Slashdot Mirror
Inventor Demonstrates Infinitely Variable Transmission
ElectricSteve writes with this excerpt from Gizmag:
"Ready for a bit of a mental mechanical challenge? Try your hand at understanding how the D-Drive works. Steve Durnin's ingenious new gearbox design is infinitely variable — that is, with your motor running at a constant speed, the D-Drive transmission can smoothly transition from top gear all the way through neutral and into reverse. It doesn't need a clutch, it doesn't use any friction drive components, and the power is always transmitted through strong, reliable gear teeth. In fact, it's a potential revolution in transmission technology."
The real icing on the cake is (as mentioned near the end) the secondary drive doesn't require a whole lot of power so it can be run by a flywheel. Infinite torque? Frictionless? This is almost too good to be true, there has to be some catch. Like the primary input drive requires more energy than they expected but I can't see it--although I'm not a mechanical engineer.
... these are the kind of news stories an engineer loves to read about.
This is the kind of thing you like to see -- I hope this man has all the capital he needs and gets that prototype up and running for demonstrations. Plus it's a small time plumber inventor
My work here is dung.
I think the weakness in this design is the need to rotate the "bottom" shaft at a speed equal to the input shaft for neutral. While indeed it doesn't need a lot of power, it's a lot of rotation where, in competing designs, a clutch disengages or the drive motor is idling. I could see a lot of things going wrong if the synchronization was imperfect, or if something went wrong.
How do you start this up from a dead stop? Somehow you have to exactly match the shaft rotation speeds to keep it in neutral before you start moving forward, otherwise there will be a lurch.
I look forward to seeing how this is developed further. It has a lot of potential.
Well, one potential flaw is the eccentrically mounted components. Unless properly counterweighted, at high speed this will cause a lot of vibration.
BTW, couldn't you do this sort of thing with a differential?
When our name is on the back of your car, we're behind you all the way!
It works as a demo very well I , as an ME agree.
The big issue in science and engineering is ALWAYS reduction to practice. The inventor acknowledges this and is working with an engineering firm to make a practical pseudo-production testing model. When you have no clutches, the lack of shock loading means the size of gears and the housing can be substantially reduced, since there won't be an engine load shock issue. There can be issues of loads when parked, though, when another car bumps yours. The other issue is how do you tow such a car when the engine fails or you want to tow it behind a motor home? There may still need to be a "cog" connection for towing.
Issues involved in getting it into a small, produceable and cost effective prototype will tax the engineers. If they can do it, there will be applications in many different fields.
Given that the gear ration can be set by controlling the small electric motor speed, it can be integrated with other electronic control systems easily.
I have to hand it to the guy for coming up with a very clever implementation. This is why we need to support the math, science and physics departments everywhere, because in the end, the world is a physical place and the countries who prosper the most will be the ones with the most technologically up-to-date innovators.
At first blush, I'd say that both Toyota and John Deere have already produced something similar. What he appears to have, however, is a system that can smoothly transition (with power) through neutral and reverse. That indeed could be the cool, patented part, as the rest of his transmission is pretty well understood and actually in production already in many of the applications they list for their invention. I don't see any patent application listed, so I can't tell for sure exactly where his breakthrough is.
Here's the fundamental principle by which his transmission works, though: Basically the idea is you supply driving power to a planetary gear system and then use another variable system such as an electronic motor or, in John Deere's case, a hydraulic motor, to take speed (but not power) away from the output shaft by spinning part of the planetary system. If you understand how a planetary gearbox works, this makes sense. So in John Deere's case, the less-efficient hydraulic motor uses a tiny amount of power to control how the actual, geared, power is transmitted to the wheels. Using this system JD has a completely variable system with a particular gear range (this is a tractor after all) that has a powered neutral stop. In the pictures and video you'll note he has two electric motors that control the ratio.
Toyota does something similar with their hybrids, although it's more of a way to efficiently (and brilliantly, I might add) blend the gasoline motor's power with the electric system in an infinitely variable way.
