As you might have seen from our build blog, we are building a telescoping climb. However, we really had a hard time finding resources about it. We also do not have the opportunity to get CFS easily in Turkey. So, we are really curious about the thinking process that goes into choosing CFS for a telescoping climb.
How did you guys choose CFS for your climb? How should we measure the force needed and how much tolerance should we have?
Our design member who is CADing the climber chose this. He was curious if it was a bit overkill. Also, we were curious if having overkill CFSs would have any negative effects.
Here are our box profile measurements:
Stage 0 = 60x60x2 mm Length= 500 mm
Stage 1 = 50x50x2 mm Length= 500 mm
Stage 2 = 40x40x2 mm Length= 450 mm
Stage 3 = 30x30x2 mm Length= 400 mm
The springs need to be strong enough to lift;
-The stage theyâre directly attached to
-The subsequent stages
-overcome friction
If you know the weight of everything and the friction is low, something like 1.25X of the stage weight might work, but Iâd recommend something more like 1.5-2X stage weight to be conservative.
Also remember that any force you have extending your arms will be extra force fighting your robot âliftâ when you go to elevate your robot, because you will need to recompress the springs.
^^ This. If itâs for a climb, you should be more worried about the constant force springs helping the robot up rather than helping the lift up. If you make the springs strong enough, you wonât need a brake, but it means that your motors (or whatever) will do the work in raising the lift; the robot rising will be âfreeâ and can be pretty fast if desired. Though, if youâre going this route, gas springs probably make more sense than CFS.
Assuming that the design you are working on is spring extended single cable/rope to retract as @AdamHeard described, here is how we chose the Constant Force Springs.
We wanted to have an âorderlyâ extension where the stage 1 would extend first, stage 2 second and stage 3 last. In order to do this, the stage 1 springs have the highest force with decreasing force in each stage. Note: the main reason we wanted an âorderlyâ extension was because we attached our Wheel of Fortune device onto stage 1 and we wanted to raise that stage to engage the wheel without exceeding the height limit.
The springs that you are looking at are pretty hefty. Remember that you have 2 springs per stage, so you are actually getting double the force per stage. The weight of each stage should be pretty light (unless you are adding a skywalker mechanism) so you really donât need that much force. Friction is the wildcard, but assuming you find a way to manage that, you should be able to use springs with lower force that the one you linked to.
Iâm assuming you are going to use guide bearings at the top of each stage to guide the next stage in and out? As the stages get smaller, the space between the guide bearings gets smaller, so you probably need narrower springs to fit in the remaining space between the bearings.
This combination worked well for us once we got the friction sorted out (although we are still working on a better solution to keep the friction low).
Moving onto your box profile measurements. I am not sure why stages 2 and 3 are getting shorter than the base stage and stage 1. I assume that you want to keep the overall package low enough to go through the tunnel, and also want to have sufficient reach to hook onto the hang bar in various tilted scenarios. In order to get the reach you are going to want, you will want each stage to be as long as possible. Each stage is going to stick out from the stage below it as they nest which makes it challenging to keep the whole package under the height limit. You are also going to want to mount the free end of the CFS a bit up from the bottom of each stage in order to maintain enough overlap to maintain stability which makes it challenging to get the maximum reach from each stage. Given the way the stages nest at the top and the bottom, the optimal package will have the same length for each stage.
Choosing springs with a low number of cycles and high force is good. This year, we used a like this one: McMaster-Carr
It has 10lbf pull force, and we used them in pairs to extend our 2lb elevator. They lost their extension force pretty quickly, so it was good that we chose such powerful springs. Maybe aiming for 5-10x force is very safe just in case you have too much friction. An extra 40lb load to climb with is not that much compared to the 150lb robot.
Basically, a worn-down 3k cycle spring is smaller and stronger than a fresh 25k cycle spring.
This is interesting. Have you actually measured this reduction in force? Do you know what the current pull force is and how many cycles you have put on them?
