I’ve been involved with FIRST since 2005, but never been involved in building any kind of telescoping arm or forklift device.
I got to thinking about it and couldn’t think about simple ideas on how to solve the main problems:
what slides past what? for telescoping I would think it would be a square tube within a square tube with some wheels or rollers to allow movement. But most forklifts use two sets of tubes to improve rigidity so how do you link them and allow them to slide?
I can’t imagine drawer slides are used extensively since they weigh so much.
regarding opening the telescopes and lifts: I’m guessing the cable+drum+pulleys system is the standard but with the telescoping arms some things just don’t add up regarding the way the cables need to wrap.
I would appreciate some explanations and also pics of said systems would be very much appreciated. I couldn’t find any closeups.
I know FRC233 (Pink) is very strong with telescopes. any other teams to look out for?
Thanks for your help in educating myself and hopefully others!
It’s been a while since 330 did one of their blue lifts…but the last time one of those broke, it was self-inflicted and it merely bent.
Lifts do not have to have their extending sections inside each other. In fact, it’s a lot easier to do a cable-powered lift if they are not inside each other.
I’ll mention cabling routes here…
There are 2 types: continuous and cascading. See this thread for explanation. (post #3) If cabling is used, it is best to ensure that there is a “gravity” cable to make sure that your lift can retract.
There are, of course, other means of telescoping.
If a 2-tube-set lift has the sections outside of each other, each “level” has a partner joined to it in the other set. Joints are typically at the bottom and on one side.
You might want to take a look at 25’s section in the 2007 Behind the Design book, as well. There is another robot in there as well–19??–that had a telescoping arm, IIRC, but it was an internal-section one. Oh, and don’t forget to look at 357 in the same book…
On the cascading rig, how do you determine the diameter of the two pulleys. My guess was that it is determined by the ratio of lengths between the lifting cable and retracting cable, can someone confirm this?
The diameter of the pulleys in a pulley system have absolutely no effect on the reduction of that system. Also, you do NOT want to use drawer slides. They have to much friction, don’t extend far enough, and aren’t nearly strong enough. I’ll have a couple cad sketches up of systems of bearings and cascades in a bit (once I get done with a work issue…)
Edited to add:
That’s a picture of the 254/968 setup from a while back (the blue tube in the picture below). It works well if you want only 1 stage. If you want multiple stages (per the PDF) you can make rectangular frames, then use C-channel with some bearings in them at the tops and bottoms of each frame to guide them.
In '08 we used IGUS slides to do it. It worked remarkably well, however the slides were certainly not meant for that application and we had to do pretty regular maintenance on them. Video of that robot here:
In addition, many hi-lo setups have counterweights to counteract the weight of each stage. That year, we used bungie cord and pulleys so that each stage effectly weighed “nothing” - in fact the motor had to keep the stages down because we decided to also counteract some of the weight of the ball.
One thing my old team used to do a lot was build telescoping sections by interfacing 8020 with a delrin slider (you can see the result relatively well here). Overall, it was simple to implement and mounting pulleys was easy.
If someone optimized the weight, and figured out how to reduce binding more, this design could work pretty efficiently.
Others have addressed cable/chain routing so I will not. Besides when my team did it we used something close to 22 ft of chain for an 8 ft lift. (Not at all efficient routing.)
What I will bring to the discussion is full extension draw slides. They are on the heavy side but made for extending. Nothing to build, just drill holes and mount. What they cost in weight is made up for in simplicity, availability of spares and speed of assembly. Think about side loads and getting hit with them up when looking at load ratings.
Also think about how you will determine full extension or retraction. Our system had a shuttle that ran up and down a tract on the top stage and two sets of draw slides on aluminum tubing to get to 8 feet. With binding, any stage could move or stick during the lift. We counted stripes on the drive axle to determine amount of motion. This was great accept we could not get the controller to remember the position between restarts. This was solved with a home switch in the lowest possible position for the upper stage of the lift.
That sounds very similar to what we considered. We could have used an encoder in the toughbox (relative) with a home switch at the bottom of travel, however we opted for a 10-turn potentiometer riding on the bottom chain so that we had an absolute value and could minimize wiring.
Does anyone have detailed pictures of Hammond from '07? That is probably one of the best examples I can think of, but I can’t seem to find any close up pictures of their elevator on here or their website.