Another way of implementing an IVT, though I don't think it is as efficient, is to use a differential. Power comes in the normal part of the differential (IE spinning the entire gear assembly), and then power comes out one side, and an electric or hydraulic motor attached to the other side (Where the wheels would normally go). You can then use the motor to change the apparent gear ratio, and even reverse it.
It is indeed similar to the planetary gear coupling boxes in parallel hybrids. And yeah, you are basically right - a 6-gear box holds you sufficiently close to the optimum rpm anyway for practical use. CVTs really shine in heavy machinery, but are not that important for personal cars. Still nice technology, though. To hell with practical importance - all hail those engineering efforts done for the heck of it!
Ubi solitudinem faciunt, pacem appellant.
The Thompson coupling was invented not long ago, and I remember being amazed that there was anything new to be done in the area of mechanical power transmission. And now this. Are we all done now, or is there more still?
Fuck the system? Nah, you might catch something.
Hate to reply to my own post, but here is a fairly detailed explanation of John Deere's IVT: http://salesmanual.deere.com/sales/salesmanual/en_NA/tractors/2006/feature/transmissions/8030_option_code_1127_1137_ivt_trans.html . The relevant part is "The John Deere IVT uses a hydromechanical, power-splitting design where a portion of the power is transmitted mechanically and a portion hydrostatically. A hydromechanical transmission is more efficient than a purely hydrostatic transmission because gears carry power more efficiently than a hydraulic pump and motor. By careful selection of the gearing, the John Deere IVT carries a maximum of the power mechanically both at normal field working speeds and at transport speeds, taking maximum advantage of the higher mechanical efficiency while providing the control and versatility of a hydrostatic." And of course this power-splitting is done via a planetary gear system.
I say this not to take away from the D-Drive's awesomeness (John Deere doesn't do reverse without shifting a gear), but to help offer explanations of how it actually works.
Fluid friction losses. Recirculating a fluid via a pump in a closed system actually makes a bit of heat, especially when there's a bit of load on it. Works great when something can be built big and doesn't need to go very fast (like the tractor application you mentioned, also used a lot in earth moving equipment and fork-lifts), but when having something that goes fast - not so much. Also if you go too fast, you're either going to have some kind of undesirable hammering or cavitation at a certain point depending on what kind of pump you use to provide hydraulic power.
Some air motors use a tilt-block that does something similar as well in regards to infinite variable speeds, but they're not so much about efficiency as about being able to control speed in industrial environments where electric motors aren't always desired. (Like working around water or in a no-spark environment.)
I watched with interest through 3/4 of the video as they continuously refused to show the back side of the model, just loosely discussing the "control shafts" and couldn't get it out of my mind
"pay no attention to the man behind the curtain".
Then finally at the end they showed the back and surprise, there's another motor there, but trying to explain it off that this motor requires far less energy than you're going to gain by using the rest of the system. Maybe this is true, but that's a poor way to present the design, by hiding a serious concern until the last second.
As they wrapped up the video they did admit that this little kink is going to be the determining factor in whether or not it's a useful design. "Why can't they just tap some of the power off the input shaft to manage the control rods?" I thought. Then it occurred to me, the speed would need to be continuously variable, and that's the whole problem they're trying to solve. So, what we have here is a continuously variable mechanism, so long as we can already provide a continuously variable mechanism. (all his D-Drive needs to complete it is, another D-Drive, which would of course need another D-Drive....) Sounds terribly recursive to me. But he didn't go into any detail as to the requirements of this control system, but from what I can tell, it needs to be continuously variable also. He dismissed it as being easy to achieve with something such as an electric motor, which one could argue the same is true of his entire invention...
We'll see. I'll remain skeptical until his design is complete, including the nagging little details of running the control shafts. But really it's an excellent idea even with this problem. It's solved the larger portion of the problem. One other thing that also came to mind is balance. The orbital gears could really get whipping around the sun gear, they'll have to be balanced. Using orbital gears itself at high torque will create new problems also. I'm no mechanical engineer but I also see a potential problem there with torque on the position of the planetary gears since the shaft isn't fixed. You don't usually see floating gears in transmissions.
