[HELP] Worm Gears or Normal Gears

[PKW]

HI to all! the video itself explain all I want to say about the difference between the same ratio obtained with worm gears or with normal gears, the first are really useful in lego and allow to build really particular stuff, but.. they have a lot of friction. i was really surprised from how much difference there is in the power output of the two mechanism, and it is directly proportional to the force the system has to generate: i test them with an !kg weight, and while my version works fine, the worm geared-one fail to lift it!

thank you for watching, hope that this video will help someone of you to make better mocs!

[BusterHaus]

Very good demonstration. The worm gears have only two advantages: size, and being able to lock the follower gear in place.

[Didumos69]

I understand your point. As far as I know worm gears are not just useful for amplifying torque, they are a means to transfer torque in one direction - from worm gear to gear and not the other way arround. Isn't that the actual point of using worm gears?

The video indeed shows that worm gears have more friction. The video would be even more valuable if you could elaborate on why the worm gears have more friction. It must have something to do with the fact that in a worm gear - gear mesh the contact faces of the individual gears move in orthogonal directions, whereas in a normal gear - gear mesh the contact faces of the individual gears move in the same direction.

Edited May 16, 2016 by Didumos69

[bonox]

The primary cause is that unlike a typical involute gear tooth profile which gives a rolling contact between faces, a worm gear operates purely under sliding motion between contact faces and hence lubrication is critical to a high performing worm gear set. This lubrication doesn't exist in lego and gives significant friction. It's also equally likely in lego though to be the lack of a low friction thrust bearing means a significant load ends up driving the worm or bush against the stop (frame/liftarm/technic brick etc) with a lot of friction involved. You'll get the same on any single helical cut gear pair as well.

[Didumos69]

The primary cause is that unlike a typical involute gear tooth profile which gives a rolling contact between faces, a worm gear operates purely under sliding motion between contact faces and hence lubrication is critical to a high performing worm gear set. This lubrication doesn't exist in lego and gives significant friction. It's also equally likely in lego though to be the lack of a low friction thrust bearing means a significant load ends up driving the worm or bush against the stop (frame/liftarm/technic brick etc) with a lot of friction involved. You'll get the same on any single helical cut gear pair as well.

:thumbup: That's a very informative clarification, thank you very much!!!

Besides causing friction, the worm driving against the stop may also dearrange your gear casing, if you didn't anticipate this. Good to know all this.

Edited May 16, 2016 by Didumos69

[Victor Imaginator]

PKW, thanks for making this video. Worth watching, very nice explanation)

[Zerobricks]

Nice showcase of the inherent inefficiency of the worm gears. You loose some 50% of power right there.

[PKW]

Very good demonstration. The worm gears have only two advantages: size, and being able to lock the follower gear in place.

[...] they are a means to transfer torque in one direction - from worm gear to gear and not the other way arround. [...]

[..] a worm gear operates purely under sliding motion between contact faces and hence lubrication is critical to a high performing worm gear set. This lubrication doesn't exist in lego and gives significant friction.[...]

all that you are saying is strictly linked to what bonox said: there is so many friction in lego worm gears that the seems to lock the power transferring, but real worm gears, have a good lubrification and move particular gears like this, with rotated teeth:

so their friction is so reduced that it is also possible to invert the torque transferring: for example hypnodisk on robot wars has a worm gear mechanism to power its wheels and it hasn't got the "handbrake" problem that lego one cause.

The reason that why i haven't talk about the lock on place function is that i'm making this kind of video for a Lego robot wars group, and it demonstrates that also at low gear reduction (5:1 instead of the 40:1 that is possible with a worm gear) the unpowered M motor resistance is enough to power a winch without slipping problems.

IMHO the main goal of worm gears is to save space and to be easy to reinforce, in fact it is difficult to make a tough gear chain without making something really heavy and big, they are also useful while using mindstorm for the lock in place function, but mainly because the give a 1:1 ratio with the motor rotation and the gear teeth moved.

PKW, thanks for making this video. Worth watching, very nice explanation)

Nice showcase of the inherent inefficiency of the worm gears. You loose some 50% of power right there.

Thank you!