We have not noticed any reduction in the pull force of our springs. We have probably only 50-100 cycles on our springs so far (we have not been counting), and we have not taken any actual measurements, but I was running them in and out by hand recently while demonstrating the design to a new sponsor and at least to my completely un-calibrated arm, the pull seemed about the same as when we did our initial testing as we were assembling it during build season.
I havenât measured it, it just feels weaker and doesnât extend quite as fast. Probably have a couple hundred cycles on them between assembly, messing with them, and actual climbing. Are you also using 2k or 3k cycle springs?
Itâs been a while since Iâve been in the lab, so they may have gotten also weaker recently while they were left stretched.
That is an interesting theory about storing the springs extended. I know that our climber has been sitting on the workbench with the stages extended such that the springs have been coiled back up into their retracted state since shortly after our first competition. It did probably spend a total of a week or two stored with the springs in their extended state during the period before, during and after our competition, but probably not much more than that.
I would be interested in hearing whether this is something we need to worry about. Would it be possible for you to measure the pull force somehow to confirm that the springs have in fact relaxed?
I havenât tested them since they were left stretched. Once we get back in the lab Iâll make some kind of a rig to test this with a fresh spring, as itâs been a perennial worry for me every time a robot with CF springs goes to champs.
The friction to overcome comes from pretty much every part of the assembly - the coil might have some friction, the bearings arenât 100% perfect at transferring energy (and can be far worse if thereâs a misalignment somewhere), anywhere aluminum is sliding against aluminum, etc. - ideally this friction doesnât amount to too much, but in FRC applications it is usually significant enough to warrant accounting for.
The majority of the friction comes from each of the stage tubes sliding inside the larger tube. While the guide bearings at the top of each stage offer good low friction solution for that end of the stage, it is typically impractical to install bearings at the other end of each stage inside the nested tubes, so you are generally incorporating some sort of sliding plate type system at that end of each stage. This sliding plate needs to be close fitting enough to not result in too much side to side play in the system or the whole arm gets wobbly but still loose enough to allow everything to slide.
Use of various low friction plastic end caps seems to be one of the preferred solutions. PTFE helps cut down on this friction and you can machine it pretty easily. Various 3D Printed plastics also do pretty well. Metal plates can work, but metal sliding on metal is not the preferred solution if you can avoid it.
This year was our first time building an arm like this and we relied heavily on previous designs. I will say that we learned a lot of valuable lessons and we are still working the kinks out. So others may have some better suggestions for how to keep the friction low.
We actually have our CF springs the other way for this reason. The motors are fighting the springs to extend the climber, and then assisted by the springs to retract it again. It meant that we could design the gearing for a lower maximum load, since the motors only ever have to move about 80 pounds.
I assume that the telescoping mechanism starts to look more like an elevator mechanism with pulleys and rope to lift each stage? And you need a retract rope as well as an extend rope, right? So you probably end up with a double spool that pulls in the up-rope while paying out the down-rope when you are extending the arm and vise-versa when retracting the arm?
There is certainly a trade-off to be made between an elevator-like spring-retract telescoping arm and a spring-extend rope-retract telescoping arm. I would absolutely encourage anyone considering a telescoping arm (including the OP) to evaluate both options.
In general, I would say that you are correct, that gearing for this lower lifting load will result in a faster climb. However, with a spring-extend arm, the extend motion will not load the motor at all (the rope is actually holding back the extend action), so the motor will be running close to free speed (divided by the gear ratio of course). And since the climb only needs to lift a few inches, the retract motion at a higher load and therefore lower speed can still be pretty quick in terms of elapsed time. With a spring-retract arm, the motor will be heavily loaded on both the extend and retract motions and therefore will take longer to extend. Since you canât extend the arm (or at least fully extend the arm) until you are in the Rendezvous zone, the total climb time needs to include both the extend action and the retract action. While the retract time on the spring-retract arm will be quicker, the difference in the total time to climb may not be as great as the difference in gear ratios alone would suggest.