Clutchless shifting a dog gearbox. 2

[daveylynx]

We are currently thinking of incorporating a full throttle gear shift system to our racing transmission. The system detects when a shift is about to be made and cuts the spark to the engine during the shift. The idea is that the clutch isn't used during up-shifts and shift time can be reduced to less than 100ms, giving obvious advantages.

What I would like to know is this: how much extra stress is placed on the transmission when shifting without the clutch? If you shift with the clutch, the drive is obviously taken up slowly enough to minimise shock loadings as the dog faces meet. However, if the clutch isn't used, the gearbox has to mechanically force the engine to lose maybe 2000rpm to match the next gear - in an instant! This will obviously put shock loadings through the entire transmission, but are the forces involved likely to be more than the forces created by the engine torque?

I appreciate that if the rotating inertia of the engine is greater and the speed mis-match between gears is greater, the shock loading will be higher, but can anyone guestimate which might place the greatest forces on the transmission - the engine torque of 200ft/lbs or shock loads due to engine being forced to reduce speed as the dogs engage. Remember that as the dogs engage, the engine isn't producing power but is spinning perhaps 2000rpm too fast to match the next gear.

 

[Milanoguy]

Hi Dave

Well I don’t have the precise answer for you, but cluchless upshifts are why people use dog engagement transmissions in the first place. Almost all racing motorcycles, and a lot of street ones upshift without the clutch. As I understand it, a downshift produces more stress on a transmission than an upshift. In an upshift the amount of torque multiplication is reduced, in a downshift the torque multiplication is increased, thus the downshift is more stressful.

More to the point, when you cut the spark, the engine is effectively in a no load state. In such a state, engine rpm can change dramatically over a tenth of a second.

Bye for now

 

[NickE]

Shock loading can be VERY harmful to the gears in the transmission. Just google: shock loading USDM Subaru WRX transmission, you will find many anecdotal references to stripped 2nd gear and etc.

There is a redline gear oil called HeavyWt Shockproof. It has some funky qualities. Its really slippery, viscous as all get out, refuses to release from metal surfaces, has a suspension of solids to absorb shock, a beautiful strawberry color, and smells to high heaven!

I'm on my third mix with their (redline's) 75W90NS (keeps some synchro action when cold) and have no trouble. I was reccommended this by several performance rally teams.

I'm sure that someone who knows more about dog-boxes will be along presently. Keep the shiny side up.


Nick
I love materials science!

 

[evelrod]

Nick, I have a rather laaaaaarge history with broken gearboxes from some 50 years of racing...I think I have broken just about every part that will fit in a manual box at one time or another. "Dog boxes" are like all other manual transmissions, some good and some a lot better. I have had a passing aquaintance with a Jack Knight four and a JK five speed...both shifted with a conventional lever and all upshifts were clutchless as it's extrememly difficult to do a proper up with a clutch...down I always used the clutch with a little 'heel toe'. Some don't and they seem to get by...quien sabe! I have used ATF in my gearboxes for many years with no problems related to the lubricant aside a minor leak or two.

I had the chance to take my crew chiefs much modified WRX out for a few laps at Las Vegas last year...handling acceptable...impressive power...gearbox sucks.

Rod

 

[NickE]

rod- yep its not the highpoint of the car... Reminds me of my old AMC Eagle.. More like shifts in a truck than sports car... Ahh... someday to afford the new design 6-speed.
(It is however quite manageable if you're patient, and kind.)

 

[kenre]

I think everyone has missed the obvious!

A dog clutch gearbox will only go into gear when Engine rpm and box rpm are syncronised! there cant be a 2000 rpm difference.

There is no slip of Syncro rings trying to match speeds to allow gear engagement.

Ken

 

[patprimmer]

A dog clutch will engage reliably and with durability with an 800 rpm speed difference. This is evident by their extensive use as clutches in inboard motor speed boats.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[daveylynx]

Ken - no, that's absolutely incorrect. A dog gearbox has a quicker change than a syncro box precisely because the speeds don't need to be syncronised! This is the reason why the box is subjected to increased shock loadings. The mismatch in speed as the dog faces meet forces the engine to match the speed of the next gear, hence the shock load.

Just as an example, F1 gear changes take less than 20ms - during this extremely short period of time, the engine will not slow down by any measurable amount of its own accord - even F1 engines have significant inertia, especially when spinning close to 20,000rpm! The engine is literally forced to match the next gear when the dogs engage. Greater rotating inertia and greater speed mis-match meang greater shock loads - just how much greater is the question I asked in the original post.

 

[evelrod]

Ken, full power shifts were being made with modified gearboxes even in the early 60's. We took every other 'tooth' off the gear synchro gears and ground ever other 'spline' out of the synchro hub, deleted the synchro rings (clutches) for a drag racing T-10 ! You never(sometimes I would lift bit, depending on how close the race) lifted your foot from the accelerator pedal and just 'reefed' on the T handle for the next gear...hard on gearboxes? I would imagine so...we really did not care as long as it was quicker. I had a 283 Vette and shifted at 7000 so my guess is there was ~1500/2000 rpm spread---Oddly, I never broke a box or the 4.56 final drive in three years of drag racing. I still remember my best time...12.27 @ 105.57 in 1963. Ahhhhh, the 'good ole days' before I knew how much this hobby/addiction was going to cost.

