Measuring Power Loss in Gear Trains

[Hrafn]

Has anyone attempted to measure the power loss from gearing motors up or down? What I'd really like to find is something like Philo's excellent work on motor specifications, batteries, tire traction, and so forth. Ideally it would compare all common pairs of gears (8/8, 16/16, 12/20, 8/24, 8/40, etc.) for at least a couple of motors.

[DrJB]

First, let us define Power ...

Power = Force × Velocity = Torque × Angular Velocity

Thus, ideally, you'd want a couple of mated gears (same number of teeth, spinning at the same RPM, then you want to measure the torque supplied to one, vs. the torque available on the other gear. What is difficult here is to estimate 'parasitic' losses'. For example, there will be friction between the axles holding those gears and whatever holes they go through. I bet the losses from friction there are higher than losses due to gears meshing. After all, gear meshing is 'almost' a pure rolling between the teeth involute surfaces. The other complication is that such power loss most likely will depend upon the actual RPM of the teeth and torque applied. Any Gears expert on the forum?

[Zerobricks]

The higher the speed the more friction, more power lost. Also bigger gears are more efficient on same axles than small ones due to axial forces.

[Hrafn]

What is difficult here is to estimate 'parasitic' losses'. For example, there will be friction between the axles holding those gears and whatever holes they go through. I bet the losses from friction there are higher than losses due to gears meshing.

That's a good point. Intuitively I would think the same, but it would be nice to have some experimental data to check intuition against.

Also bigger gears are more efficient on same axles than small ones due to axial forces.

Zblj, could you expand on this?

I would also expect larger gears to be more efficient than smaller ones because of reduced backlash.

I'm also curious to know if power losses depend on whether the gears being meshed are spur to spur, bevel to bevel, or spur to bevel. The bevel gears seem to be made of a softer material. I think the profile of their teeth is essentially the same as those of the spur gears, though.

[Zerobricks]

A small gear has the axle much closer the teeth than a big gear, meaning the force created by same torque is much higher on the axles than by big gears.

[Lipko]

It's hard to come up with actual data, because everything depends on how you build the geartrain. Where, how you bear the axles, how flexible is the frame, etc. In many cases, even the rotation direction can make a huge difference. It's possible to use the same gears in the same sequence (and paying attention for not tightening stuff) and achieve very different power loss.

As for big vs. small wheels: Teeth are not straight, so whenever they transmit force, the for will have a radial component (the gears want to push apart each other). Smaller gears mean smaller radius. You know, that the torque is tangential_force x radius. So smaller gear needs greater force to transmit the same torque. Greater tangential force means greater radial force. Greater radial force means greater friction at the nearby bearing hole.

Sorry for my early-morning English.

Edited January 30, 2014 by Lipko

[andythenorth]

I'm happy to be shown wrong, but lubrication (or lack of) is likely the most significant factor in gear-train power loss.

As an example, 8043 excavator was very non-performant for me, despite fresh batteries and the correct LAs. Applying olive oil to all gears and most axle holes resulted in a much slicker performance and more fun toy.

[JGW3000]

As an example, 8043 excavator was very non-performant for me, despite fresh batteries and the correct LAs. Applying olive oil to all gears and most axle holes resulted in a much slicker performance and more fun toy.

Great tip - but wondering if you need to use 'extra-virgin' olive oil or just regular olive oil ? Any non-food oils better, what about the old standby WD40?

[DrJB]

I'm happy to be shown wrong, but lubrication (or lack of) is likely the most significant factor in gear-train power loss.

As an example, 8043 excavator was very non-performant for me, despite fresh batteries and the correct LAs. Applying olive oil to all gears and most axle holes resulted in a much slicker performance and more fun toy.

Olive Oil works wonders ... I use it in all my GBCs and it makes for smooth and QUIET operation.

Great tip - but wondering if you need to use 'extra-virgin' olive oil or just regular olive oil ? Any non-food oils better, what about the old standby WD40?

Not sure I'd try WD40 as it's a 'penetrating' oil with some synthetic additives ... unless I'm willing to sacrifice lego parts...

[JGW3000]

That is what I was wondering, what synthetic lubricants are compatible with ABS plastic?

[Hrafn]

It's hard to come up with actual data, because everything depends on how you build the geartrain. Where, how you bear the axles, how flexible is the frame, etc. In many cases, even the rotation direction can make a huge difference. It's possible to use the same gears in the same sequence (and paying attention for not tightening stuff) and achieve very different power loss.

