Jump to content


Photo

Seamless Gearboxes


  • Please log in to reply
631 replies to this topic

#401 Magoo

Magoo
  • Member

  • 2,538 posts
  • Joined: October 10

Posted 10 March 2012 - 01:00

I'd like to take a moment to commend myself for not participating in this discussion.

Advertisement

#402 Tony Matthews

Tony Matthews
  • Member

  • 17,499 posts
  • Joined: September 08

Posted 10 March 2012 - 01:18

I think 'discussion' is giving it a greater status than it warrents. But I know what you mean...

#403 bigleagueslider

bigleagueslider
  • Member

  • 879 posts
  • Joined: March 11

Posted 10 March 2012 - 02:16

I think 'discussion' is giving it a greater status than it warrents. But I know what you mean...


I would disagree in principle. From a purely theoretical standpoint, the topic of discussion is actually quite interesting. However, I would agree that many of the posts seem pointless.

The question posed was can a gearbox shift truly be "seamless"? The conventional wisdom would say no, but that does not mean it's not possible. A definition of what "seamless" means would be a good start. Does it mean uninterrupted torque transfer during shifting? No change in input/output shaft speed during a shift? Does seamless mean instantaneous, or does it imply some finite period of time? I have never yet seen a manual gearbox with a truly instantaneous shift and no loss of torque transfer. But with concepts like Zeroshift and DCT's that give very rapid shifts, the whole issue may be academic.

An interesting historical analogy is that of the 17th century Scotsman James Ferguson. Ferguson is famous for Ferguson's Paradox. In order to prove his philosophical point that it's possible to do things that seem to defy logic, he invented a gear mechanism that caused an output shaft to turn backwards, come to a complete stop, then turn forwards, as he rotated the input shaft in one direction at varying speeds. It's now known as the Ferguson Mechanism, and the seemingly impossible principle behind it is known as Ferguson's Paradox. Definitely worth a read: http://chestofbooks....al-Paradox.html

slider


#404 cheapracer

cheapracer
  • Member

  • 10,388 posts
  • Joined: May 07

Posted 10 March 2012 - 04:05

I'd like to take a moment to commend myself for not participating in this discussion.



:lol: :lol:


I have never yet seen a manual gearbox with a truly instantaneous shift and no loss of torque transfer.


You must be kidding? This is a setup right? :lol:

You should read this recent thread ... http://forums.autosp...howtopic=161648


So Sam, it realy should be up to you to prove there is no gap, not for me to prove there is one.


Actually it is up to you to disprove a claim.


If you tried it regardless of any torque control or controlled clutch slip you would seriously damage the shift mechanism.
I am certain that this is the case, so you are way wrong.


Once again not relevant to your stance which is that seamless isn't possible and the damage is at no different a level than abusing normal dogs that millions of pimple faced young blokes do every weekend on road and dirt bikes - and many of them 'flat change' too, I for one won't own a dirt bike that won't flat change (some will, some won't). Engine and driveline are balanced during gear changes to prevent loss of traction more so than concern of damage.

You may be "certain" but I am stating facts. "certain" = opinion.

Edited by cheapracer, 10 March 2012 - 04:28.


#405 gruntguru

gruntguru
  • Member

  • 5,415 posts
  • Joined: January 09

Posted 10 March 2012 - 05:51

I to am well aware of how it 'works'.
Doesnt tell you if torque is being constantly transfered though does it.

If you were aware of how it works you would know there is no gap so - sorry - epic fail.

#406 gruntguru

gruntguru
  • Member

  • 5,415 posts
  • Joined: January 09

Posted 10 March 2012 - 06:00

That has nothing to do with Zeroshift's seamless changing which can be done regardless of positive, negative or constant input torque load, up or down the gears - there is no reliance on anything other than gear rotation to engage/disengage.


If you tried it regardless of any torque control or controlled clutch slip you would seriously damage the shift mechanism.
I am certain that this is the case, so you are way wrong.

That is patently incorrect. There is nothing in the Zeroshift mechanism that is any more fragile than a dog ring box. In addition the mechanism is fail-safe regarding selection of two gears simultaneously - as soon as a gear is overdriven it disengages.

#407 gruntguru

gruntguru
  • Member

  • 5,415 posts
  • Joined: January 09

Posted 10 March 2012 - 06:11

Nice explanation shows the mechanism very well.
Omits to explain what the torque is doing through the process however.

1st to 2nd shift as follows.

Engaged in first gear the clutch disengages (to a carefuly controlled slip state if you wish), as engine control reduces torque from the engine.
The second gear selector assembly is moved along its slot in the output shaft hub into the 2nd gear position and the 2nd gear shift buttons apply pressure to 2nd gear through their springs.
BOTH SELECTORS AND FORKS ARE IN THE ENGAGED POSITION AT THIS POINT.
At the same time the torque has reduced sufficiently from the input shaft via the laygear to 1st gear, which now allows the 1st gear selector fork to be moved to the neutral position.
The reduction of input torque and the spring detents on the fork mechanism, cause the 1st gear buttons to spring out of semi engagement to full disengagement, just before the 2nd gear buttons fully engage with the drive faces on the 2nd gear.

