PDK Shifting Technique
#16
Nordschleife Master
This is similar to what others have said, but it begs my question. What I'm trying to learn is the delivery of torque during the shift. Is there a discernible interruption of torque delivery in a high-rpm shift? And does that sensation change when you switch to sport mode?
Though no one has directly answered that,
Gary
Even my sequential manual in a Cup, no
The reason is obvious; in upshifting the car is accelerating, the excess force simply pulls the trans and wheels up with it. In downshifting the inertia is backwards and the drive gear must absorb the excess force and transmit it.
This isn't as difficult as it's being made. Take a car, any car with a manual trans and try this;
Shift up or down immediately after an acceleration interval, don't wait for it to "spin down" and shift. While you can make it jolt if you try hard enough, typically any upshift is reasonably smooth. Now downshift with the same pre-conditions. You will find that only when the revs are matched is the shift smooth. Any car...
#17
Race Director
Nope, several of us have answered that, The answer is NO, not on the street or the track. As we all have noted, even while apexing at high RPM, no. Sport shift is a little more forceful, but every thing is programmed to go fast and that leads to a loss of smoothness and perhaps a small amount of what you are talking about.
#18
Rennlist Member
Thread Starter
Join Date: Dec 2009
Location: Southern California
Posts: 1,429
Likes: 0
Received 15 Likes
on
15 Posts
The reason is obvious; in upshifting the car is accelerating, the excess force simply pulls the trans and wheels up with it. In downshifting the inertia is backwards and the drive gear must absorb the excess force and transmit it.
This isn't as difficult as it's being made. Take a car, any car with a manual trans and try this;
Shift up or down immediately after an acceleration interval, don't wait for it to "spin down" and shift. While you can make it jolt if you try hard enough, typically any upshift is reasonably smooth. Now downshift with the same pre-conditions. You will find that only when the revs are matched is the shift smooth. Any car...
This isn't as difficult as it's being made. Take a car, any car with a manual trans and try this;
Shift up or down immediately after an acceleration interval, don't wait for it to "spin down" and shift. While you can make it jolt if you try hard enough, typically any upshift is reasonably smooth. Now downshift with the same pre-conditions. You will find that only when the revs are matched is the shift smooth. Any car...
Very impressive transmission. Will discuss later,
Gary
#19
Nordschleife Master
#21
Rennlist Member
I find that the easiest way to describe PDK shifts is to say that, unlike changes with a manual transmission, there is no sensation of chassis-jacking. Up-shift or down-shift, the car stays neutral in pitch as torque is fed more-or-less continuously to the wheels. This makes the car feel less powerful. Obviously, it is not.
How is this done? You might want to ask Porsche that question. Or look at the wiring diagram to see what sensors do what tasks. The ECU will still be a mystery, though. I'd hazard a guess that acceleration, engine RPM, road speed, barometric pressure, and other factors are fed into a smoothing algorithm, then tested against a history of the driver's intervention by shifting. In this way the car 'learns' how one wants to drive it. Of course, that leaves holes in the program's ability to always know exactly what one wants to do. Passively down-shifting into first gear while in Sport mode is a bit abrupt sometimes because of this.
So, which one is better: PDK or manual? If you want that 'slam', 'slam' sensation felt on one's backside as the wheels hook up with each shift, the PDK will probably not work for you. OTH, if you value smooth, precise moves, PDK could be your friend. I value smooth.
Marshal
How is this done? You might want to ask Porsche that question. Or look at the wiring diagram to see what sensors do what tasks. The ECU will still be a mystery, though. I'd hazard a guess that acceleration, engine RPM, road speed, barometric pressure, and other factors are fed into a smoothing algorithm, then tested against a history of the driver's intervention by shifting. In this way the car 'learns' how one wants to drive it. Of course, that leaves holes in the program's ability to always know exactly what one wants to do. Passively down-shifting into first gear while in Sport mode is a bit abrupt sometimes because of this.
So, which one is better: PDK or manual? If you want that 'slam', 'slam' sensation felt on one's backside as the wheels hook up with each shift, the PDK will probably not work for you. OTH, if you value smooth, precise moves, PDK could be your friend. I value smooth.