I work for the Department of Redundancy Department.
The new aspect is that this planetary gearset actually has TWO inputs, and the output is determined by the *difference* in speeds between the two. That's how it can go from reverse to forward seamlessly. V1 > V2 is Forward, V1 V2 is Reverse, V1 = V2 is Neutral. Assuming there are no practical limits on the velocity of either input, the possible difference between them is infinite.
Personally I find this really exciting, because i've always been in love with the idea of a variable transmission. Ignoring electric motors for a minute, there are some absolutely INSANE things you can do to a small motor with cams, turbocharging, etc, to extract absolutely massive amounts of power from teeny engines. Like, 1000+hp from sub 2 litre motors. The problem is they end up being extremely peaky (power is only made at a narrow RPM band, or a terribly high one)... but with a variable transmission you can let the engine hunker down in it's sweet spot and let the tranny worry about all the fiddly bits. Hell, you can even do the same thing with a big engine... I wonder if its possible to make five figures of power from a 7 litre? With this we just might find out.
BTW, couldn't you do this sort of thing with a differential?
Yes. That's basically what the Prius does: it uses a differential (actually an epicyclic, which is a flattened differential) as a mixer, and drives one input with a gas engine and the other with an electric motor, giving not only an infinite number of speeds but also a way to use the engine to charge the motor with excess power, or use the motor for braking. But then you need both an engine and a motor. Managing an infinite drive from a single input is pretty cool.
Nostalgia's not what it used to be.
Actually, the transmission in the Prius is completely different from this. The Prius takes two full power inputs (the engine and the electric motor), and adjusts the power output from the two (balancing them) to achieve the end ratio. This takes a single full power input (and two factional inputs, perhaps a very small fraction if friction losses are small enough), and produces a variable end ratio. Quite a big difference between them. For the Prius transmission to work, both engines need to be of comparable power (A 100 hp gas engine would need somewhere near a 100hp electric motor). This would likely work with a 100hp engine and a pair of 1/2 hp (or less, depending on precision and friction) electric motors.
And FYI, an OTTO cycle engine is not most efficient at 2000 rpm. It's most efficient at its horse power peak RPM, and at full throttle. Anything less than that (RPM or throttle), and you lose volumetric efficiency. And when I say efficient, I'm saying the power/fuel use is the maximum. It's all about the intake and exhaust design (you can tune them for maximum efficiency at a particular RPM for a particular engine design). That's why hybrids typically use smaller engines. So that you can run it closer to its peak power for longer (40hp at full throttle would be plenty to cruise on the highway and still be able to charge the batteries without needing to be throttled back).
If a man isn't willing to take some risk for his opinions, either his opinions are no good or he's no good
Not to worry about 'infinite' or 'frictionless' - these characterizations are not the intent of the device so we can just evaluate it as a normal continuous transmission being controlled by the ratio of speeds between a control shaft and the drive shaft. Efficiency for low-torque cases can be quite decent as eccentric bearings, gear-qualities and diameters can be controlled well with current techniques.
So... with a real torque, there will definitely be significant forces between the two shafts. Clearly, the full torque will be on the central, driving shaft, while some smaller fraction will be on the upper shaft. As we bring the distance between them down, then the torque between them can decrease, but then there will be more stress on the smaller pinions' teeth. Planetary gears are great for this class of problem and he's throwing decent diameter eccentric bearings in where he can too. The bloke seems honest, and has clearly thrown a fair amount of time and energy into the problem.
There are other approaches to controlling gear ratio via the speed differences between two shafts - he's not trying to do something impossible, he's just trying to do something difficult, successfully. Whether the cost of the bearings and gearing will be favourable when compared to the other approaches is the question. I think his system will work - and decent sealed bearings, high strength pinions, planetary systems - these already exist and are stable tech in current transmissions, even in relatively dirty industrial environments where the transmissions aren't as protected as in cars. In particular, the cost of electronic control for motors has fallen massively over the last years, so if nothing else, the general class of solutions using differential speeds of low-torque motors to control a high-torque transmission is more appealing now.