WRONG!
The first gear buttons do not disengage until the second gear buttons are fully engaged and OVERDRIVING FIRST GEAR.

IT IS CLEAR YOU DO NOT UNDERSTAND HOW IT WORKS!

#408 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 10:06

WRONG!
The first gear buttons do not disengage until the second gear buttons are fully engaged and OVERDRIVING FIRST GEAR.

IT IS CLEAR YOU DO NOT UNDERSTAND HOW IT WORKS!


Dear grunt I am sorry to say, it is impossible for both gears to be driving at the same time through fully engaged buttons.
The 1st and 2nd gears are rotating at DIFFERENT speeds and both are constantly engaged through their gear teeth to the SOLID laygear.
Any shift overlap between full engagement between these gears HAS GOT to be by using slip, disengagement or damage and wear.
In this case it is through the use of spring detent buttons, sprung selector forks and slip between components.
These springs are being asked to do far to much for a road car application.

It is clear that it is you who do not understand how it works.

p.s sweetheart, if you were correct which you are NOT, why bother to disengage the 1st gear in 2nd at all???????????????????????????????????
Guess what all you clever engineers, THERE IS A GAP.

Edited by 24gerrard, 10 March 2012 - 10:20.


#409 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 10:10

That is patently incorrect. There is nothing in the Zeroshift mechanism that is any more fragile than a dog ring box. In addition the mechanism is fail-safe regarding selection of two gears simultaneously - as soon as a gear is overdriven it disengages.


WRONG WRONG WRONG

#410 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 10:11

If you were aware of how it works you would know there is no gap so - sorry - epic fail.


IT IS YOUR FAIL GRUNT
WE CANT BE RIGHT ALL THE TIME CAN WE


#411 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 10:50

I would disagree in principle. From a purely theoretical standpoint, the topic of discussion is actually quite interesting. However, I would agree that many of the posts seem pointless.

The question posed was can a gearbox shift truly be "seamless"? The conventional wisdom would say no, but that does not mean it's not possible. A definition of what "seamless" means would be a good start. Does it mean uninterrupted torque transfer during shifting? No change in input/output shaft speed during a shift? Does seamless mean instantaneous, or does it imply some finite period of time? I have never yet seen a manual gearbox with a truly instantaneous shift and no loss of torque transfer. But with concepts like Zeroshift and DCT's that give very rapid shifts, the whole issue may be academic.

An interesting historical analogy is that of the 17th century Scotsman James Ferguson. Ferguson is famous for Ferguson's Paradox. In order to prove his philosophical point that it's possible to do things that seem to defy logic, he invented a gear mechanism that caused an output shaft to turn backwards, come to a complete stop, then turn forwards, as he rotated the input shaft in one direction at varying speeds. It's now known as the Ferguson Mechanism, and the seemingly impossible principle behind it is known as Ferguson's Paradox. Definitely worth a read: http://chestofbooks....al-Paradox.html

slider


Well written slider.
You have hit the nail (or shift buttons) right on the head as usual, if I ever start a team I will give you a call (and gruntguru funnily enough).
My statement is that these layshaft stepped boxes are quick shift, not 'zeroshift' or 'seamless'.
The definition is not scientifically correct.
You are absolutely correct when you say it could be academic after all the driveshaft torque shows a smooth shift which is felt far less than in a conventional manual shift, unless you change very carefully with a syncro cone mechanism.

The issue that makes me so adamant to show this marketing description up for what it is does not directly come from the search for the gap in torque delivery.
It is the way the engine control and clutch release needed to achieve success with the machanism is carefuly ignored or played down in any descriptions or videos of how the system works.
These shifts ar NOT an efficient way to transfer torque through a gear change, unless it is a power OFF coast or deceleration shift.
They are a good way to smooth out a manual gear shift and save money in production. A good auto box is much more expensive.
I have had sight of shift mechanisms for F1 manual layshaft gear boxes and these have to be totaly different to the mechanism under discusion.
These do not waste time disengageing and slipping the clutch for long periods and the engine control is now more difficult this year because of the banning of exhaust blown diffusers. Gearbox problems are predicted through the season because of this and other reasons.

PS. I love the paradox that is not a paradox slider. When is a gap not a gap, hahaha

Edited by 24gerrard, 10 March 2012 - 10:58.


#412 GeoffR

GeoffR
  • Member

  • 497 posts
  • Joined: January 04

Posted 10 March 2012 - 11:28

expect me to respond in kind. You will find me better at it and with a lot more experience.

