How did you scale?

[frcguy]

As we're thinking about adding a scaling mechanism to our robot this offseason I'm curious to see how other teams accomplished scaling this past season. Post details of your team's scaling mechanism below! Thanks in advance.

[ttnn58]

a telescopic arm with a dragon shaped hook on the top:
https://youtu.be/gwu25xif6Tw?t=2m11s
to open it we use 4 pistons (2 for each level)
and to close it we also have 2 window motors to pull the robot up and another window motor to fold the robot

[cad321]

An idea we were seriously considering was a mechanism powered by stored energy. The idea was to have 2 gas springs (normally retracted) and have a winch to pull it open. The springs would have enough force between the 2, to lift the bot (~85lbs-100lbs of force each). This would alow us to take advantage of the extra 5 seconds after the buzzer goes to climb. Theoretically giving us an extra 5 seconds of match play.

The biggest reason we left the idea behind was the lead time on the springs. It is very easy to find gas springs that are normally extended, however we had a very tough time finding a supplier with normally retracted springs in the size we needed and that could have them in under 2 weeks (preferably less). I'd love to see this pulled off. Perhaps a good off season project for someone. I heard people talk about this concept, but I don't recall seeing a team actually implement it this year. If anyone knows of a team that did, id love to see it.

[KohKohPuffs]

Originally Posted by cad321

An idea we were seriously considering was a mechanism powered by stored energy. The idea was to have 2 gas springs (normally retracted) and have a winch to pull it open. The springs would have enough force between the 2, to lift the bot (~85lbs-100lbs of force each). This would alow us to take advantage of the extra 5 seconds after the buzzer goes to climb. Theoretically giving us an extra 5 seconds of match play.

The biggest reason we left the idea behind was the lead time on the springs. It is very easy to find gas springs that are normally extended, however we had a very tough time finding a supplier with normally retracted springs in the size we needed and that could have them in under 2 weeks (preferably less). I'd love to see this pulled off. Perhaps a good off season project for someone. I heard people talk about this concept, but I don't recall seeing a team actually implement it this year. If anyone knows of a team that did, id love to see it.

I recall talking to 254 about potential climber designs at SVR. They did talk about using two gas springs, which would be retracted and held like that using pancake cylinders. When the pistons were activated, the gas springs would extend, and a winch gearbox would be used to retract the gas springs to pull the robot up. However, looking at their technical binder, it seems they used surgical tubing instead, I guess since it's more cost-efficient than gas springs.

By the way, have you tried mcmcaster.com for finding the gas springs you need? I believe that website has a wide assortment of those in various sizes.

[TheBoulderite]

We used gas springs. Watch the end of this video to see it in action. https://youtu.be/8w9TLPjWI4Q

[NShep98]

Our scaler sucked pretty badly. The only match we successfully got up high enough was never recorded. The arm the we used to manipulate defenses was also a 2 stage linear slide elevator with a spring-loaded hook on the end of it. Once the hook got in place, the arm would pull away, leaving the hook attached to a seatbelt on a winch that was supposed to pull us up. Being that the winch was situated at the back of the robot, it would pitch forward, so we added 2 lengths of paracord to the winch that looped around the front of the robot to keep us level.

In theory, this would work, except for a number of issues including the hook being knocked out too early in a match, the seatbelt and/or paracord getting tangled or frayed, the winch jamming, and plenty of other fun setbacks. The moral of the story for this approach is planning.

[AlexanderLuke]

Team 696 used a pair telescoping arms that were assisted out by constant force springs, and then retracted back by high tensile strength cord that ran through the arms and became wound around a spool for each respective arm. The entire assembly was driven by a custom winch gearbox with a pneumatically actuated ratchet and pawl.

You can see the arms in our 2016 robot release picture and thread, as well as in this picture taken from the Los Angeles regional. This is also a video taken from the Ventura Regional that shows us on the red alliance in qualifications scaling (I do not take credit for the video).

They were very successful for us and often worked with little issue.

