high level road runner gearbox

[Ian Austin]

Have built my first gearbox and it runs well under power. My problem is that now the motor/gearbox combination is fitted to the driving axle the gear box folds around the idler shaft and consequently stops turning. If I lift it with my fingers it runs again. What techniques do you use to overcome this?

[Tim V]

Sounds like you're referring to "torque reaction" which has been covered in a number of posts on this forum.

There are various different solutions, I won't single out one (I think there's one in my workbench thread, other solutions are available).

[Will L]

The issue here is the lower separate segment of the Gearbox which is free to pivots around one of the lay shafts. Several of the Highlevel gearboxes are like this, and it is very useful as it allows you can configure the gearbox to suit your chassis. But you must to fix the relationship between motor, gearbox and driven axle before you drive the loco through it.

What you do about this will depend on the installation in the loco. If its on a fixed axle, then so long as you mount the motor rigidly to the chassis you'll be OK. If its on compensated or sprung axle, you will need to fix the two parts of the Gear box together in a configuration that suits your loco, You can do that with glue or solder or in more complex ways.

I've posted about doing this before, in this post I'm doing things here which are more complicated than you need but it does cover your query too if you skip down to the end.

Once that is sorted out then it is advisable to add a torque reaction link as well. What that is about is explained here

Will

[billbedford]

Have built my first gearbox and it runs well under power. My problem is that now the motor/gearbox combination is fitted to the driving axle the gear box folds around the idler shaft and consequently stops turning. If I lift it with my fingers it runs again. What techniques do you use to overcome this?

Fix the motor in the frames. Ideally the the last layshaft should be in the same horizontal plane as the driven axle.

[Will L]

I take it that's assuming it is attached to an axle free to move up and down vertically. This is a logical possibility but I didn't cover it because
1. As axle slot/horn blocks will only allow the axle to move strait up and down, and the gearbox will only allow it move in an ark around the lay shaft, the amount of movement available before one constrains the other will be small, possibly within the bounds of the normal expectations of a suspension system, but only so long as the motor is fixed in exactly the right place.
2. More critically, while the axle may be able to go up and down, the gearbox will keep this movement the same on both sides, i.e. the axle will not be able to rock so wheel can't be higher one side than the other. So you springing/compensation won't work properly.

Will

[Russ Elliott]

Will is correct, as is his 'thinking sequence'. Let me try to put some figures on the situation, taking a graphic from another thread:



Take a sprung 6-coupled 200g tank loco producing say 40g (0.4N) maximum on its drawbar. That requires 0.4N over the 3 axles at the railhead. Say each axle contributes 0.13N. The torque required on a 20mm diameter wheel axle to produce 0.13N at the two tyres of that axle is 0.13N x 10mm = 1.3Nmm. The distance between the final shafts of the pivoting segment of the High Level Roadrunner is about 10mm, and the gears on those shafts have the same number of teeth, so we can assume the motor torque at the gearbox's final fixed shaft is also in the 1.3Nmm maximum region. If there is a vertical 0.7N (say) on the axle, and the gear torque can alter this by 0.13N, then the resulting vertical force on the axle will vary from 0.57N to 0.83N.

This admittedly is a limiting case of maximum drawbar pull, and the above change to an axleload force is above what the axle could exert anyway in normal circumstances (statically a typical 200g/3 = approx 0.7N), so if the Roadrunner's pivoting output stage is left 'unfixed', such a gearbox torque will either lift the driven axle (against the chassis spring) until it has no effective downforce (and hence rail grip) on it, or lift the chassis until the arrangement of motor to output stage to chassis reaches some kind of mechanical limit. (Both cases are to an extent complicated or changed by which axle is being driven, the orientation of the pivoting output stage, and of course the existence of the other two coupled axles, but hopefully you get the drift.)

The upshot is that our motors are capable of producing a level of torque, and can develop a rate of power input to the system, that has to be reacted against to the chassis, otherwise the motor, or an 'unfixed' output stage of a Roadrunner, will simply try to rotate itself with respect to the chassis against the resistance of the loco (which hopefully is small) and its train (which can be large). If there is no train to pull, or the train is short, say 10g max drawbar, the above sums become more engaging, because the motor torque is now down to 0.4Nmm, and thus its effect on a 1.4N/mm axle becomes far more debatable in the context of our '0.5mm' vertical axle deflections. Nevertheless, even at that lower torque level, control of motor and gearbox torque reaction remains good practice IMHO.

[David Thorpe]

I don't understand the maths, but recognise the desirability in a CSB system of ensuring that the gearbox has some limited scope for both horizontal and vertical movement. I wonder if the following would work:
Decide the position of your gearbox in the chassis.
Take a rod and tube, the latter a bit bigger than the rod.
Put the rod through the tube and fix it between the chassis frames in a position where it will come up against the body (rear) of the gearbox.
Fix the rear of the gearbox to the tube. The tube, and therefore gearbox, will be constrained by the fixed rod but, the tube being wider than the rod, will still have limited all-round movement.