Perhaps you should try to find something that I posted here http://www.pentaxfor...hotography.html a while ago before you start making 'keyboard warrior' threats, and you may get an understanding as to why they mean f*** all to me!
When is an idiot not an idiot, ha ha!
As for this thread, I've seen/read enough!

Edited by GeoffR, 10 March 2012 - 11:45.


#413 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 12:17

Perhaps you should try to find something that I posted here http://www.pentaxfor...hotography.html a while ago before you start making 'keyboard warrior' threats, and you may get an understanding as to why they mean f*** all to me!
When is an idiot not an idiot, ha ha!
As for this thread, I've seen/read enough!


You instigated the conflict Geoff.
I am cool about it.
You obviously do care otherwise you would not have posted again saying you have seen read enough.
The self aggrandisement now becomes obvious.

I am sure you are very capable in what you do, so am I.

#414 Fondles

Fondles
  • Member

  • 50 posts
  • Joined: July 11

Posted 10 March 2012 - 13:29

Dear grunt I am sorry to say, it is impossible for both gears to be driving at the same time through fully engaged buttons.


Quite fortunately, that doesn't happen with the Zeroshift mechanism.
So where's the problem?


#415 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 13:50

Quite fortunately, that doesn't happen with the Zeroshift mechanism.
So where's the problem?


I do not have a problem at all.
Gruntguru does, he is saying both gears are driving at the same time.
He thinks that the 2nd gear bullets 'drive' at the same time as the already 'driving' bullets engaged to 1st gear.

#416 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 14:02

Posted Image

propshaft rpm 1100rpm start of graph, 1125rpm (start of shift bracket, not overlap) 1125rpm at end of shift bracket, 1200rpm end of graph.
propshaft rpm 1000rpm at shift overlap.


2nd gear selector engages at 6.54, the 2nd gear bullets do not, the 1st gear selector is already engaged and 1st gear bullets are driving.
Ist gear selector disengages between 6.54 and 6.55, the 1st gear bullets remain engaged until torque comes off their drive surfaces.
1st gear bullets disengage at 6.55 coincident with the drop in output shaft rpm.
2nd gear bullets engage fully and drive at 6.57 coincident with the torque spike on the propshaft trace.

The shift actuation window is 0.03 seconds
The shift overlap is 0.02 seconds.
The powertrain shift control envelope is shown by the brackets but actualy extends off the graph to reach full power on drive.
It is a way way slow gear shift if everything is included.

Edited by 24gerrard, 10 March 2012 - 14:54.


#417 NTSOS

NTSOS
  • Member

  • 692 posts
  • Joined: February 05

Posted 10 March 2012 - 14:13

I have never yet seen a manual gearbox with a truly instantaneous shift and no loss of torque transfer.

slider


Would you please clarify?

A. Does "and no loss of torque transfer" mean less than 100% torque transfer, but no complete interruption or gap?

or

B. Does "and no loss of torque transfer" mean less than 100% torque transfer, but implying an interruption or gap?

Thanks!

John

Edited by NTSOS, 10 March 2012 - 14:22.


#418 Spoofski

Spoofski
  • Member

  • 262 posts
  • Joined: April 10

Posted 10 March 2012 - 15:23

These do not waste time disengageing and slipping the clutch for long periods and the engine control is now more difficult this year because of the banning of exhaust blown diffusers. Gearbox problems are predicted through the season because of this and other reasons.

Care to elaborate on this?


#419 NTSOS

NTSOS
  • Member

  • 692 posts
  • Joined: February 05

Posted 10 March 2012 - 16:27

1st gear bullets disengage at 6.55 coincident with the drop in output shaft rpm.


However, I think that the overall output speed average does not drop. If you take a ruler to the output shaft speed on the graph, you will see a slight drop followed by three rises. The average is that there is no overall drop, just a wobble caused by the sudden drop to the 2nd gear torque level.

John

Advertisement

#420 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 16:31

Care to elaborate on this?


Blowing the exhaust gas over the rear diffuser required that the engine management set the engine so as to burn fuel to increase mass flow without noticeable torque output. This made it easier to reduce torque flow through the shift components in most of the required gear shifts.
It also gave the teams a clever way to create traction limiting control subject to scrutineering.

#421 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 16:37

However, I think that the overall output speed average does not drop. If you take a ruler to the output shaft speed on the graph, you will see a slight drop followed by three rises. The average is that there is no overall drop, just a wobble caused by the sudden drop to the 2nd gear torque level.

John


I agree it is close and difficult to read accurately.
It would be subject to how much the clutch was biting and how much torque was made available from the engine.
After the bracketed area it is obvious there is reasonable acceleration to off the graph.
I dont think there is as much leading up to the shift but again that is variable depending on EM.

#422 Spoofski

Spoofski
  • Member

  • 262 posts
  • Joined: April 10

Posted 10 March 2012 - 19:26

Blowing the exhaust gas over the rear diffuser required that the engine management set the engine so as to burn fuel to increase mass flow without noticeable torque output. This made it easier to reduce torque flow through the shift components in most of the required gear shifts.
It also gave the teams a clever way to create traction limiting control subject to scrutineering.