[DrJB]

... but real worm gears, have a good lubrication and move particular gears like this, with rotated teeth:

... so their friction is so reduced that it is also possible to invert the torque transferring ...

This could not be farther from the truth. Some worm drives ARE reversible (such as those operating the tray of DC/DVD/BD drive. However, MOST worm drives are NOT reversible.

The simplest driving factor is the relationship between the friction angle (tan mu) and the angle of the helix making the worm gear. For the gear drive to be 'reversible, you need:

Helix angle > mu

If the above relationship is NOT satisfied, then you'll enter a 'lock-up' and no matter how much lubrication you add.

Edited May 16, 2016 by DrJB

[PKW]

This could not farther from the truth. Some worm drives ARE reversible (such as those operating the tray of DC/DVD/BD drive. However, MOST worm drives are NOT reversible.

The simplest driving factor is the relationship between the friction angle (tan mu) and the angle of the helix making the worm gear. For the gear drive to be 'reversible, you need:

Helix angle > mu

If the above relationship is NOT satisfied, then you'll enter a 'lock-up' and no matter how much lubrication you add.

thank you for your clarification! only one thing: the original name of the greek letter is mi, not mu, why everyone spell it wrong? (this is the only thing i learnt from my classical liceum ahahah)

[Didumos69]

The charm of all this is that in the end we've all come to a thorough understanding of the operation of worm gears. I truly think that's a good thing. Cheers!

[bonox]

If the above relationship is NOT satisfied, then you'll enter a 'lock-up' and no matter how much lubrication you add.

not really - any helix angle is bigger than zero. All that remains is to get your lubrication and material properties good enough. Often a significant issue here relates not to choosing low friction pairs but in making one element, most often the gear, sacrificial to protect the one hardest to replace, most expensive etc.

Multi-start worms are typically very easy to back drive.

As a point of interest though, most worm drives are not made purely for efficiency and the oil ends up being used for cooling as much as direct lubrication.

[DrJB]

not really - any helix angle is bigger than zero. All that remains is to get your lubrication and material properties good enough. Often a significant issue here relates not to choosing low friction pairs but in making one element, most often the gear, sacrificial to protect the one hardest to replace, most expensive etc.

Multi-start worms are typically very easy to back drive.

As a point of interest though, most worm drives are not made purely for efficiency and the oil ends up being used for cooling as much as direct lubrication.

I care to differ. I stand by my statement that the helix angle must be larger than the friction angle for the gear set to be reversible? Now, maybe you have some super-fluid with zero (or negative) friction coefficient, in such case, care to enlighten us?

[suffocation]

thank you for your clarification! only one thing: the original name of the greek letter is mi, not mu, why everyone spell it wrong? (this is the only thing i learnt from my classical liceum ahahah)

Off-topic, but educational - in English we refer to the letter as "mu". Phonetic transcriptions may vary across languages - a notion not always taught at liceo classico.

[Didumos69]

not really - any helix angle is bigger than zero. All that remains is to get your lubrication and material properties good enough. Often a significant issue here relates not to choosing low friction pairs but in making one element, most often the gear, sacrificial to protect the one hardest to replace, most expensive etc.

Multi-start worms are typically very easy to back drive.

As a point of interest though, most worm drives are not made purely for efficiency and the oil ends up being used for cooling as much as direct lubrication.

I care to differ. I stand by my statement that the helix angle must be larger than the friction angle for the gear set to be reversible? Now, maybe you have some super-fluid with zero (or negative) friction coefficient, in such case, care to enlighten us?

I found this easy to understand article in this matter, which states that under static conditions, "the friction angle must be larger than the lead angle to prevent back-driving", with the lead angle being the opposite of the helix angle. EDIT: When traversing the internet, the helix angle and lead angle appear to get mixed up quite often, especially in the context of worm gears. I suppose DrJB was referring to the same condition with his statement: "helix angle must be larger than the friction angle for the gear set to be reversible".

The article also states that "It is usually impractical to design irreversible worm gearing with any security". If the statically self-locked worm gear is subject to shock and vibration, the friction coefficient between worm and gear may suddenly drop, causing the friction angle to drop below the lead angle. During the short time the friction angle drops below the lead angle, the gears are no longer self-locking and back-driving can occur. Once started, back-driving usually continues because the friction coefficient decreases once the gears are in motion. This effect has actually caused accidents in cases where lift devices relied on self-locking worm gears to hold the cargo in a desired position.

