# Linkage Design With Toss Up Potential

I’ve been seeing a lot of teams basing their initial designs off of parallel linkage designs that have worked in the past, and ignoring the 40" tall elephant in the game for the time being. I originally came up with this design near the end of the Gateway season, and thought it was the logical extension of the Watt’s 6 bar design that’s become so popular. Seeing as Sack Attack did not require teams to try to reach higher with their lift systems, it never showed up independently. I’ve decided to post it to give teams a starting point for more sophisticated linkages to reach new heights, and try break the mindset that “good enough” is always the way to go.

It’s an 8 bar system, composed of three parallelogram 4 bars connected to each other in a geometry which resembles the Watt’s Parallelogram 6 bar, though not the doubled up 6 bar that’s been typically referred to as an “8 bar” around here. It has a similar effect to the 6 bar design, but allows for much greater height to be gained from the vertical section in the middle.

Even without taking into account extra height from an endeffector, this linkage on its own comfortably reaches the 40" hanging bar while crossing under the barrier. It does so with such comfort that you aren’t forced into very specific geometry as with some other lifts, and can modify things a bit more to suit your design’s exact needs. It does not rely on chains or gears to maintain its motion, which can be sources of slop, and doesn’t have the poor transmission angles or numerous joints of a scissor lift. It should be able to be constructed in a similar manner to other linkages, though I’ll leave implementation to the teams.

But the most interesting thing you’ll find about this linkage if you actually build it is that as drawn, it falls into a small category of linkages known as Gruebler’s Paradoxes. This means that according to Gruebler’s equation, which predicts the degree of freedom of a linkage, it should lock up with zero degrees of freedom. However, due to the exact geometry of the linkage, one pin joint in each of the small diamond shaped regions is redundant. Keeping these redundant joints in place adds more friction, but does an incredible job at keeping the linkage rigid throughout it’s motion, as any slop or deformation will fight these redundant joints. You may want to remove the joints, however, if you want to offset one or more of the stages to produce non-parallel motion.

Why don’t you put a 12 inch bar attached to your lift? Like everyone else? Are you trying to hang at 28 inches for a reason we don’t know instead of the logical 16?

Options and insurance. All drawings of traditional 6 bar or chain bar lifts I’ve done give only a couple inches insurance in reaching the bar, and that’s making some pretty generous assumptions about reaching vertically, actual part geometry, and so on. I recall a lot of similar drawings in the first few months of Sack attack of designs that could go under the trough and reach the high goal for scoring/descoring with the main manipulator; on paper they worked, in practice, very few were built. While I don’t doubt that a 4 bar/6 bar/chain bar hang can be done, it’s clear that restricting yourself this way is going to cause many teams to design themselves into a corner, with extremely specific lift geometry dictating choices on bar latch, intake, and so on, rather than optimizing these elements and building a more flexible lift that can be modified a bit as needed. As shown, this lift with manipulator height included can easily reach 50+ inches from a starting 12, allowing a designer the flexibility to do a number of things, including:

-Use a larger, self-centering latch to reduce operator error while hanging.
-Use a wide intake which necessitates pushing the front of the linkage back in the robot.
-Mount the latch anywhere they choose on the endeffector, rather than being limited to the top.
-Have a less near-vertical max resting point for the linkage, increasing stability at maximum height.
-Use an unorthodox drivetrain which restricts room for lift to reach floor level.
-Shorten the robot further for CoG reasons.
-Use uneven links to produce variable manipulator angle at the expense of lift height which would put the latch out of the bar’s range on a 6 bar.

Another advantage of a high reaching lift like this for toss up is that at the 24" height for locking up balls, you are at your maximum of horizontal reach.

And, I don’t think toss up has been around long enough to say what “everyone else” is doing. Now, with 10 months left, is a better time to be sharing and exploring new options, than building to mirror the few designs that are already out there.

Without special gearing (probably a winch, planetary, or top support) your axles will probably twist under the weight. I’m just guesstimating that you would place a hanging system maybe 24 inches away and assuming your going to use combination gearing. With that, your drivetrain, tower, and part of your lift can only be 8 pounds before axles give way.

either way though, you can still add a hanging mechanism like you would with the middle extension. That way you get more “fake height” with less required torque

I agree! I remember in early Gateway, the most common designs included 1103-style claw robots with vertical lifts, wheel-leg top-rollers, and dual-tread robots, all of which had been successful designs in past games. None of those designs ever came close to being as common, or as effective, as the 6-bar, side-roller combo that was optimized for all facets of that game (high capacity, speedy pickup and lifting, covering goals, scoring across to the other 20" goal, etc.).