As for big vs. small wheels: Teeth are not straight, so whenever they transmit force, the for will have a radial component (the gears want to push apart each other). Smaller gears mean smaller radius. You know, that the torque is tangential_force x radius. So smaller gear needs greater force to transmit the same torque. Greater tangential force means greater radial force. Greater radial force means greater friction at the nearby bearing hole.

Sorry for my early-morning English.

Your early-morning English is clearer than my pre-coffee English, and it's my native language, so no apologies necessary!

I thought it might be too complicated to really measure, but it's good to hear that from someone who's been building with Technic for a while. Your explanation of the axial and tangential forces makes sense, thank you (or köszönöm, as my mother would say).

[aol000xw]

Silicon oil/grease work best with ABS if I remember right.

[allanp]

Well I can tell you the power loss of the 42009 outriggers going up and down.....all of it . Worm gears and lead screws are the most inefficient mechanisms made by Lego IMHO, closely followed by their current gearbox designs.

[torso]

I like measuring, as it makes it easier to develop a feeling for how things work. So, I hooked up a couple of gears to a medium motor and measured the speed with a bike tachometer. It was far from an ideal setup, and someone with more knowledge about mechanical things might point out some errors I'm not even aware of. (I'm a software guy after all )

The result is that I couldn't measure any difference between the gear combinations. Small changes in how well the pieces were put together did make a measurable difference though.

[Hrafn]

The result is that I couldn't measure any difference between the gear combinations. Small changes in how well the pieces were put together did make a measurable difference though.

Thanks for doing that! Lacking a tachometer myself I couldn't do the test here. Do you remember roughly what the percentage losses were when the parts weren't put together well? Was it a question of misalignment, rubbing, flexibility in the support structure, or what?

[torso]

Do you remember roughly what the percentage losses were when the parts weren't put together well? Was it a question of misalignment, rubbing, flexibility in the support structure, or what?

It was a studded build and the problem was that the axle coming out of the motor that got pinched. Just pressing down on the structure made a difference that only just big enough to measure measure. (Numbers come out at 0.5%, but I doubt that it is especially accurate.)

The time when the misalignment was bad enough to be audible the loss was about 10%. I didn't actually notice it until I had written the numbers down and saw that it didn't make sense. With more experience with Lego motors I would probably have noticed it right away.

[imajor]

Not sure I'd try WD40 as it's a 'penetrating' oil with some synthetic additives ... unless I'm willing to sacrifice lego parts...

Everyone says that about WD40, and I have no doubt that it is true. But I used WD40 on my 8043 (because I could not find silicon oil anywhere), and after I disasembled it, I saw no visible damage on the parts.

[Lipko]

I'm not sure that measuring the revolutions is a good way for determining the torque. It depends on the torque vs. revolutions characteristic of the motor. And did you put any real load on the geartrain? Like, driving a car, where there is quite a big radial force on at least one axle? Or did it just run freely?

Maybe one could try it with winches too. Check how much weight on the input winch can move a certain weight on the output winch. EDIT: too much influence by the winches.

Edited January 31, 2014 by Lipko

[torso]

I'm not sure that measuring the revolutions is a good way for determining the torque. It depends on the torque vs. revolutions characteristic of the motor. And did you put any real load on the geartrain? Like, driving a car, where there is quite a big radial force on at least one axle? Or did it just run freely?

The gears ran freely, as I thought there were enough uncontrolled variables involved already. This was really just a quick test to see if there would be any obvious difference or some unexpected phenomenon.

At least it made me stop worrying about gear combinations (worm gear still untested), and devote my attention to gear bracing instead.

[Boxerlego]

It is all about proportion and the worm gear proportions is x:1. Having a x:1 ratio means the every revolution of the worm gear/drive shaft is turning exactly one gear tooth so inorder to turn a 24 tooth gear around one time would take 24 revolutions of the worm gear/drive shaft.

Since the question is based around power loss. I will begin by developing a hypothesis. I'm going to say there is no power loss through the gear train and everything is 100% efficient and the power that is put in is going to be the same amount of power that comes out. Now it is time to build a gear train and develop some kind of test to measure the actual power output. Just by calculating the hypothetical output and the actual output then it will be possible to find the difference in power loss.

[MrNumbskull13]

This topic is quite interesting! Would ABS lubrication be best suited for technic?