So, 'genius', no. Hard-working, self-taught engineer? Yes.
Assuming 1000 miles per month (which is what most leased cars are allocated) :
The difference between a 150mpg car and a 250mpg car is 32 gallons of gas per year per car.
The difference between a 30mpg car and a 40mpg car is 100 gallons of gas per year per car.
The difference between a 20mpg car and a 30mpg car is 200 gallons of gas per year per car.
The difference between a 15mpg car and a 25mpg car is 320 gallons of gas per year per car.
The difference between a 12mpg car and a 22mpg car is 450 gallons of gas per year per car.
The difference between a 10mpg car and a 20mpg car is 600 gallons of gas per year per car.
Read from bottom up, you see the point of diminishing returns.
If car companies would focus on the right range (forget about exotic expensive 150+ mpg carbon fiber hybrids that hold two people, focus on 30+mpg vehicles that hold a family and gear) they would have a LOT more impact. I don't necessarily agree with the way cash for clunkers was handled, but in the cases where people traded in a 12 mpg car and drove off in a 22mpg car - it makes a BIG difference.
Glonoinha the MebiByte Slayer
The way I understood it (could be wrong), the Prius drive is only one half of what this guy came up with. The clever bit is the other half. The Prius transmission would not work well without significant torque input/output(electric breaking) on the electric side. The way this works, there is almost no load on the ratio selection element, the only input it needs is enough to create a difference in speed.
Mind the frickin' laser...
I tried it for a few minutes, and the Prius never suddenly accelerated. Clearly the simulation is flawed.
It's loud. Plus you have to have a heavy hydraulic system (pump, oil reservoir, valving, etc). In practice an electric motor gives you most of the same benefits of a hydrostatic system, but it's a lot lighter and doesn't require an oil system. Of course batteries are heavy, but aren't strictly needed (as in a locomotive).
A few years ago I heard of a design that used small hydraulic motors connected to each car wheel via clutches that would efficiently overcome the propensity of a differential to send power where you don't need it. Basically when slippage was detected, the clutches would engage and the faster wheel would act as a pump, sending fluid to drive the other wheel. The beauty was that if you tied all four wheels into the same system, you could get on-demand four wheel drive as well.
Another prototype I heard of used a hydrostatic system to charge up a nitrogen accumulator in a form of regenerative braking system.
Hydrostatics also has limitations in the amount of power you can transmit. Every large combine harvester we've ever owned has had a hydrostatic transmission, but no tractor ever has had. Combines typically don't pull things; driving power is minimal compared to the power consumed by threshing. Whereas in a tractor, it's all about driving power (outside of PTO applications). You just can't really put 500 HP of pulling power through hydrostatics. Most hydraulic motors are gear motors, which means the oil spins little gears. Under high load, oil slips past the gears without turning them. Compare that to electric where on a daily basis Locomotives pass thousands of horsepower from big diesel engines to the wheels with electric motors.
No, the most efficient point is at peak torque. That's where the engine is able to produce the most energy for a given amount of gas. The horsepower peak is where the engine is producing the most power (energy/time). It is not necessarily it's most efficient point unless they coincide which is rare.
The largest value of this device is in its "wow, how does that thing work?" design. By baffling the onlooker and also describing the widget very carefully the illusion of a wonderfully useful device can be created.
It has a problem in the real world, though. The reaction torque is equal to the working torque - and the reaction torque path runs through that "secondary control shaft." This will become obvious as soon as he tries to transmit some significant power through his device. What he's showing isn't a new invention at all, it's just a mechanical "summer" that adds the inputs from two input shafts. All that's new here is some fancy handwaving and creative description.
It might be good enough to fool some people but Mother Nature and those who paid attention in school aren't fooled. Maybe if / when he actually tries to transmit some power through his "invention" and the control motor just spins backwards he'll "discover" a source of electrical energy?
hunker down in its sweet spot and let the tranny worry about all the fiddly bits
I've spent too long on the Internets, apparently.