Not sure where you got the first bit from as it makes no sense.. The throttle is under ECU control during downshifts and capable of managing all torque between full braking and as much as is needed.

The second bit: you might mean some stability control on entry rather than traction control in the traditional sense, but this is old hat and well predates seamless boxes. I no longer deal with the SECU so don't know what the limitations are on doing this but they were considerable even before the SECU and I'd be surprised if they had been relaxed since.

#423 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 21:15

Not sure where you got the first bit from as it makes no sense.. The throttle is under ECU control during downshifts and capable of managing all torque between full braking and as much as is needed.

The second bit: you might mean some stability control on entry rather than traction control in the traditional sense, but this is old hat and well predates seamless boxes. I no longer deal with the SECU so don't know what the limitations are on doing this but they were considerable even before the SECU and I'd be surprised if they had been relaxed since.


It is possible to program gear shift control (with the right components fitted), so as to modulate upshifts to reproduce at least part of a TL power on result.
I believe this is possible within regulations.

There are still gains to be made in balancing engine control to assist shifts as well as mass exhaust gas requirements on overrun.
With high downforce, it will remain relatively simple to prevent car unbalance during shifts.
It is the KERS that will pose the biggest balance problem into the future.
With its increasing use and higher energy operation, keeping the rear end under control will get more difficult.
If conventional brakes are retained and KERS remains only on the rear axle, 2014 is going to be interesting.

#424 Spoofski

Spoofski
  • Member

  • 262 posts
  • Joined: April 10

Posted 10 March 2012 - 21:36

It is possible to program gear shift control (with the right components fitted), so as to modulate upshifts to reproduce at least part of a TL power on result.
I believe this is possible within regulations.

There are still gains to be made in balancing engine control to assist shifts as well as mass exhaust gas requirements on overrun.
With high downforce, it will remain relatively simple to prevent car unbalance during shifts.
It is the KERS that will pose the biggest balance problem into the future.
With its increasing use and higher energy operation, keeping the rear end under control will get more difficult.
If conventional brakes are retained and KERS remains only on the rear axle, 2014 is going to be interesting.

I suppose I was more wondering why you thought gearbox reliability would suffer with the EBD 'ban'.

#425 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 10 March 2012 - 22:45

I suppose I was more wondering why you thought gearbox reliability would suffer with the EBD 'ban'.


I see what you mean.

To be honest having found out just recently what some of the control strategies have been for shift control, I am now not sure that this years changes on their own will effect reliability. The limit on number of units is more likely to.
Changing over to 8 ratios in 2014 will help spread loads and smooth things a bit but a I still feel that IMO 2014 will see major powertrain problems.

My main criticism is that from 2014 it will be almost impossible to reduce downforce through regulation without changing everything else radicaly.
With high downforce maintained I do not see any further development scope for mechanical systems after that date.
Energy recovery will then be the only relevent and interesting part of F1 cars and that will be strangled by aero requirements.

The red bulls keeps winning from mainly aero tuning and by 2014, IMO the interest from the public will be much less than today.
I see developments and regulations in sports car racing and rallying overtakeing F1 if the teams Bernie and the FIA dont take care of business.

Moto GP has already taken the danger and skill aspects away from F1 in the minds of many and the corruption from people like the Murdochs causing total disruption to TV coverage could be the final nail.

Edited by 24gerrard, 10 March 2012 - 22:49.


#426 gruntguru

gruntguru
  • Member

  • 5,415 posts
  • Joined: January 09

Posted 11 March 2012 - 01:41

Dear grunt I am sorry to say, it is impossible for both gears to be driving at the same time through fully engaged buttons.

Forget about clutches, dampers, ignition cut etc. I am going to describe a full power racing shift in a zeroshift box with no torque reduction techniques applied. I refer you to Cheapracer's excellent illustration which demonstrates clearly that both gears are driving simultaneously. Look at figure 2 where the red hub (second gear) which is travelling faster than the blue hub (first gear) has just made contact with its "bullet".
1. At this instant first gear is still doing all the driving since second gear has only just contacted its bullet and is not yet exerting any force.
2. A microsecond later the picture still looks the same. Both bullets are still in contact with their hubs however 2nd gear is now taking some load and the 1st gear load has been reduced by an equivalent amount.
3. The progress described in the previous step continues for a small but finite period. The load on the 2nd gear "bullet" continues to increase while the load on 1st gear correspondingly reduces until . . . .
4. The load on 1st gear finally drops to zero and 2nd gear is doing all the driving. The picture still looks like "2" in Cheapy's illustration but it is about to change.
5. Now that 1st gear is fully "unloaded" the 1st gear hub starts to outrun its bullet and we move to diagram 3 and 4 in Cheapy's illustration.