Marshal
#22
Nordschleife Master
Actually, I'm short of time right now so I'll have to postpone explaining why, and I rarely tell another poster he is simply wrong, but you are. Not as a matter of opinion, but physics. I understand what intuitive feeling you're expressing, but in fact the inertia you describe works with the intended direction in both cases. If you don't feel a jolt on upshifts that don't pause for the engine to slow, then... No, out of time. Must run, but that does tell us something interesting about the PDK transmission. Basically, it is a lot more sophisticated than people give credit for, but also more complex in terms of the electronic control, just as you said earlier.
Very impressive transmission. Will discuss later,
Gary
Very impressive transmission. Will discuss later,
Gary
an accelerating vehicle is taking up energy and translating it into velocity. A slowing vehicle is giving up energy, as in the recharging systems on Hybrids or heating up brakes. So the direction of energy is opposite in the two examples.
Now, if you wish to contort that I'll be happy to watch, but just as an fun exercise, take it to redline and shift 2-3 as fast as you can and tell me what you feel, don't wait as you discuss in your opus, just do it smoothly and rapidly. Then just slam a downshift similarly, no blips or anything and tell me the engine/drivetrain are working together and inertia is the "same".
There is indeed inertia present, it's just that the force is moving in opposite directions. So in absolutes, yes there is inertia, just of opposite sign... big difference!
#23
Rennlist Member
Thread Starter
Join Date: Dec 2009
Location: Southern California
Posts: 1,429
Likes: 0
Received 15 Likes
on
15 Posts
an accelerating vehicle is taking up energy and translating it into velocity. A slowing vehicle is giving up energy, as in the recharging systems on Hybrids or heating up brakes. So the direction of energy is opposite in the two examples.[...]
There is indeed inertia present, it's just that the force is moving in opposite directions. So in absolutes, yes there is inertia, just of opposite sign... big difference!
There is indeed inertia present, it's just that the force is moving in opposite directions. So in absolutes, yes there is inertia, just of opposite sign... big difference!
First, I hope you'll forgive my brusque post this morning. A tough medical day was facing me and my planned relaxation with the forum beforehand evaporated with make-work chores before I realized it. I didn't notice the time until I hit 'reply' on your note.
When I do a 2-3 shift in that experiment you suggest, the driveline gets a serious thump, as do the driver and passenger. I'll get back to that later, but first an explanation of why some of that energy flow you picture is actually moving backward. An eddy if you like.
When we accelerate in a gear, we are building momentum in the car. The engine is turning chemical energy into kinetic energy. So far so good, but we have to remember that just like those hybrid Porsches you're picturing we have energy storage aboard, just not as complex as theirs. The new 918 has this elaborate flywheel where the passenger would go, and we have a smaller one. The whole point of a car's flywheel is to store energy in kinetic form to smooth out the impulses provided by a cylinder-style engine and with internal combustion to compensate for the lack of torque at low rpm. But designers are opportunists. If we must do something for certain reasons, we try to turn the inevitable into advantages elsewhere. Thus it is with the flywheel which now serves mutliple purposes. In particular, the flywheel helps us change gears.
If the engine is built with very low internal rotating mass, then it loses rpm very quickly when the throttle is closed. At that point, the engine is an air pump and nothing more. Without a store of kinetic energy to slow the fall, engines decelerate very rapidly. Race cars are a mild form of that limit case. They do have kinetic energy stored in the driveline, but not nearly so much as a road car and their rhythm is much faster as a result. When you shift one, the engine rapidly drops rpm, so you best not dawdle. When you start off from rest, the engine's slow rotation provides very little torque and so little kinetic energy is stored that you can't stumble off the line as road cars manage to do. If you fail to get the rpm up first and then apply strong throttle as the clutch engages, you will embarrass yourself with a stall. Alright, no one will blame you for rapid departures on a race track. But just as obviously, that tactic is awkward when Aunt Susie is leaving a light in front of you and her acceleration is rather limited for your Carrera if you give it "strong throttle." In road cars, we need a reasonably hefty flywheel to let us smoothly move away from rest without running up her back end.
