Next step on the journey on the worm lifter. We did a lifter module to integrate with our frame system some adjustments were necessary but its moving along. Slowly due to all kinds of school things like parent teacher conferences etc.
What I don’t get is why you need one to lift worms? They’re all wriggly, sure, but usually small enough to pick up by hand. Maybe wear gloves? Heck, in Georgia you can even get them canned, I think: I keep hearing folks saying they’re about to open a can of worms. (Me, I prefer them fresh).
No, seriously: A great inspiration to other teams showing what one can do with some effort. 1989 is lucky to have you!
If everything goes as planned it should be “live” at the Brunswick eruption on Nov. 6th. If not we go with the “old” climber from the 2020 bot. Problem with that one is it does not always latch and stay up after the power is off. Hence a worm gear design as worms dont backdrive
Can you enunciate on your statement of “worm gears -do- back drive if their ratio is low”? I am a student and am interested in hearing how this is possible and why. Are you referring to a situation in which one of the gear’s teeth break along the pitch diameter and are no longer transmitting power to one another? Is this related to the lead angle?
Worm gears always backdrive, but by how much in what time scale is the key. In the majority of applications with a single-start worm, the pitch angles are close enough to right angles and geartrain efficiency is low enough that you can wave your hand at the remainder and say it doesn’t backdrive and for a minute or hour you’re plenty correct enough. Plenty correct enough to collect hang points 5 seconds after the match is over, for example.
What if you want it to backdrive? A nice greasy 4-start worm will happily backdrive, GRT 192’s Joey Milia proved it in about 2012 or so with a worm drivetrain gearbox that packed the CIMs into the drive rail footprint for maximum bellypan space, and still could get pushed around by hand. (AndyMark sells a version of it as the RAW Box - AndyMark, Inc)
The other experience I’ve had with backdriving is when we didn’t think our mechanism would backdrive, couldn’t measure any backdrive in hours-long tests under constant load in the lab, and then left it unpowered buffeted by the wind each night for several years. Turns out, a tenth of a degree here, a hundredth of a degree there, from wind buffeting “walking” the mechanism, will add up! Fortunately you don’t have to design for those kinds of timescales in FRC.
Well let me be more specific then . this one does not back drive in 5 seconds on bench tests with a CIM connected as a load to any significance that might make us not get the points.
As to how well will it work or reliably climb at all - well I withhold judgement on that until it actually climbed successfully with a robot connected etc etc under real competition conditions. It has happened before that something looked really good on paper and on the bench and then miserably failed under real world conditions. So something I often say “Looks great - fantastic - but will it work on the bot as intended”. So all I can say for now is this design made it to the point where we feel its worth actually putting it on a bot and trying it. We got a cycloidal in the works that does backdrive for sure and is more efficient by a big margin but there are still some engineering things to work out and that one has not made it to the bench test yet. Goal is to have a lifter that does and one that does not back drive under FRC style conditions and one that might even survive a buddy climb. Now on paper this design will lift 150 lb give or take at around 30-40 Amps depending on a yet to be determined efficiency with a CIM as we do not have the funds to go NEO yet. and the robot is to have 2 - one on the left and one on the right. to help balance the beam by operating them independent and shifting the center of G of the dangling robot. Well we will see. Biggest concern RN is to get everything done in the time we can work on the robot
Now the cad will be available on the grabcad and if someone is interested I can post the work in progress now - just ask
Not as far as I can tell. There is always a little shrinkage and it might be as we are tuned in pretty well. So it is quite possible that the “tuning” is covering up some shrinkage. But then Our tolerances are probably pretty low. On a gear system shrinkage will give you some backlash and I am not claiming we are without but its considerably less than a toughbox mini - not that this is a difficult yardstick to achieve. On a 12 in piece we come in +/- < 1mm so that is close enough for me for FRC things. Most items are “printed solid” without the need of infill. So for example with a .8 nozzle and .88 outside perimeters, 1.2-1.6mm inside perimeters and 5 to 20 perimeters depending on the part gives a solid part. Here is an inside look at one of our parts (and a hotend mishap)
So it might also be that most of our parts are “grid work” as you can see on the pics above and that might help. Which then gets reinforced with 1/2 in square al. 16 gauge tubing which is still light and strong. we also found that “going solid” brings the strength in the layerd direction (z) close to 80% of “laying down” (x,y) And maybe its the material we use. Mostly HIPS and ABS and some Nylon for some gears and the occasional PETG. As for accuracy ?