Just an idea - might try it on my next loco.

DT

[billbedford]

I take it that's assuming it is attached to an axle free to move up and down vertically. This is a logical possibility but I didn't cover it because
1. As axle slot/horn blocks will only allow the axle to move strait up and down, and the gearbox will only allow it move in an ark around the lay shaft, the amount of movement available before one constrains the other will be small, possibly within the bounds of the normal expectations of a suspension system, but only so long as the motor is fixed in exactly the right place.

Would you like to show us your trigonometry for this?

2. More critically, while the axle may be able to go up and down, the gearbox will keep this movement the same on both sides, i.e. the axle will not be able to rock so wheel can't be higher one side than the other. So you springing/compensation won't work properly.

If that really worries you then leave off the link etching and just mesh the final gears. Providing the gear are not mega fine then there will be enough tolerance to accommodate the small amount of twist in the axle.

[Andy W]

"Fix the rear of the gearbox to the tube." Good idea Davey, and I'm sure it would work, however, one of the major pluses of CSBs is the ability to drop the wheels and gears out of the chassis - and this would negate this.

I find all the theory and maths very interesting, but I've found, in the small Victorian locos that I tend to build, that the motor is such a tight fit inside the boiler and against the cab rear that there's not much room for the box to move anyway. The choice of the right gear box and its layout is the vital thing, and the last segment of the gearbox should be fixed to the rest of the gearbox frames.

[billbedford]

This admittedly is a limiting case of maximum drawbar pull, and the above change to an axleload force is clearly way above what the axle could exert anyway in normal circumstances (statically a typical 200g/3 = approx 0.7N), so if the Roadrunner's pivoting output stage is left 'unfixed', such a gearbox torque will either lift the driven axle (against the chassis spring) until it has no effective downforce (and hence rail grip) on it, or lift the chassis until the arrangement of motor to output stage to chassis reaches some kind of mechanical limit. (Both cases are to an extent complicated or changed by which axle is being driven, the orientation of the pivoting output stage, and of course the existence of the other two coupled axles, but hopefully you get the drift.)

This is complete bo**ox.

No axle will lift. Gravity will ensure that all axles will stay firmly on the rails.
Since there will be an equal and opposite reaction between the final drive gears not only will there be a change in the load on the driven axle, there will a equal and opposite change in load on the frames. The result of this is that the total weight on the rails will not change. The CSB spring will transfer any change of load on the driven axle to the undriven axles. This will result in a displacement, with probably a vertical and rotational component, of the body on the springs. Because the change in load due to the motor input will be small compared with the total weight of the model, the displacement will be imperceptible in most situation.

Going back to the OP. You have a choice of either removing the final swing link and fixing the motor in the frames or fixing the swing link to the gearbox body and using a torsion link to hold the motor. The former has the advantage that the motor and gearbox need not be fitted to the loco until the frames are known to run sweetly, and the disadvantage that some care will be needed in the fixing of the motor/gearbox to ensure a satisfactory mesh on the final drive.

[Russ Elliott]

The theoretical axle load variations explored above are not as much as I had expected, but I find it difficult to generalise the effects of gear torque on axle loads because the number of tractive axles involved varies between locos, the train loads make a big difference, and the torque 'capacity' (is that the right word?) of motors does vary a lot - in respect of wanting to 'wrap itself around an axle' under load, what a chunky 1624 in a Portescap might want to do is very different to what a 6mm diameter micromotor might be capable of doing. For example, the stall torque of a Faulhaber 1624 is about 4mNm (4Nmm), so could exert a theoretical 140Nmm through its Portescap 35:1 gears.

[Will L]

I take it that's assuming it is attached to an axle free to move up and down vertically. This is a logical possibility but I didn't cover it because
1. As axle slot/horn blocks will only allow the axle to move strait up and down, and the gearbox will only allow it move in an ark around the lay shaft, the amount of movement available before one constrains the other will be small, possibly within the bounds of the normal expectations of a suspension system, but only so long as the motor is fixed in exactly the right place.

Would you like to show us your trigonometry for this?

Not sure about trigonometry but a picture should do it. The highlevel gearbox illustrated isn't the one that started this thread but is similar enough not to make any difference. We are assuming the primary gearbox frame and motor are rigidly attached to the frames, and the axle is free to move up and down.

bill rep 1.jpg

In a perfect world the axle will not be free to move up and down at all, as to do so it must also move sideways in the horizontal plain. In practice the various clearances between axle/hornblock/gearbox bearings etc will allow the axle a limited amount of vertical movement, possibly even enough to allow the 0.5mm either side of centre we want for our a working suspension. While it must be possible to work out how much clearance between the various moving parts is necessary to achieve this, possibly using trigonometry, I'm not going there because it isn't my primary objection.

