So my team was planning on using a Dart Actuator from AndyMark, but they appear to be out of stock now.
So we are looking for a similar alternative.
I saw this one suggested in another thread: https://www.firgelliauto.com/products/linear-actuators
But it looks like it will only fit a BAG-sized motor. We are planning on building a “tube” which holds 5 power cells and tilts up and down. It will weigh around 15lbs and the actuator will be connected around the middle of the “tube” to the back of the robot. I’m worried that that one will not be fast enough.
I also found this one: https://www.firgelliauto.com/collections/linear-actuators/products/deluxe-rod-actuator
Does anybody have any experience modding these to fit FRC-legal motors?
How difficult is it?
I’ve used, funnily enough, both of these linear actuators on a university robotics team for a robotic arm. Over the course of 3 years we’ve used some combination of both of those actuators amongst other actuation methods (thanks, AndyMark PG188). The second link is definitely not super friendly to modification. The first one MIGHT be adaptable, but I have a hard time believing it’d be simple or better than rolling your own solution.
EDIT: If you want specific details about how we used them as well as potentially further details about the construction of them, let me know and I can try to get some hands on stuff. I still live in the area of my University’s team and have a few contacts that would let me stop by for a bit to get pictures/video.
Thanks for the fast reply!
If you think they are hard to modify, how would we go about building our own? My team has essentially no machining capabilities besides basic saws, so a solution as close to off-the-shelf as possible would be great.
EDIT: more details would be great if it’s not too much of a hassle for you! Thank you!
I don’t know much about those other linear actuators, but we have had poor experiences using linear actuators on large arms. In 2016, we used 2 Dart Actuators with CIMs, and it felt like were constantly having to rebuild them after every match. It was more often that they needed rebuilt than they didn’t. They didn’t like high shock loads and the only way we could prevent them from breaking so often is by machining a new base plate and adding a massive bearing. I can’t speak for those, but that’s just the experience we had. In general, it felt pretty slow also.
We did use a custom homemade linear actuator in 2017 for holding our gear, and had no issues with that, but that was clamping. The machining for that was not ideal, and required two lead screws (left hand & right hand for what we were doing) to be welded together, built up, and then machined to a hex. For a normal actuator, that wouldn’t be necessary of course.
I would recommend using a large sprocket at the pivot point if possible. 1678 in 2016 is a great example of this.
Interesting, thank you for the reply. For a team that has never used chain before (I know), how difficult would implementing something like that be?
Also, a four bar linkage was another suggestion we got for moving the “tube”.
Huh thats really funny! We are basically doing the same thing and just brought out our old dart actuators from 2015 to see how they worked.
We asked our former lead build mentor what his experiences were and his response was that he hated them for 2 reasons. 1. if using 2 at once, getting the timing perfect on both was a pain and they would wear down unevenly which caused them to jam (not a problem for our or, similarly your case), and 2. they were fairly large/heavy and somewhat slow (2-3 second top to bottom with an old, used CIM motor). Both of those factors did not deter us from considering it for our robot this year.
We just bought our first NEO’s the other day and they just came in (ikr a little late to the game lol) and I was wondering the same thing. I will update you tomorrow night when I can get the chance to mount one to the dart. I’ll try and post a video with the difference in speeds too.
Also, there is an included potentiometer that we have installed from awhile ago. I was thinking of possibly replacing it with a better encoder like the US Digital MA3 that looks to fit so I will be trying that out soon.
Looking forward to the responses in this thread 
4201’s DARTs are currently great doorstops, one of which is covered in red Sharpied text:
“No. Stop it. Get some help”
“This is not a legitimate robot part”
“No. Plz don’t. Don’t do this to yourself. Put it down”
etc.
I would highly recommend almost literally any other motion system over COTS or semi-custom linear actuators in this context.
Just out of curiosity, why is this? It’s been years since we’ve used them so we have no idea. What did you not like about them?
Awesome, looking forward to seeing the results from a NEO.
As others have mentioned, they’re bulky, heavy, slow, inherently super inefficient, wear down and require rebuilding often, and expensive.