The key point here is that diagram 2 (both gears driving) is a state that exists for a finite period of time. It has to, because the disengagement of 1st gear relies on the engagement and "overdriving" of the mainshaft by 2nd gear.

NOTE. If the components were infinitely rigid (not possible in the real world) the "overlap" period with both gears driving would be zero. (There would still be no gap in torque transmission).

The secret to this system is that the pawls can work independently of each other although being on the same hub.



Posted Image


Edited by gruntguru, 11 March 2012 - 04:03.


#427 Kelpiecross

Kelpiecross
  • Member

  • 928 posts
  • Joined: October 10

Posted 11 March 2012 - 05:59

[quote name='bigleagueslider' date='Mar 10 2012, 13:16' post='5578429']

- gear mechanism that caused an output shaft to turn backwards, come to a complete stop, then turn forwards

BLS - in the the normal arrangement of this "paradox" there are three stacked apparently identical gearwheels as the output - one does not rotate, one rotates slowly in one direction and the other slowly in the opposite direction.
It is not so much of a "paradox" as an illusion or trick as the three stacked gearwheels, although of identical diameter, have different numbers of teeth. The different number of teeth is not noticeable.
The general principle of the gear train is interesting as very high reduction ratios (of 1000:1 or so) can be generated by just two gearwheels. This type of gear layout is used sometimes in aircraft undercarriage and flap mechanisms.

PS - 15000:1 in the FW190 undercarriage.

Edited by Kelpiecross, 11 March 2012 - 06:04.


#428 Tony Matthews

Tony Matthews
  • Member

  • 17,499 posts
  • Joined: September 08

Posted 11 March 2012 - 08:59

... Cheapracer's excellent illustration...

I have to admit that I find the illustration no help at all, although I don't have any problem understanding the principle. I still think rachael's post of months ago describes the seamless system best - I know her two examples are different from the Zeroshift system, but surely it only takes one mechanism to work to prove the point.

#429 Fondles

Fondles
  • Member

  • 50 posts
  • Joined: July 11

Posted 11 March 2012 - 13:29

Gerrard, could you please give us your definition of 'seamless' ?

To me it would mean that the torque output from the gearbox would not alter by more than a reasonable percentage and with no interruptions. If I had to put a number on that I'd say +/- 20% and no zero reading for any length of time.

And you?

#430 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 11 March 2012 - 16:28

Forget about clutches, dampers, ignition cut etc. I am going to describe a full power racing shift in a zeroshift box with no torque reduction techniques applied. I refer you to Cheapracer's excellent illustration which demonstrates clearly that both gears are driving simultaneously. Look at figure 2 where the red hub (second gear) which is travelling faster than the blue hub (first gear) has just made contact with its "bullet".
1. At this instant first gear is still doing all the driving since second gear has only just contacted its bullet and is not yet exerting any force.
2. A microsecond later the picture still looks the same. Both bullets are still in contact with their hubs however 2nd gear is now taking some load and the 1st gear load has been reduced by an equivalent amount.
3. The progress described in the previous step continues for a small but finite period. The load on the 2nd gear "bullet" continues to increase while the load on 1st gear correspondingly reduces until . . . .
4. The load on 1st gear finally drops to zero and 2nd gear is doing all the driving. The picture still looks like "2" in Cheapy's illustration but it is about to change.
5. Now that 1st gear is fully "unloaded" the 1st gear hub starts to outrun its bullet and we move to diagram 3 and 4 in Cheapy's illustration.

The key point here is that diagram 2 (both gears driving) is a state that exists for a finite period of time. It has to, because the disengagement of 1st gear relies on the engagement and "overdriving" of the mainshaft by 2nd gear.

NOTE. If the components were infinitely rigid (not possible in the real world) the "overlap" period with both gears driving would be zero. (There would still be no gap in torque transmission).


Very good diagram of the basic mechanical operation

However grunt you have chosen to ignore the essential modulation of the operation and the damping.

Your description of the operation would produce a graph nothing like the one shown (Zeroshift POWER ON).
Your graph would show a propshft torque trace running to the center of the graph with a slight rise dependent on acceleration.
The trace would then drop verticaly at a right angle until it reached the next higher gear torque level.
It would then simply turn through another sharp 90 degrees to horizontal and rise slightly to off the graph.
You are saying there is no break in torque, so the shift overlap would take zero time.
If it took any time at all there would still be a GAP.
The shift would also be a sudden bang of engagement and not at all smooth like the real zeroshift 'slow' torque reduced shift is.