2. More critically, while the axle may be able to go up and down, the gearbox will keep this movement the same on both sides, i.e. the axle will not be able to rock so wheel can't be higher one side than the other. So you springing/compensation won't work properly.

If that really worries you then leave off the link etching and just mesh the final gears. Providing the gear are not mega fine then there will be enough tolerance to accommodate the small amount of twist in the axle.

You mean like this.

bill rep 2.jpg

If I've read Russ a-right (I've had several goes and I think I understand now) if you attempted this and because there is now nothing to restrain the axle from moving away from the gearbox, the reaction between the gears will either force the axle up or down, and, even if there isn't enough force around to allow this single effect to force the gears out of mesh, all those little clearances which allowed us to even consider the first case will now ensure un-meshing does happen.

Will

[Russ Elliott]

This is complete bo**ox.

No axle will lift. Gravity will ensure that all axles will stay firmly on the rails.
Since there will be an equal and opposite reaction between the final drive gears not only will there be a change in the load on the driven axle, there will a equal and opposite change in load on the frames. The result of this is that the total weight on the rails will not change. The CSB spring will transfer any change of load on the driven axle to the undriven axles. This will result in a displacement, with probably a vertical and rotational component, of the body on the springs. Because the change in load due to the motor input will be small compared with the total weight of the model, the displacement will be imperceptible in most situation.

I probably didn't express myself clearly. (I will rewrite the offending paragraph.) What I meant was the axle will lift in respect to the other axles. I agree the total weight cannot change. I agree the change in load from the effects of the motor should be 'small', but we come back to the situation reported all those years ago which started the whole motor torque reaction debate, as exemplified by Ted: "...Portescap fitted 7F. It lifted the driven (rear) axle under load in reverse, because the torque reaction of the motor bore down on the bottom of the boiler, 'jacking' the axle upwards." Or, as MarkS reports: "The problem I am having is that when power starts to be applied, the loco will rise (or lower when direction is reversed) on the rear (powered) axle."

The theory indicates we should be in safe-ish territory, but practice seems to indicate otherwise.

[billbedford]

Not sure about trigonometry but a picture should do it. The highlevel gearbox illustrated isn't the one that started this thread but is similar enough not to make any difference. We are assuming the primary gearbox frame and motor are rigidly attached to the frames, and the axle is free to move up and down.

This is one of those cases where 'common sense' tells you one thing, but only working out the actual dimensions will tell you if you are in or out of tolerance.

If I've read Russ a-right (I've had several goes and I think I understand now) if you attempted this and because there is now nothing to restrain the axle from moving away from the gearbox, the reaction between the gears will either force the axle up or down, and, even if there isn't enough force around to allow this single effect to force the gears out of mesh, all those little clearances which allowed us to even consider the first case will now ensure un-meshing does happen.

Again 'common sense' gives you the wrong answer. There are two limiting cases here. If the axle is constrained not to rotate, the it would indeed move up or down in the guides, however without that constraint the axle would rotate. Since we want our axles to rotate, the question becomes 'under what circumstances does the vertical reaction become visibly significant?'

[Andrew Ullyott]

I use High Level boxes and CSB's. The solution I've settled on is to drill two holes in the sides of the gear box adjacent to the plate which faces the motor. A U shaped piece of 0.45 wire is fitted, the tails of which are fitted through holes in an adjacent frame spacer and then soldered.
The tails are as near to verticle as possible which means that I usually have to fit the spacer to suit.
I solder the wire to the gearbox where it passes through the hole with the tiniest amount of solder.

In this way the final gear drive can rotate, the motor is effectively fixed in the frames but the wire allows just enough lateral movement. I've use it in 0-4-4, 2-8-0, 0-6-0 configurations.
If you need to drop the motor out you just cut the wires.
Don't know about the theory but it works for me. I'll post a picture later if I can find one.

[Will L]

Again 'common sense' gives you the wrong answer. There are two limiting cases here. If the axle is constrained not to rotate, the it would indeed move up or down in the guides, however without that constraint the axle would rotate. Since we want our axles to rotate, the question becomes 'under what circumstances does the vertical reaction become visibly significant?'

I'm not convinced I do have a wrong answer but I do agree your summary. In the best traditions of CLAG "go on then". I look forward to seeing a loco with coupled wheels, and a sprung driven axle meshed to a drive gear fixed rigidly to the chassis.

Will

[Ian Austin]

Hi all,
Thanks for incredible amount of information. Lots to mull over. Shall digest and progress. Thanks again
Ian Austin