity to improve the efficiency of ICEs around the world.
There was one “problem without a solution” that did not get solved in time. That was a method to feed oil to the rods/pistons. The only solution that was found was to use high-pressure oil from the actuation cylinder, and it required feeding the oil across a seal with an ultra-high Pressure-Velocity parameter. Time had run out and no other solution was found. That was the one item that dominated the test time and limited the data collection.
A solution to the rod oil feed problem has been found—three or four months after the conclusion of the testing and reporting. Fortunately, the problem has not killed the concept’s viability but it has limited the amount of data that proves the concept’s performance. The CVD can now move forward with more funding/investment