I’ve seen maaaaybe two robots that used DARTs effectively in ~18 events. Much like COTS treads, they’re a trap.
humm… what is another type of motion do you recommend? Our team is kinda running out of time for build and we are running out of ideas for precise linear motion. We talked about air but for our bot design, it will not work. I am a Build lead on 4557 and I really don’t want to use a mechanical device (LA) that is trashy as you seem to be implying. The issue I am having rn is time, resources, and cost. The LA get over all these issues, because we have one on hand, but might cause many more down the line. Thanks for your help! I am looking forward to seeing any ideas you might have.
Pretty much every implementation of the DART seems to be in an arm, or something that is fundamentally rotational and not linear motion. The classic sprocket bolted to an arm driven by a high reduction gearbox was my team’s go-to, both for reliability and cost. The OP described what is essentially an arm, but I’m not sure what your actual mechanism is.
Ok, so we did consider that design but were worried about how much flex would be in the chain. Our design is building a “tube” which holds 4-5 power cells and moves up and down. It will be long and we are worried about the small “play” in the chain being magnified because it is so long, which would have a large effect on our shooter accuracy. I do not have much experience with the amount of play in a chain setup like that, but I am assuming it will be large.
Why does the arm need to rotate? My first instinct is to say if you don’t have machining resources more precise than saws, you should not be trying to make an adjustable angle shooter. We fell into this exact trap in 2016 (literally the same thing – trying to use a linear actuator to move an arm up to shoot at variable angles) and it didn’t end well for us. We had issues with sensor calibration, lead screws failing, low tolerance machining resulting in inaccurate shots, etc – and we could have performed a lot better that year had we just either gone for the low goal or a fixed high point shooter. Consider how many variables you have to tune to get this working even assuming no part failures.
If you really want to stick with this concept, the large sprocket on the pivot point could work. Since the arm doesn’t have to rotate by too much, you could use a turnbuckle on one side of the chain, similar to how you may have seen teams tension elevator chain in the past couple of years. So long as that turnbuckle doesn’t pass over either sprocket in the full range of motion of the arm, you should be able to take out the slack in the chain with this method.
In general, try to look to resources such as 118’s Everybot, not just for their strategic concepts, but also because they utilize manufacturing techniques that are available to teams in your position. Even if you decide you can do better strategically than what they put out (and really make sure you’re right if you’re thinking this), you can look to the past years of Everybot and MCC designs to get an idea of how teams do similar game tasks and actuations.
Through a bit of trial and error on C-C distances, you can get that backlash to almost nothing (see our 2016 robot).
An easier way to accomplish your desired motion might be a a rope and pulley system that can raise and lower the tower relative to a fixed vertical near the exit for the power cells, though I do see the draw of a linear actuator. You’ve designed yourself into a bit of a corner imo.
Loved your guy’s robots in the past (sad to see you guys go 
) and especially your 2016 robot. Are there any close ups of that chain and gear setup? I know its been done time and time again but surprisingly we have never worked with chain before! Has the chain expanding been an issue over time, especially for slower movements? I was looking at team 5803’s 2018 robot’s arm for inspiration but yours is very similar. Thanks for the advice, we have been talking about it all day on our slack channel and we will post later tonight with updates!
Do you need more than 2, 3, or 4 angles with this design? Pneumatics would be an easy way to accomplish changing the angle with such a mechanism. I would figure out the points you plan to shoot from on the field, see how many angles you actually need, and then design a cylinder based solution around it.
If your shooter is adjustable because you plan to shoot from “anywhere”, I would take a step back and do some strategic analysis to figure out where you are actually going to be firing from.
We want it to rotate for a few reasons including staying within the frame and also aim adjustment depending on location of robot on the field. It allows us to shoot high, low and pick up from the loading station on the ground and up high because of the variable arm extension.
After reading through this thread, it has sparked yet another discussion on our team and brought some things to our attention so thank you for commenting.
We do not have access to a massive machine shop with cnc and such but we do have a fair amount of resources including an old Bridgeport and a couple band saws and the likes. We have the capabilities to do a lot more, we are just trying to find the simplest and easiest solution to our issue that will work well.
Thanks for the comment again!
Wow! That design looks really similar to ours. Great minds think alike
In that CAD, it looks like the “tube” is being driven from a four-bar linkage. Is that something you guys are still considering?