There are a number of reasons why you cannot simply ignore the control needed on engine torque delivery and the clutch disengagement (slip).
The engine rpm has to be suddenly reduced to the new higher gear level during the shift and within the shift window.
It cannot be achieved in the zero time you state.
The shift cannot be achieved without controlling engine torque, because the road engine shown has a heavy flywheel and clutch with a great deal of rotating inertia.
Even just engine control is insufficient, so the clutch has also to be disengaged (slipped).
You are close in your description to over run shift operation in F1 and similar though.
This is because F1 cars are very light and the flywheel and clutch is also very much lighter than a road cars.
There is also no need for a smooth shift in an F1 box.
Unfortunately a spring detent mechanism such as this one shown in the videos and diagrams would not last for long shifting at very high speed.
AFAIK F1 shift mechanisms are all ratchet types.

Edited by 24gerrard, 11 March 2012 - 16:42.


#431 carlt

carlt
  • Member

  • 1,076 posts
  • Joined: June 09

Posted 11 March 2012 - 17:00

Very good diagram of the basic mechanical operation

However grunt you have chosen to ignore the essential modulation of the operation and the damping.

Your description of the operation would produce a graph nothing like the one shown (Zeroshift POWER ON).
Your graph would show a propshft torque trace running to the center of the graph with a slight rise dependent on acceleration.
The trace would then drop verticaly at a right angle until it reached the next higher gear torque level.
It would then simply turn through another sharp 90 degrees to horizontal and rise slightly to off the graph.
You are saying there is no break in torque, so the shift overlap would take zero time.
If it took any time at all there would still be a GAP.
The shift would also be a sudden bang of engagement and not at all smooth like the real zeroshift 'slow' torque reduced shift is.

There are a number of reasons why you cannot simply ignore the control needed on engine torque delivery and the clutch disengagement (slip).
The engine rpm has to be suddenly reduced to the new higher gear level during the shift and within the shift window.
It cannot be achieved in the zero time you state.
The shift cannot be achieved without controlling engine torque, because the road engine shown has a heavy flywheel and clutch with a great deal of rotating inertia.
Even just engine control is insufficient, so the clutch has also to be disengaged (slipped).
You are close in your description to over run shift operation in F1 and similar though.
This is because F1 cars are very light and the flywheel and clutch is also very much lighter than a road cars.
There is also no need for a smooth shift in an F1 box.
Unfortunately a spring detent mechanism such as this one shown in the videos and diagrams would not last for long shifting at very high speed.
AFAIK F1 shift mechanisms are all ratchet types.

i don't think anyone has argued that there is not a reduction in torque at the change , it just doesn't go to zero and therefore there is no seam

#432 gruntguru

gruntguru
  • Member

  • 5,415 posts
  • Joined: January 09

Posted 11 March 2012 - 23:22

However grunt you have chosen to ignore the essential modulation of the operation and the damping.

The modulation and damping are not essential. The Zeroshift can be operated without in the same way that a dog ring box can be operated with full throttle and no clutch. And YES there will be an upwards spike in torque if the Zeroshift is operated without clutch-slip, dampers and engine modulation as there is with a dog ring box.

Clutch-slip, dampers and engine modulation are old technology that have nothing to do with the Zeroshift patent. I am sure it suits you to focus on these irrelevant matters to avoid focusing on the real issue of whether there is a torque gap in the zeroshift upshift.

You are saying there is no break in torque, so the shift overlap would take zero time.
If it took any time at all there would still be a GAP.

This indicates a total misunderstanding of the mechanism and the physics involved. The "overlap" is a period where BOTH GEARS ARE DRIVING due to elasticity of the components. The overlap does not represent a gap in the shift - drive is continuuous before, during and after the overlap.

#433 gruntguru

gruntguru
  • Member

  • 5,415 posts
  • Joined: January 09

Posted 11 March 2012 - 23:23

i don't think anyone has argued that there is not a reduction in torque at the change , it just doesn't go to zero and therefore there is no seam

There is no reduction in torque at the change.

#434 kikiturbo2

kikiturbo2
  • Member

  • 657 posts
  • Joined: December 04

Posted 12 March 2012 - 00:08

Gerrard, could you please give us your definition of 'seamless' ?

To me it would mean that the torque output from the gearbox would not alter by more than a reasonable percentage and with no interruptions. If I had to put a number on that I'd say +/- 20% and no zero reading for any length of time.

And you?


have tried to get his definition of seamless for a while with no success... Even mentioned a production box that has an increase in torque on the output shaft during shifting... to no avail.. :)

#435 NTSOS

NTSOS
  • Member

  • 692 posts
  • Joined: February 05

Posted 12 March 2012 - 00:52

There is no reduction in torque at the change.



Interesting, so there is zero torque drop during the shift and the trace is simply tracking the inevitable drop to 2nd gear level torque?

John

#436 gruntguru

gruntguru
  • Member

  • 5,415 posts
  • Joined: January 09

Posted 12 March 2012 - 01:00

The raw zeroshift mechanism would produce an inertial "spike" in torque followed by a drop to the 2nd gear torque level. Any variations from this as seen on the graph are artifacts of the "smoothing" systems added to the transmission or disturbances and oscillations caused by elasticity and masses in the system.