But are those useful qualities this year?

I’ve heard that top rollers and side rollers with four bars were nearly identical to side rollers in match play. Plus, Commonality doesn’t necessarily mean it’s a good thing though However, i’ll admit that i think there were a severe disadvantage for top rollers in skills challenges.

True, there were some teams that would vary the design and get similar results, but my point was agreeing with JoeG that it is too early in the season to know what “everyone else” is doing. The common design at the start of the season is usually not the common design at the end.

And while it is true that common doesn’t necessarily mean good (especially this early on), many advanced teams think alike by the end of the season. When people all get an idea of which strategy is best, they optimize their designs to fit that strategy; usually, all of the optimized robots (all designed around one set of tasks) end up looking very similar.

I managed to pull together a quick CAD model of that design. Didn’t use quite the right lengths so far(so it makes it to 35" atm), but it certainly is quite the arm. As for having a pin in all 4 spots on each diamond, is it a noticeable amount of friction increase? Also, would rubber bands be placed similarly to other linkages? I can’t wait to play with different setups of this arm

I’ve seen this done before, I think mostly by teams who didn’t realise that they were using redundant joints and bits of metal.

I don’t think the problem with 8-bar linkages previously was that they weren’t rigid enough and they needed to be braced more. I think the problem was that they are impractically long and impractically heavy. Even a maxed out six bar is pushing things a little bit, particularly if your manipulator or the objects it carries weigh a lot.

Efficiency in design can be a good thing. If you can do something a simpler way with less materials, why wouldn’t you?

I’m not quite sure how this doesn’t resemble the 8 bars we’ve seen a bunch of teams build before. It contains an 8 bar, it just exploits the fact that the whole linkage is parallel to add another joint between the bottom and middle and between the middle and top. I mean it’s not exactly the same, but it’s similar enough that I’m pretty sure I’ve seen teams build this and think they were building a “normal” 8 bar.

Then you just have a standard non-reinforced 8 bar again.

I mean I’m not saying 8 bars are never the right design, but they seem kind of like overkill and the extra reach isn’t free.

Thanks for the great explanation!

I am wondering whether it would be worth mounting a manipulator for normal game play on this arm or whether you could somehow fit one of these for hanging on a robot with another arm for scoring…

I think the “doubled up 6-bar” JoeG refers to might be this?

Key differences being only 3 horizontals not 4 and quite a substantial increase in potential lifting height from a given starting size.

Depends what you mean by “special”… the method of bolting the arm straight to a gear so that you are not transmitting the arm torque I was adequate for hanging in Round Up, provided the axle was supported properly and bearings were close to the load.
A simple improvement to that is to pivot the top high strength gear on a bolt using the round inserts. I would be surprised if this wasn’t adequate. I think things would bend or break in other places first.

My first mistake could be equating a reversed double 4 bar with an 8 bar. Though they have diferent structures, I believe they use similar amounts of torque.

The second assumption I think I was making was that he’s using combination gearing. Therefore, the 12 inch gear couldn’t be attached to the mechanism.

If not, and if I remember correctly, teams didn’t use a mechanism as long as an 8 bar in round up. They would just extend the vertical bar at the end of the linkage. If it’s truely longer than round up systems, I believe that would be enough to cause axles to twist in the second stage of combination gearing.

As for other things breaking, I remember with the rd4b, I would place strips of 1 by 25 metal to stop axle bending, reduced jumps in power to reduce jerk, doubled up on HT gears, screwed bolts to metal, and used elastic to minimize force. Either way, the axles in our center stage gave way.

You do not need to use axles at all… 2" screws with green inserts. Then screw all the gears to other gears so no torque is transferred. [ATTACH]7539[/ATTACH]

That’s an example of my robot from sack attack it had a transmission so the towers had to be wider. Because of this I had gear skipping and had to use shaft locks to hold the axles together.

But I am not sure why you assume with compound gearing your axles will twist. You can work around this.

well first i should correct myself:

Meant to say 12-tooth.

Using ratios with 12 teeth gears required axles without modifications. However you did bring up a good counterexample of combination gearing past 1:5 without using axles.