But then if everything is printed and everything shrinks - but at the same ratio then you are still on and for “interfacing” with other items… A smooth, tight fit for a 608 bearing (22mm OD nominal) is a 22.6mm or 22.8 loose. etc so .6 to .4 bigger on a hole. The 1/2 in tubing we get measuers between 12.6 and 13.1 mm so a 14mm square works quite nicely to get it through with a little tapping to overcome the little sag on the bridges (top of the square). On a whole robot frame - this years is 29 5/8 x 29 5/8 nominal we are +/- 1mm which is close enough. So probably not good enough to build a space craft but well withing acceptable tolerances for FRC work. Especially with sub $500 printers
For the case where you are shocking/vibrating it with enough energy to produce some relative motion and/or convert static friction to dynamic friction, you can get a worm gear to back drive at a higher ratio than you would expect from static analysis.
A good way to understand what’s going on is to imagine an inclined plane at the contact angle between the worm and the worm wheel. IIRC this is the same as the spiral angle of the worm itself for a non-enveloping worm/wheel pair. If this angle is steep (like on a 4 start worm), then you can easily imagine a weight sliding down that slope. As the spiral angle gets shallower (like a single start worm) your ramp is shallower, and things need to be more and more slippery to get them to slide down the ramp.
A lubricated steel/bronze pair may have a static coefficient of friction of 0.15, leading to a critical angle of 8.5 degrees. Get is moving and the coefficient of friction may drop to 0.02 to 0.08, giving critical angles of 1 to 4.6 degrees. So, if it starts to slip, it will probably go all the way! Simonelli (below) suggests that a conservative choice for self locking is 5 degree lead angle or less, and 15 degree lead angle or more if you want to back drive. This is a fun chart and gives you some feel for the efficiency of power transfer vs lead angle.
Next up: ratio. For a small reduction ratio a given break-free torque in your motor multiplies into a relatively small worm gear load. This is in addition to the friction in the worm gear/worm system. So, easy to slip. However, for a large ratio worm a little friction in the drive hardware/motor translates into a much larger slip torque. Again, this is in addition to the worm’s built in friction. So, for a given worm spiral angle and a given motor break-away torque the higher the ratio, the higher the resistance to sliding. One takeaway: if you add a brake directly on the motor, you get a lot more bang for your buck!
I am pretty sure that single enveloping and double enveloping worm/worm gear systems exhibit a much more complex friction calculation, as you have multiple contact areas and some speed variation between the various contact areas. BUT, I don’t have the answers on that…
I would make your worm solid! That’s the part that gets the most fatigue abuse. You might try the worm gear with infill, but don’t use it in competition without testing!!! The cyclic stresses seem likely to create problems with the infill interface. Of course, I have no actual testing to base this on, other than my 3D printed change gears, done with infill. They didn’t get a lot of mileage, but they cut the threads I needed! Those are pretty big; you can see a CD case for scale…
Right now everything is solid. Problem we had with infill is that it creates weak spots on round things like gears, pulleys and wheels. To reduce print time and “lighten” the part we put holes in it or spokes
Like you have in some gears above (your hex holes) So you get spokes and control the layering of the perimeters and Minimize the filament usage and direct strength to where its needed. IMO a anything below 100% infill is an inefficient usage of filament
As for the worm its atached to a 42 tooth gear (that gets driving by a 19 tooth that is attached to the CIM motor there is a pair of those Looking from the gear side
you can see the holes in the gear as that gear holds a relatively small load in the rotational direction its about 2 in in diameter so with the CIM topping out at somewhere around 5 in/lb at 40 Amps (cant go higher in FRC) the gear will see less than 12 in/lb torque max
Then the Worm is printed right on top of it
And that is solid. So everything is solid yet holes. spokes etc are used to kinda simulate infill % only better IMO. The modulus on the worm is probably overkill and further tests will show as to “how low can you go” It also allows on that prototype for some inconsistencies that might evolve probably due to material creep - hence the design uses 2 worms and 2 worm gears to cut down on that. Also a solid (enough perimeters) helps with that. So somewhere you could say that I am for 0% infill and hardly a piece has infill and it seems that the use of holes and more perimeters make a stronger part at the same weight and print time than a part that appears solid yet has some % smaller than 100 as infill
I really like the dual worm design for load sharing and force balancing; I think that’s going to really help the printed design!
You may find that you want two different plastics to cut down on them thermally welding together. Also, I would look at using solid lubricants in your grease; MdS or graphite.
Igus makes special high wear filaments that might be appropriate for this application: igus® tribo 3D printer filament made of iglide® high-performance plastics Worm gearing is much more a sliding wear problem, rather than the mostly-rolling action of involute gears.
I agree. For now everything is HIPS as HIPS is cheap and makes good gears. The intent is to at the first sign of “trouble” to replace the troubled piece with Nylon (Taulman 910) a bit pricey - about 8-10 times the cost of HIPS but it was a “savior” in many cases and it works great on HIPS too. Now HIPS has not welded together on anything yet but there is always a first time