Disengaging and re-engaging gears

[wisam96]

Is it possible to disengage two moving gears then engage a moving gear with a non-moving gear?
If so approximately how much energy loss will there be in doing so?

 

[Merlin3189]

Well, it is possible, but not generally a good idea.
You can easily disengage two gears, but there will be a jolt when you try to engage a moving gear with a stationary one. If the load is light and there is some elasticity in the chain, it could work. If the load is high, you are likely to damage the gears.
Normally gear changing with live systems uses a clutch, maybe a sprung coupling and in cars for example synchromesh cones - all trying to reduce the damage caused by crash gear changes.

The nearest I can think of to such a system, is a derailleur gear on bikes. There the chain engages with a gear moving relatively slower or faster rather than stationary. The difference in speed is not very great and good technique is to reduce the load during the change. I would guess this is your best option if you must have a live gear change without a clutch.

As far as energy loss is concerned, I think that would be very difficult to quantify. More to the point is, not how much energy is lost, but where that energy goes! Some will go into noise and heat, but some will probably go into bending and possibly breaking the mechanism. If there is no damage, I would expect the energy loss to be small (but then I do expect damage!)
If you use a clutch, the main energy loss is due to slip and would be quantifiable if you know the duration and torque profile of re-engagement. Essentially the slip loss power is torque x slip speed.

 

[wisam96]

Well, it is possible, but not generally a good idea.
You can easily disengage two gears, but there will be a jolt when you try to engage a moving gear with a stationary one. If the load is light and there is some elasticity in the chain, it could work. If the load is high, you are likely to damage the gears.
Normally gear changing with live systems uses a clutch, maybe a sprung coupling and in cars for example synchromesh cones - all trying to reduce the damage caused by crash gear changes.

The nearest I can think of to such a system, is a derailleur gear on bikes. There the chain engages with a gear moving relatively slower or faster rather than stationary. The difference in speed is not very great and good technique is to reduce the load during the change. I would guess this is your best option if you must have a live gear change without a clutch.

As far as energy loss is concerned, I think that would be very difficult to quantify. More to the point is, not how much energy is lost, but where that energy goes! Some will go into noise and heat, but some will probably go into bending and possibly breaking the mechanism. If there is no damage, I would expect the energy loss to be small (but then I do expect damage!)
If you use a clutch, the main energy loss is due to slip and would be quantifiable if you know the duration and torque profile of re-engagement. Essentially the slip loss power is torque x slip speed.

Do you think making a mechanical timer to shift the axle would be better, or should I use a microcontroller and a servo motor to shift the gears.
I'm not familiar with derailleur gear, I'll do some research on it, but I doubt it will help me in my project

 

[gmax137]

Search for "constant mesh transmission." Like motorcycles use, the gears are always engaged. Shifting involves sliding the gears, only the dogs mesh and unmesh. The dogs are bigger and sturdier than gear teeth. You can shift w/o the clutch by backing off the throttle a bit. Of course that won't work from a standing start, need the clutch for that.

 

[Ketch22]

Yes, a dog type constant mesh transmission would work well. Also consider that a synchromesh ring such as used in a standard transmission could easily be adapted to most applications. That would bring the accelerate the nonmoving gear to match the moving gear You would not be able to transmit much power while the synchro rings were slipping but the actual meshing process could be fairly quick.

 

[wisam96]

Yes, a dog type constant mesh transmission would work well. Also consider that a synchromesh ring such as used in a standard transmission could easily be adapted to most applications. That would bring the accelerate the nonmoving gear to match the moving gear You would not be able to transmit much power while the synchro rings were slipping but the actual meshing process could be fairly quick.

Is the constant mesh transmission a simple mechanism that I could use in a small project like mine?
I'm building an autonomous vehicle with regenerative braking, i basically want to disengage a flywheel for a second and then re engage it with a different gear.

 

[Tom.G]

The gears aren't the only thing to worry about, consider the stresses on the flywheel. An additional possibility is the approach used in automatic transmissions. Internally they use planetary gearing and prony brakes to select the ratios. And all the gears can be constantly meshed.

 

[Rx7man]

Inertia is the problem.. one of the two sides needs to have low inertia.. on bicycle deraileur, one side has a lot, which is the wheel side because of the weight of the rider and bicycle, the other side has very little (the weight of the rider's legs).. furthermore, there's a lot of elasticity in the rider's muscles, and each gear change is only a very small ratio (10%?).. that makes for smooth gear changes despite not having any clutches.

With the description of what you're trying to do, I think TomG has a good idea.. leave the gears meshed, use planetaries and brakes to smoothly apply the power...

Why not just use a clutch? How big is this vehicle? Does regenerative braking actually 'pay' for the complexity and extra weight it will add? If the vehicle is electrically powered, it might be simpler to regenerate electrically than mechanically.

If using a clutch, you could estimate the loss of power by integrating the product of force x distance exerted on the clutch plates.. when just engaging, F will be small, D will be large, as you continue engaging, F will increase, but since both parts of the clutch will start to move together, D will become smalll.
Estimating energy loss from a sudden engagement will be very difficult... there will be a certain flex in the driveline and tires to absorb shock, but the tires will probably spin a bit (F x D again), etc