We’ve tried the 80/20 extrusion and Delrin sliders as well, but did not like the binding. That is why we went with draw slides.
With the extrusion/slider set up you typically need two sliders on each ‘post’ for stability. To minimize binding you want to keep the sliders as far apart as possible. This usually means mounting one slider at the top of the lower lift member and the bottom of the higher lift member. As the lift is extended, the slider on the upper member moves with it and gets closer to the the slider mounted on the lower member. Some experimentation can determine just how close the two sliders can approach without bind. You can then put in stops.
A general rule of thumb is bearing surfaces should be separated (lengthwise) about 3 times their diameter (or width). My experience suggests this may be a little close for 80/20 hardware so you may need 4 or 6 times the width. (For a 1" extrusion this means 4 to 6 inches between the slider pairs.)
One last though on bind: Aluminum is actually have fairly ‘sticky’ material. Can you substitute another material or cover the aluminum (plating, enamel, powder coat)? Dry lube sprays may also help.
Just to add to some of the suggestions on multi-stage lifts, Team175 (Buzzrobotics) has used aluminum square tubing, aluminum round tubing, as well as PVC for the stages. Provide enough clearance for Delrin bearings on the aluminum so there is no metal-to-metal contact between stages, and also reduces friction between stages. Instead of using metal cable, we use Spectra braided Cord (1/8" - 1050 lb tensile strength) Which has lighter weight, and allows for much smaller pulley radii, saving weight and space. In addition, we used a double drum-one side to wind up cord for lift, and the second to wind down for retraction of the lift, driven off the same motor. We install an encoder or a multi-turn pot to control lift movement, and put tensioning rubber bands on the cable ends to provide some extra length that is needed due to cord overlap differences on the drums. This also keeps tension on the cord so they won’t come off the pulleys. Good Luck prototyping cable routing.
We have used manufactured slides in the past. I think our Triple Play robot used slides from McMaster. In other years we have manufactured our own using sheet aluminum bent to nest and telescope. Almost exclusively we use Spectra cable I think from McMaster. It is pretty amazing stuff as far as strength vs. size. We manufacture cable reels using delrin and string them through each section of lift from top to bottom of each section. As the reel pulls on the cable it becomes shorter raising the sections. If you design for two sides to lift, I recommend using a dual reel so that the cables are drawn at the same speed and distance. We have generally used FP motors or window motors in the lifts in the past.
From personal experience, don’t try extending a telescoping fishing pole (or anything really) by pushing bicycle break cable sheathing or similar cable into it between two wheels (either both driven or with one idling).
I’ve seen a cool rolamite-like design using a chain as a rack, in a manner similar to the metal bands in rolamite bearings, and a driven sprocket paired with an idler sprocket/roller acting like the rollers in a rolamite bearing. I don’t remember where I saw it, but I do believe it was on an FRC robot (which I saw in pictures only, so it was sometime between ~'93-'08). I’m not sure how well it worked then, but by using large chain, it could be made incredibly strong.
More from the Team 1379 First Overdrive robot experience. We had one square section of aluminum permanently embedded in another. Everyone laughed when I tried to take it out one day.
We machined blocks of Delrin as sliders, but the tolerances were still very tight even with liberal amounts of grease and had to be re-machined a couple of times if I remember correctly. I wouldn’t dismiss the 80/20 linear sliding solutions so quickly. Besides the bearings, they have pre-engineered rollers or you could make your own to fit in those slots.
330, when they used lifts, used Al C-channel. We’d weld two pieces into an I-beam if we needed extra height over what we could get out of our base pair of sections, and then put the I-beam in between the C-channels. The bottom ends would be fastened together (we used a 2-tower system); the tops would be fastened on the lower side. Each piece had a delrin roller at top or bottom, 2 rollers to a sliding pair. The cabling was steel cable, properly crimped at the ends, cascading run, IIRC. Oh, and we had a gravity cable to make sure the lift came down. 2 FPs through stock gearboxes and a winch drum.
One simple comment I will make here is to never underestimate friction. When designing something like a lift where you know you will have a surface needing to rub, roll or slide on another, do not underestimate friction. The amount of resistive force (to lifting) that can result from a poorly designed sliding interface can be staggering.
Also, try to counterbalance your lifter with springs or weight if possible. It will not only make the lift easier to operate (ie: less load on the system), it will also make the system much more consistent and repeatable.