Now get us up to the shift point in second gear and the engine is turning that flywheel (and the other internal rotating masses) 125 times per second. That is, 7500 rpm. That's a substantial chunk of kinetic energy for which we have special units of measure, but for whimsy, let's call it $1.25 worth of KE. When we get third gear engaged, the car is still gaining speed, so it's store of kinetic energy is increasing, just as you say. But the flywheel is going to be turning only 73 times per second. The car's net store of kinetic energy is still rising, you're right, but part of that energy has gone the opposite 'direction' if you like. The flywheel lost energy. 52 turns per second had to be gotten rid of somehow. We had to move $0.52 worth of KE out of that flywheel.
When the designers do their job, they choose a flywheel whose rate of giving up energy to support the engine's rotation (keeping the engine from almost instantly stopping from the pumping load) will occur at the same rate as a typical driver of that model car will achieve the shift. Your judgment will be "smooth car" or if you're a sports car type, probably more specific: "nice smooth shifting transmission."
My passenger (and my driveline) still would get that thump however. At least if I used the shifting method you suggest. You see my "fast and smooth" is going to be a lot faster than yours because I spent so many years driving race cars with zip for flywheel energy storage. Without consciously matching my shift speed to the engine's rhythm, without giving its pumping load time to absorb that $0.52 of kinetic energy from the flywheel, it's going to be dumped into the driveline in a great lump when I let out the clutch at the 'smooth' rate that I consider normal. Neither one of us is wrong, we just have to adjust our driving style to the car if we want that smoothness that passengers hope for.
Now let's get back to the subtlety they've built into the PDK. All of you here uniformly describe the full-throttle shifts as being quick and smooth. On inquiry, you indeed mean by those words that no interruption in acceleration takes place. The delivery of torque is without a notch, without the gap that a smooth high-rpm shift requires with a manual. Where is that $0.52 of KE in a 2-3 shift going for all love?
Well right now, subject to reading the technical docs I've been given, it seems very likely that the following is happening. As the shift begins, power is reduced to minimize wear on the clutch for the current torque path through the transmission. That 'old' clutch must release that path before we bring the other path for torque into action. If the rpm matched we could fully engage the other clutch immediately, but it does not. The numbers vary slightly, but for clarity, the PDK must manage to shed that $0.52 of KE, just as we must with a manual shift.
The engine is turning too fast, so the computer leaves the power low (or perhaps even with a closed throttle completely) while the other clutch is brought into engagement at a controlled rate. The flywheel energy is dumped into that torque path in a smooth way over a period I would need instruments to measure. (Perhaps one second?) As the engine slows to the new speed (roughly 4400 rpm), as it sheds that surplus kinetic energy of rotation through pumping losses but also through that 'new' clutch, the throttle is opened as well -- with very precise timing -- so that just as the surplus kinetic energy in the flywheel has been fully dumped, the engine is producing torque again and acceleration continues without interruption.
What has been done is the $0.52 of flywheel energy has been used to fill that 'notch' I asked about. The acceleration went from being a product of the engine's burning fuel delivered through the left torque path (let's say), to a bridge delivery of flywheel energy for some short period, until the engine's fuel -burning torque could be delivered through the right torque path.
That is one impressive transmission. The algorithm I just described will be obvious to any control engineer, but implementing it at road-car prices is another story. I understand now why they had to triple the 'power' of the on-board computer for this new engine to make the PDK feasible. I'm sure it also helps that they reduced the internal rotating mass for the new engine, so the amount of energy stored in the flywheel and engine that must be shed this way is not as much as it would have been with previous engines. That's important because the method I just described entails friction losses in the clutch that could be avoided if the PDK simply waited for the engine to slow from pumping load.
On top of all that, they turned the necessity into advantage. We enjoy that 'bump' when a gear changes and sure enough we feel it when the car is in sport mode. Almost certainly, they've done that by engaging the next clutch more quickly than in touring mode. That reduces the wear in what is a high-torque situation while also dumping more of the surplus kinetic energy in a lump that produces a bump that creates a grin in the driver. The opportunists.
Gary
#25
Nordschleife Master
Nice long answer, and I'm just as average guy,
but
how many "engines" of significant power are at work in a vehicle under a dv?
what is the difference between an engine producing a Torque and a flywheel with an angular momentum (accepting your flywheel proxy for the true "back force") and for that matter why do you pause to shift your street car and not in a racer? And for that matter, which "side" of the trans is your flywheel attached to?