#437 NTSOS

NTSOS
  • Member

  • 692 posts
  • Joined: February 05

Posted 12 March 2012 - 01:30

The raw zeroshift mechanism would produce an inertial "spike" in torque followed by a drop to the 2nd gear torque level. Any variations from this as seen on the graph are artifacts of the "smoothing" systems added to the transmission or disturbances and oscillations caused by elasticity and masses in the system.


Exactly! I have two new graphs, one damped and one undamped that were run back to back.....I'm trying to obtain permission to post them.

John

#438 munks

munks
  • Member

  • 348 posts
  • Joined: January 03

Posted 12 March 2012 - 01:49

The "overlap" is a period where BOTH GEARS ARE DRIVING due to elasticity of the components.


This is where you guys talk past each other. 24gerrard is going to respond the same way he always does to this statement, which is by changing it slightly: "It is impossible for both gears to be FULLY engaged" which is, of course, not what you're saying. I'm just trying to give you a head's up that repeating the same fact is going to result in the same response.

By the way, my vote is that torque never drops to zero. I can't say that there's not any dip (because the elasticities will obviously cause some complex sinusoidal-type response), but it won't drop to zero. By my personal definition, that means there is no gap.


#439 bigleagueslider

bigleagueslider
  • Member

  • 879 posts
  • Joined: March 11

Posted 12 March 2012 - 01:54

Would you please clarify?

A. Does "and no loss of torque transfer" mean less than 100% torque transfer, but no complete interruption or gap?

or

B. Does "and no loss of torque transfer" mean less than 100% torque transfer, but implying an interruption or gap?

Thanks!

John


John,

I meant "no loss of torque transfer" to mean an interruption where there is zero torque transfer.

Logic tells us that you cannot have two rigidly coupled gear sets on common rotating shafts with different ratios. And any mechanism that switches the power path from one to the other would seem to require some finite amount of time to operate, Zeroshift included. A shift event requires a change in speed of the input/output shafts, meaning there is an acceleration. Theoretically, if the shift event was instantaneous that would also imply infinite acceleration of the rotating masses, and thus infinite forces. The drivetrain would either suffer structural failure, stall the engine, or spin the tires.

However, there is one possible exception. There are some manual gearbox designs that use a complex system of selectable two-way overrunning clutches for each gear. But I have yet to see if they are reliable in operation.

slider



Advertisement

#440 NTSOS

NTSOS
  • Member

  • 692 posts
  • Joined: February 05

Posted 12 March 2012 - 02:51

John,

I meant "no loss of torque transfer" to mean an interruption where there is zero torque transfer.

Logic tells us that you cannot have two rigidly coupled gear sets on common rotating shafts with different ratios. And any mechanism that switches the power path from one to the other would seem to require some finite amount of time to operate, Zeroshift included. A shift event requires a change in speed of the input/output shafts, meaning there is an acceleration. Theoretically, if the shift event was instantaneous that would also imply infinite acceleration of the rotating masses, and thus infinite forces. The drivetrain would either suffer structural failure, stall the engine, or spin the tires.

However, there is one possible exception. There are some manual gearbox designs that use a complex system of selectable two-way overrunning clutches for each gear. But I have yet to see if they are reliable in operation.

slider


Slider,

Thanks for the clarification....I *may* be posting new graphs and I would very much appreciate your input! Not sure it's going to happen though.

John

#441 carlt

carlt
  • Member

  • 1,076 posts
  • Joined: June 09

Posted 12 March 2012 - 08:06

There is no reduction in torque at the change.


I can't see how that is possible , as munks posted , with the clutch slip , damping and drive train elasticity I would assume that torque would modulate

but I see no reason with Zeroshift for torque to drop to zero

#442 ferruccio

ferruccio
  • Member

  • 446 posts
  • Joined: July 09

Posted 12 March 2012 - 09:06

Gerrard, could you please give us your definition of 'seamless' ?


His definition is different to everyone else. Everyone else understands it as NO GAP while he understands it as 'continuous' (smooth) which really is his right. Because of this 'misalignment' all the arguments become pointless.

AFAIK F1 shift mechanisms are all ratchet types.


No they're not all ratchet. You're the expert. Shouldn't you know this?


#443 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 12 March 2012 - 09:31

This is where you guys talk past each other. 24gerrard is going to respond the same way he always does to this statement, which is by changing it slightly: "It is impossible for both gears to be FULLY engaged" which is, of course, not what you're saying. I'm just trying to give you a head's up that repeating the same fact is going to result in the same response.

By the way, my vote is that torque never drops to zero. I can't say that there's not any dip (because the elasticities will obviously cause some complex sinusoidal-type response), but it won't drop to zero. By my personal definition, that means there is no gap.