If you can find the Hotrod Mag article from 59/60 or so from their archives...that's where I got the idea back then.


Rod

 

[kenre]

OK! i was wrong! but for nice smooth shifts speed matching must be needed. must be from working on too many Roadrangers.
For racing, anything goes!

Rod, Impressive times for back then! I have heard of those tricks being used on syncros to speed shifts. Liberty cut syncros rings a bell.

Ken

 

[sreid]

In F1, the ECU controls the clutch for shifts.

http://www.f1technical.net/articles/66

I believe that the engine speed is changed by the ECU cutting the fuel.

 

[KiwiME]

In a motorcycle dog-clutch transmission the drivetrain provides a cushion for the change in speed of the clutch /and engine if reqd. The rubber damper in both the clutch inner hub and the rear wheel hub provide lots of cushion. The dogs also have a lot of slack in rotation to increase the likelihood of a clean engagement. The ratios tend to be close which helps a lot.

Just apply simple kinematics. You need to look at how many degrees engine rotation you can get from the flex you have in all the parts from the clutch disk to the tires, and assume it all acts like a torsion spring. Estimate the engine inertia. Calculate the resultant torque (T = I x angular accel) required to decelerate the engine speed from one gear to the next over that many degress.

 

[toprem]

This comes from some experience with marine sterndrives where we essentially use dog clutch full power shifting. Now the RPM difference is he most damaging to the system. Someone long time back did someting smart. The shifting is done in the lower unit and hence the engine rpm is reduced by the gear ratio. So the speed difference across the dog clutch can be reduced by half (2.0:1 ratio). Best RPM for a relatively low shock shift has been found to be 400-500 rpm. (That is 800 engine RPM) Shift shock is a characteristic of dog clutches. If you sit in the grand stand for the F1 GP you would see drivers head banging back and forth every time he shifts as he accelerated out of the pit lane. Another key aspect we learnt was the speed at which the shift was done. Faster the shift the better it is. A dog clutch shift is comprised of more than one rejection. Everytime there is a rejection between the target gear and the dog clutch, the clutch gains energy to bang in faster. Anything that you can do to have torsional compliance behind the target gear will help reduce the shift shock.

 

[patprimmer]

I have never seen a stern drive where full power shifting is used. To my knowledge, they all only have a forward neutral reverse function with only one forward and one reverse ratio.

There is very little compliance in the drive train between the motor and the clutch as it is a number of relatively short shafts connected via steel bevel gears and steel (Hardy Spicer type) universal joints. There is a lot of compliance after the clutch as the propeller is a very low efficiency (at idle speed) hydraulic coupling.

Forward or reverse is selected at idle. The reduction in the drive ratio certainly helps to a greater degree than the extra mass and surface area of the propeller as the inertia is 1/2 mass times velocity squared.

Regards

eng-tips, by professional engineers for professional engineers
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[Rob45]

Nick:
I doubt that your troublesome WRX gearbox is a dog-type; much more likely it uses synchro rings.

 

[NickE]

Yes it is a sychro transmission. In fact it has old style baulk rings. I was primarily addressing a method for lessening the effects of shock loading.

(And I dont have troubles with my WRX trans (knock wood) after 100k mi. but it isn't the best feeling I've ever driven.)

 

[Pancholin]

There are some very good posts here. I will relate my experience. Triumph motorcycles. One was a scrambler with a wide ratio gearbox, the other was a Triumph engine and a close ratio Triumph gearbox in a Norton frame. It was very difficult to shift the wide ratio without the clutch. But very easy to shift the close ratio one.
The wide ratio box was 2.91 - 2.21 - 1.43 - 1.0
The close ratio box was 1.70 - 1.30 - 1.09 - 1.0
And there were rubber pieces in the clutch to absorb some shock in both clutches. I suppose the springs in automotive clutches would absorb some of the shift shock.
Have a great day.
Pancholin

 

[Fabrico]

Final drive ratio can also be an important factor.

 

[evelrod]

Actually, Fabrico, as it relates to the general theme of this thread, the FD is not nearly as critical as the gearbox ratios (unless...I even have reservations about this... you go to extremes, e.g., ~6.0:1 for 1/8 mile or ~2.0:1 for the lakes) as detailed in the Pancholin post. The closer the ratios, the easier the shift and the lower the shock loading. Many motorcycle and formula gearboxes have no synchros OR dogs...just the straight cut gears...extremely close ratios make the clutchless shifts seamless.

Rod

 

[Fabrico]

This is why I mentioned “can” as we still do not know what it is being used on. Final drive ratios multiply all other ratios and the differences between them. The higher the numerical FD ratio is, the greater or "wider", all transmission gear ratios become. If my math is right...
Using only the close ratio Triumph gear ranges listed above as an example

Close Ratio = 1.70 - 1.30 - 1.09 - 1.0

3.0 FD = overall ratio of 5.1, 3.9, 3.27, 3.0 for a difference (upshifting) of 1.2, .63, .27
5.0 FD = overall ratio of 8.5, 6.5, 5.45, 5.0 for a difference (upshifting) of 2.0, 1.05, .5

These are not way-out ratios, 5.0 provides very nearly double the RPM difference between each gear than 3.0.