[hectorcastillo]

We used a hook made out of two pieces of 0.25" lexan that was mounted onto the end of a 1x1 which slid on another 1x1 via the VEXpro linear motion kit. The hook was tied to a strap that went down to a winch. The winch was powered by a 775pro on a 100:1 versa planetary with two 0.5" ratcheting wrenches on the output shaft so that the shaft could only turn one way. As the winch turned, the strap would be let out in order to raise the hook. A constant force spring would extend the 1x1 lift and some surgical tubing would pull the 1x1s backwards away from the robot. When all of the strap was let out and the lift was at it's maximum extension, we would line up the hook against the bar and then continue turning the winch which would then spool the strap in the opposite direction. After getting lined up on the bar, the climb would take about 3 seconds. After the match we would then pull off one of the external gears in order to reset the climber for the next match. The entire mechanism ended up weighing around 6-7 pounds.

https://youtu.be/TjlrXFp4Ru0?t=140

https://drive.google.com/open?id=0B8...UxoTDhwc2pTbnc

[Jeremy Germita]

We added a scaling mechanism for Championships. At the event, we fought some silly consistency issues that we really should have hammered out at home. The hook had a 2/10 success rate with only one of those attempts finishing in a successful scale.

After 118's reveal video came out, we began prototyping harpoon style scaling as it made the most sense for the machining resources, robot weight budget, and time available to us. I think we had a working idea of a surgical tubing powered harpoon shooter around the weekend before our first regional.

Development on a scaler slowed between Los Angeles and Las Vegas as we wanted to focus more on potential improvements to our (never fired in an official match) shooter.
After Las Vegas, we decided that scaling would give us the best chance of getting selected for alliance selections over our shooter, so we began more serious development on the scaling mechanism. We designed winch gearbox plates that were a direct drop-in replacement for our shooter motor mount plates. This allowed us to free up all of the real estate from shooting for a scaling mechanism. This was the "easy part". After calculating the proper gear ratio to lift ~140lb of robot+battery+bumpers(with some safety factor), this part of the system just worked. We used 1.9mm spectra cable as the winch line spooling onto 1/2" hex shaft. This is really good stuff for general robot use. Even when frayed and knotted, it kept on trucking. We ended up replacing the entire line when we arrived in St. Louis and again after our last qual match.

The next major component of the system was the firing mechanism. This was one of the more difficult components. We went through 8-9 iterations of this mechanism tuning different variables like barrel length, barrel dia(compared to the hook shaft dia - more on this in a bit), surgical tubing type/length/diameter/tension, and force required to lock the hook into place. We found that with this design, shot consistency increases as you increase the length of the barrel as the surgical tubing seems to have more time to send the hook flying in our desired direction. This component was built with a combination of 1x1 tubing, versaframe gussets, a few bits of 8020 brackets and two 3D printed parts(which could easily be made from delrin or some other plastics or even wood by hand). It's actually very simple. We'd probably do something similar again if we ever need a one-shot stored energy mechanism. Here's a 1/4 speed video of rev 6 firing.

The final component is the hook. This one saw ~20 iterations with variations in hook dimensions, materials, locking/closing mechanisms, and shaft length. Huge thanks has got to go out to team 842. We saw their scaling mechanism at Las Vegas and they've very graciously explained how their hook worked. Their mechanism was a HUGE inspiration for our version. This is the best picture I could find of the hook mechanism, you could also see one of the early revs of the hook in the 1/4 speed video of the firing mechanism. This is another very simple part. The "shaft" is a length of fiberglass driveway marker with a length of thunderhex shaft pressed onto the end. Early revisions of this component used Andymark churro shaft for the entire length. We moved away from this as it added a lot of unnecessary weight to the hook. This component was the most difficult component of the system. Many of the early issues was getting the hook to close on impact with the bar. Early on, we played with a bar that holds the hook open that gets pushed out of the way, allowing it to snap shut on the bar. We had difficulties with consistency and weight with this concept. After we saw 842's overcentered surgical tubing method, we knew this was the simplest and lightest concept we could implement. The version of the hooks that took the field on Galileo ended up being very similar in geometry and construction to 842's.

For this system, we developed a set of procedures to follow to ensure safe loading and transport of the robot. Without the winch line constraining the hook's flight, the stored energy in the system was enough to send the hook into our shop ceiling, piercing the ceiling tiles 1/2" deep. For robot transport, we NEVER stored the hook while energized. While in queue, we verify that the robot's air pressure is >90% (the locking/release mechanism relied on air pressure) and we don't load the hook until our robot's position was settled on the field. As far as mechanism controls goes, we have implemented a system where the operator had to hold an "enable" button before the firing button could activate the system. This prevented accidental firing on the field.
We learned from 118 and Bane in 2015 and this scaling mechanism in 2016 that stored energy systems should NOT EVER be taken lightly and teams must take utmost care in designing, handling, and operating such systems.