Finally, a mass responds to the net F, your example seems to be leaving out a lot of F's of both signs. However net F is in opposite directions when accelerating and -accelerating.
but
how many "engines" of significant power are at work in a vehicle under a dv?
what is the difference between an engine producing a Torque and a flywheel with an angular momentum (accepting your flywheel proxy for the true "back force") and for that matter why do you pause to shift your street car and not in a racer? And for that matter, which "side" of the trans is your flywheel attached to?
Finally, a mass responds to the net F, your example seems to be leaving out a lot of F's of both signs. However net F is in opposite directions when accelerating and -accelerating.
Last edited by allegretto; 08-05-2011 at 08:54 AM.
#26
Rennlist Member
Marshal
Last edited by Marshal_Mercer; 08-05-2011 at 04:42 PM.
#27
Rennlist Member
Thread Starter
Join Date: Dec 2009
Location: Southern California
Posts: 1,429
Likes: 0
Received 15 Likes
on
15 Posts
In the example, my point is the same one Porsche alludes to without rubbing it in. If you don't manage those secondary forces, you end up with "a high level of discomfort." Not to mention wearing out the clutches quickly. Subtlety rules in a successful design.
Incidentally, I can't quite see why you guys are so hostile in this thread. I am praising your transmission you know.
Gary
#28
Rennlist Member
Thread Starter
Join Date: Dec 2009
Location: Southern California
Posts: 1,429
Likes: 0
Received 15 Likes
on
15 Posts
I think you're thinking of inertia, Marshal, which is an inherent property of matter. Momentum is another word for the kinetic energy of an object with direction included. It is computed as mass times velocity, so as velocity builds, so does the momentum of an object.
Or you may be remembering that in Newtonian physics momentum is conserved. That is, in a closed system, the net momentum remains constant. A car is not a closed system.
Gary
#29
Nordschleife Master
Okay, I guess we do have to pause for a physics lesson.
I think you're thinking of inertia, Marshal, which is an inherent property of matter. Momentum is another word for the kinetic energy of an object with direction included. It is computed as mass times velocity, so as velocity builds, so does the momentum of an object.
Or you may be remembering that in Newtonian physics momentum is conserved. That is, in a closed system, the net momentum remains constant. A car is not a closed system.
Gary
I think you're thinking of inertia, Marshal, which is an inherent property of matter. Momentum is another word for the kinetic energy of an object with direction included. It is computed as mass times velocity, so as velocity builds, so does the momentum of an object.
Or you may be remembering that in Newtonian physics momentum is conserved. That is, in a closed system, the net momentum remains constant. A car is not a closed system.
Gary
#30
Rennlist Member
Okay, I guess we do have to pause for a physics lesson.
I think you're thinking of inertia, Marshal, which is an inherent property of matter. Momentum is another word for the kinetic energy of an object with direction included. It is computed as mass times velocity, so as velocity builds, so does the momentum of an object.
Or you may be remembering that in Newtonian physics momentum is conserved. That is, in a closed system, the net momentum remains constant. A car is not a closed system.
Gary
I think you're thinking of inertia, Marshal, which is an inherent property of matter. Momentum is another word for the kinetic energy of an object with direction included. It is computed as mass times velocity, so as velocity builds, so does the momentum of an object.
Or you may be remembering that in Newtonian physics momentum is conserved. That is, in a closed system, the net momentum remains constant. A car is not a closed system.
Gary
Inertia is described as the quantity of a mass's resistance to changes in velocity (Newton's First Law of Motion), where Momentum is the product of mass and velocity (Newton's Second Law of Motion); Velocity is the vector of an object's speed in some direction.
Velocity, like Momentum, does not "build".
Change either the speed or the direction of an object having mass and the object's velocity changes. The word that we use to describe the rate of change in velocity over time is Acceleration. Force is the product of Mass and Acceleration. Rearranged, a = F/m.
Although a car is not a closed system, Acceleration is; it is assumed that Mass is a constant. I believe that you were looking to Acceleration, not Momentum, in your thinking.
Marshal
Last edited by Marshal_Mercer; 08-05-2011 at 05:34 PM.