Now back to LEGO. It is clear that with the lack of lubrication and the amount of friction involved, a LEGO worm gear mesh can be regarded as irreversible. However, as the OP's video showed us, the amount of friction involved in a worm gear mesh is a practical problem. This brings me to the following question: How to build a heavy load lift from LEGO (as in the video) that does not suffer from this progressive friction, but does prevent back-drive? My idea would be to combine both gear arrangements from the video (both with the same ratio, one with worm gear one without) and use them in parallel. The normal gears will take over where the worm gear would suffer from too much friction and the worm gear will prevent back-drive. Would that work? EDIT: Propably not, the normal gears will drive the worm gear.

Edited May 17, 2016 by Didumos69

[PKW]

[...]Now back to LEGO. It is clear that with the lack of lubrication and the amount of friction involved, a LEGO worm gear mesh can be regarded as irreversible. However, as the OP's video showed us, the amount of friction involved in a worm gear mesh is a practical problem. This brings me to the following question: How to build a heavy load lift from LEGO (as in the video) that does not suffer from this progressive friction, but does prevent back-drive? My idea would be to combine both gear arrangements from the video (both with the same ratio, one with worm gear one without) and use them in parallel. The normal gears will take over where the worm gear would suffer from too much friction and the worm gear will prevent back-drive. Would that work? EDIT: Propably not, the normal gears will drive the worm gear.

I think that the best way to make an auto-locking system is to use motors that has a good embowered turning resistance so if they are geared down in the correct weight they won't be back turned from the weight lifted, anyway this means that maybe you can't lift all that the motor can expecting that the locking function will work in any7 case, because the unpowered torque is really lower than the powered one. there could be implemented some auto locking mechanism in the wire, or something like the one used in catapults but they are quite difficult to build in a reversible way.But as I said i think that the space saving of worm gears and the ease of the worm gear mechanism will compensate the lack of outputed power.

[bonox]

I care to differ. I stand by my statement that the helix angle must be larger than the friction angle for the gear set to be reversible? Now, maybe you have some super-fluid with zero (or negative) friction coefficient, in such case, care to enlighten us?

Whilst you might stand by your statement what you said was "ANY amount of lubrication" and I addressed that. You can never reduce the helix (or lead) angle to zero or you won't achieve anything, but tribology improves rapidly. In any case, I don't have to share any stupid negative coefficient with you (what a lunatic idea); we've already got back driving setups - would you care to enlighten us to your knowledge that materials science will not improve at all in future?

[DrJB]

I'm not here to enlighten you or anyone. I have a PhD in mechanics and my statement was not based on 'lego' but more fundamental laws of physics. As for my comment on negative friction coefficient, I was being playful, and if you got offended, time to reconsider what this hobby is and does for you. Yes, we constantly see crazy ideas on this forum (e.g. perpetual machine) and some members have NO background in engineering/physics. Yes, I do come here to share/learn, and I also speak out when I read what you call 'stupid' comments ... *peace*.

[Lipko]

I don't think that your comparison is quite right. The worm gear setup has a quite significant gearing up at the end: 32 vs 8 teeth. That may introduce significantly more friction than the gears in the other setup.

[PKW]

I don't think that your comparison is quite right. The worm gear setup has a quite significant gearing up at the end: 32 vs 8 teeth. That may introduce significantly more friction than the gears in the other setup.

Yes I also thought that, the "no load" (or more accurate "really low load") test makes me see that there is more or less a 4% difference lifting speed with the low weight of the forks(i was expecting the worm gear lifter just a bit faster), so i test the static friction (removing the motors and the worm gears) and the gearwheel-only mechanism has 3-4 more friction than the 5:1 gear ratio in the other lifter (that has 2 gear reduction), so i think it didn't affect the test. anyway my test don't want to take out numbers but only to show that the difference between worm gears and gearwheels is noticeable and high because not everyone knows this (while i came out with a 50% loss with more or less 0.5 kg mass other with different mechanisms found a 40% loss, but they used a worm gear and a gearwheel vs a big gear chain).