You are correct munks.
Nothing has changed.
The zeroshift components shown are not capable of sustained shifting without modulation to the levels shown.
The operation has to be modulated to a point where there is no torque transfer from input to output.
Torque has to be reduced at input to disengage the low gear bullets and with modulation there is a point after this where no torque from input is reaching output.
Without modulation the mechanism might work a couple of times and create a shift impact like a sledge hammer, that would shake the car enough the make the drivers teeth rattle.
There would still be a seam at the point of overlap caused by the impact.
It would be a result of the drive train winding up as engine torque came on with a sledge hammer bang at the high revs it was turning in low gear and letting go as the engine was dragged down to the lower revs required for the new high gear.

#444 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 12 March 2012 - 09:43

Interesting, so there is zero torque drop during the shift and the trace is simply tracking the inevitable drop to 2nd gear level torque?

John


It is but it is read at the propshaft and does not show the torque at the input and output parts of the shift mechanism.
The readings there are masked by the output damper.

#445 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 12 March 2012 - 09:48

have tried to get his definition of seamless for a while with no success... Even mentioned a production box that has an increase in torque on the output shaft during shifting... to no avail.. :)


Seamless means without any gap in torque transfer at any point along the torque path.
Reading the torque at the propshaft does not show the whole torque path from the crankshaft to the rear wheels.

It is a simple matter to build a gearshift to achieve an increase in the torque read at the propshaft.
In fact you could easily modify the shift graph shown to achieve it with a different rating of damper.
This still does not make the gearbox seamless on its own.

#446 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 12 March 2012 - 09:51

The raw zeroshift mechanism would produce an inertial "spike" in torque followed by a drop to the 2nd gear torque level. Any variations from this as seen on the graph are artifacts of the "smoothing" systems added to the transmission or disturbances and oscillations caused by elasticity and masses in the system.


I agree, there would still be a gap, much smaller but still a gap.

#447 cheapracer

cheapracer
  • Member

  • 10,388 posts
  • Joined: May 07

Posted 12 March 2012 - 09:54

i don't think anyone has argued that there is not a reduction in torque at the change , it just doesn't go to zero and therefore there is no seam


Interesting, so there is zero torque drop during the shift and the trace is simply tracking the inevitable drop to 2nd gear level torque?

John


If the gear ratio is the same then there will be a seamless, uninterrupted torque change during the entire process.



"It is impossible for both gears to be FULLY engaged" ........


Logic tells us that you cannot have two rigidly coupled gear sets on common rotating shafts with different ratios.


That is because people do not factor the dog's lash into their thinking, both gears are fully engaged but the power flow for each ratio has some time of separation due to the dog's lash - do not confuse the word "seperation" meaning a torque gap, while engaging one ratio has separation then the other gear has separation while disengaging but the changeover between the 2 ratio's applying drive is instantaneous (no gap).


And any mechanism that switches the power path from one to the other would seem to require some finite amount of time to operate,


Well I for one am not sure how to define the 'no measurable time' that it takes for example a 'sprag clutched' gear that's sped up to take over from the previous gear or the Zeroshift mechanism, these 2 are probably best...

instant

: an infinitesimal space of time; especially : a point in time separating two states

instantaneous

1. Occurring or completed without perceptible delay:

Edited by cheapracer, 12 March 2012 - 10:01.


#448 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 12 March 2012 - 09:59

His definition is different to everyone else. Everyone else understands it as NO GAP while he understands it as 'continuous' (smooth) which really is his right. Because of this 'misalignment' all the arguments become pointless.

No they're not all ratchet. You're the expert. Shouldn't you know this?


A gap in torque transfer, is a gap, is a gap, is a gap.
Cant see how that is difficult to understand.

I said AFAIK they were all ratchet types.
There are other methods of operation of which I am also aware.

If you know different please enlighten us.
Technology is the purpose of this thread.

#449 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 12 March 2012 - 10:03

If the gear ratio is the same then there will be a seamless, uninterrupted torque change during the entire process.







That is because people do not factor the dog's lash into their thinking, both gears are fully engaged but the power flow for each ratio has some time of separation due to the dog's lash - do not confuse the word "seperation" meaning a torque gap, while engaging one ratio has separation then the other gear has separation while disengaging but the changeover between the 2 ratio's applying drive is instantaneous (no gap).




Well I for one am not sure how to define time that it takes for example a 'sprag clutched' gear that's sped up to take over from the previous gear or the Zeroshift mechanism....

instant

: an infinitesimal space of time; especially : a point in time separating two states

instantaneous

1. Occurring or completed without perceptible delay:


Result BANG drive train windup and rebound creating a GAP in torque to the rear wheels.
Most likely to also result in major geartrain damage.

#450 24gerrard

24gerrard
  • Member

  • 2,008 posts
  • Joined: March 11

Posted 12 March 2012 - 10:06

Exactly! I have two new graphs, one damped and one undamped that were run back to back.....I'm trying to obtain permission to post them.

John


This would be great.