~ C.
Not true. Volumetric efficiency is measured as the the volume of air taken in on each stroke vs the displacement of the cylinder. So if 1 liter of STP (Standard Temperature and Pressure) gets drawn into the cylinder, and the dispacement of that cylinder is 1.2 liters, the total efficiency is 1/1.2 (or about 83%). At 0 RPM and 100% throttle (well, any throttle position that isn't completely closed), volumetric efficiency is always 100%. But as the engine starts turning (at full throttle, otherwise the vacuum drawn by the throttle restriction will reduce efficiency), the actual efficiency will depend on intake design. Considering that OTTO cycle engines use valves, air is only drawn in 25% (about) of the time. So the vacuum drawn trying to draw that air in will cause the efficiency to drop. However, intake runners are designed for this. So basically, when the valve closes, the momentum of the air causes a pressure build up behind the valve. That pressure will cause the air to reverse direction. This leads to a harmonic wave in the intake runner. The frequency of the wave is dependent on the design of the intake runner (cross-sectional area, cylinder volume and length mainly, but curves and other obstructions do play a part). If the valve opening is timed properly with this frequency, the incoming pressure wave from the harmonic will actually force air into the cylinder. That's how some racing engines can actually achieve a higher than unity volumetric efficiency at a specific RPM. It's all relative to the design of the engine. Some engines may be designed for 2000 rpm. And increasing the RPM over that WILL decrease VE. But you cannot say as a general rule that VE is inversely proportional to RPM, because it isn't. And pumping losses are directly proportional to VE (in fact, the pumping losses are DUE to VE below 100%).
You do have a point that thermal and mechanical losses do increase with RPM (Mechanical due to friction, thermal due to the increased movement of air around the parts). However, your reasoning behind diesels being more previlent is flawed. It's not because they operate at a lower RPM. It's because of a few reasons. First off, diesel is denser (energy/volume) than Gasoline while still having a similar stoichiometric ratio with air. Secondly, diesels are typically built without a throttle blade. That means that even at idle or lower power settings, there is no restrictive plate to draw a vacuum (and hence harm VE). Since diesel doesn't behave as bad as gasoline when run lean, they typically control power output by controlling the fuel flow. Third, diesel engines tend to burn much hotter than gas engines (the flame front is significantly hotter), so there is a more complete burn. You combine these effects, and you can see why they are more efficient (and it's not because they run slower). The reason that diesel engines typically run "slower" is two fold. First, since diesel engines don't use spark plugs, timing is controlled by the mechanical fuel injectors (direct injection). They were simply not fast and accurate enough to time at high rpm. The second reason, is that diesel is slower burning than gasoline. So at higher RPMs, there's a large chance that combustion won't be complete when the exhaust stroke starts (resulting is a large drop in efficiency and a large increase in mechanical stress).
If a man isn't willing to take some risk for his opinions, either his opinions are no good or he's no good
It's quite different from the HSD in that it has three inputs, contrary to what GP said - one power input, and two control inputs, both of which ought to require just a fraction of the input power to control the input/output gear ratio.
Real engineers disagree on the "inventor's" website:
http://infinitelyvariabletransmission.com.au/wp-content/uploads/2010/05/dDrive-Transmission-Report.pdf
"The torque provided by the Control shaft will typically be of the same magnitude as the torque provided by the Input shaft."
"The Control shaft (and associated mechanical elements) should be sized to this torque requirement
accordingly – the Input and Control should be considered as parallel power paths rather than as ‘power’
and a ‘control’ elements respectively."
So this whole thing isn't very useful. To add this as a transmission to a power motor, you need
one ore two additional motors of same power with variable speed and enough torque at any speed.
Yeah but eliminating clutches removes a common point of failure. So even if the performance benefits aren't that great it may increase the life of the car/machine.
it may increase the life of the car/machine.
Sadly, this may kill the project.
It seems that manufacturers don't want to build things that last forever. Planned obsolescence is the current fashion.