Hi guys, I have designed a backlash-free geared system for an application where I don’t need but max a turn of the driven gears and I believe it can be used in light-CNC type machines too so I am posting it here so maybe you can add your opinion on it or suggestions on how to make it better.
I am not sure this exist in another form but in my searches I haven’t found anything similar (maybe for a reason).
I will be using a metal strip/tape to provide tension in the system so at all times the motor gear will be engaged with the driven gears like in the pic below. When rotating one direction there will be no backlash because the motor gear doesn’t loose contact with the driven gear and the other driven gear will be rotated by the cable.
The metal strip will carry a quite heavy platform so my idea is that once tightened so all parts make contact the way showed in the picture, the system will not loosen unless the torque/load/acceleration combo exceeds the plastic deformation threshold of the metal strip.
I am wondering if there is any way to provide the same operation but with no limitation on the number of rotations that the driven gears can turn in one way, so the payload can travel longer distances. I am planning to use double helix herringbone gears instead the spur gears shown below but those were more difficult to sketch.
Now: For 3D printers this could be adapted so that the cable is Spectra and the circumference of the driven gears is longer than the travel of the effector along one side of the printing bed. So if the bed is 250x250 for example, the diameter of the driven wheels should be ~80mm in diameter and one would only have to calculate then from there to get the diameter of the motor gear to get the correct number of steps necessary for the desired resolution.
What do you think?
PS: can’t wait to have the MK2S arrive so I can start printing… so many projects left in the dark…
Edit: this will also avoid the string walking that happens with the usual spools that are mounted on the motor shaft -because it is essentially one loop at most.
@P_Grimm Thanks. English is not my native language but as far as I understand, “hey/hi guys” is gender-neutral and is customarily seen just like “hey friends!”.
@Florian_Ford if you using a looped belt or ribbon around both big gears – the upper side will be stressed and the lower part would be loose … so you always have tension on the teeth and could move indefinitely. But keep in mind that the tension always need to be bigger than the force you apply to move it - else you get backlash.
There’s a similar-looking approach with rack-and-pinion drives where the small driven gear is mounted on a spring pivot or slotted mount, so it moves tangentially relative to the two big gears meshing with the rack.
@Jonathan_Buford There are several ways I could have solved this payload transporting application, including split-gear spring-loaded gears as you correctly suggest. Frankly I would have opted for hyperbolic worms or split worm gears but all seem a bit too complicated for my actual application. I am also concerned that split gears (spur or worms) settle somewhere on a middle position when changing directions, only reducing backlash to about half the way I understand it.
With any of this, it really depends on the final application. As it was pointed out above, there isn’t a perfect solution, only things that reduce the overall tolerances. With the tradeoff usually being increased losses due to friction within a system. The only way to eliminate backlash would be to have a feedback system that directly tracks the moved object and compensates as appropriate or to just directly drive something. Even systems with only belts and pulleys will have elastic backlash or possibly deformation of the belts at some level. Your general concept will work, but with the spring loading, you will either have high force against the drive pinion (and friction) or with lower spring load, you have to watch your acceleration to not overload the spring and stretch it more.
It all comes down to what tolerances and loading you need in your system and designing to suit that.
@Jonathan_Buford Totally agree with you. For my application the concept is sound. I am drooling though after a cheap/achievable feedback mechanism but I am either not qualified to consider such thing or it is more than I would pay for.
The solution I have designed uses cheap (3D printed?) spur/herringbone gears and the mechanism to achieve backlash-less operation is the same as the one delivering the function (moving a payload/effector) …
I guess nothing is quite new under the sun. Look what I’ve found (at least I avoided infringing on the patent by not using the torsional spring hehe): https://www.google.us/patents/US20070034034
That is a pretty poor patent. Actually, your design would probably infringe on Claim 1 of that, but they might have needed to not do a more generalized case if they had prior art that is more similar to your solution. You might look at the references to see what the prior art is.
Like you said, for basic concepts like this, someone has probably thought of it before and might have bothered to patent it.
@Jonathan_Buford What a pita this patenting is … I don’t plan to commercialize this nor to patent it but I never understood why would anyone grant patents for more than 2years for the really meriting ones … 2years is plenty for that concept to evolve anyways …
This patent I have found is very general and my opinion is that you can’t really patent obvious stuff … what I did is quite general and shouldn’t be allowed to be patented in the first place.