The scaling system was the most intensively prototyped system we've ever created. Ultimately, consistency issues bit us in the translation between the practice frame and the competition bot, but we learned so much about prototyping and which variables to watch for and tune. It was an incredible amount of fun to see this project grow from rev 1 of the firing mechanism all the way to fully scaling on Galileo Quals 60.

[Zac]

This past season with 1768 we used a 2 stage telescopic elevator for scaling. My favorite part of the mechanism was the fact that the entire thing, was driven by just one motor. The climbing sequence involved hooks that expanded in 2 directions, rotating the climber up from the stowed position, extending the elevator, and finally retracting the elevator to climb.

The hooks were folded into a closed position and held with stops, the entire lift was rotated down into a stowed position (with gas shocks springing it back up) and held down using a bearing. The advantage here is that the bearing actually facilitates elevator extension as the entire arm rotates out of the stowed position.

Additionally we used seatbelt strap to hold the arm once rotated up. We didn't want to put the gas shocks in tension. We used constant torque springs to create an automatic winding drum so that we didn't have to rewind the seatbelt strap every match.

I attached some pictures that show everything much better than I can explain it with words.

You can also read more about the climber on page 9 of Greg Woelki's 1768 Tech Binder

One of my favorite videos of the climb is this one (Lower right hand corner of the frame) which includes our favorite play of the season, the SLAM DUNK!

~Zac

[ollien]

Originally Posted by KohKohPuffs

By the way, have you tried mcmcaster.com for finding the gas springs you need? I believe that website has a wide assortment of those in various sizes.

Looks like they have them, but they're mighty expensive. I'm not sure how other springs stack up (admitantly I've never bought them) but they seem high at first glance.

http://www.mcmaster.com/#9502t5/=13q7aou

[frcguy]

Originally Posted by KohKohPuffs

I recall talking to 254 about potential climber designs at SVR. They did talk about using two gas springs, which would be retracted and held like that using pancake cylinders. When the pistons were activated, the gas springs would extend, and a winch gearbox would be used to retract the gas springs to pull the robot up. However, looking at their technical binder, it seems they used surgical tubing instead, I guess since it's more cost-efficient than gas springs.

Here is the explanation of why 254 didn't use gas springs if anyone is interested.

Originally Posted by Ashwin Adulla (from 254 Technical Binder thread)

Initially, the idea of utilizing gas shocks as the means to scale the tower after time had ended seemed like a feasible idea in CAD. When it came to assembling them on the hanger however, we encountered the issue of both gas shocks releasing at different times as was evident in a test which consequently bent the hanger weldment. This problem could have either been attributed to a difference in the force exerted by each gas shock or a difference in force exerted by the pistons with a spring coiled around the rod that were used to compress the gas shocks. We knew that a winch gearbox was both a proven and feasible option, and thus we decided to implement it following SVR.

[Zac]

On the topic of a spring powered climb, I am convinced that there is potential for a constant torque spring solution. Vuclan calls this a contorque motor. It can be seen at the bottom of this page

I had toyed around with this idea for the 1768 climber. Given that the system was already driven by just a single motor, there was already a single shaft in existence that could be used for the contorque motor. The trouble was really that the system needed to both extend and retract, which meant that implementing this would have just meant extending would take longer.

Had we initially started designing with this in mind I think the idea would be to use pancake cylinders as brakes to hold the contorque motor locked until it was time to release it and climb. Another pancake cylinder could be used to hold the entire system down in the stowed position, upon releasing this cylinder the mechanism would rotate up and the elevator could extend.

In this implementation there was no synchronization to worry about as it was a single driven shaft.

~Zac

[C.Lesco]

We didnt, although if you dig around cheif delphi there was a cool magnet graple design in the photos a while back. Kinda wish i was on BREAD sometimes, you guys always seem to be doing something cool. Good luck with the climber!

[BrendanB]

Our climber was a little "Rube Goldberg" this year.

The main system was two 30" pieces of box tubing that with two gas shocks creating an over center linkage so the climber could be "locked" down. A 2" throw piston pushed up on one end of the stowed climber to get it started. The other side the base of the box tube assembly was attached to an 8020 slide. This got us up close to the bar and we used a torsion spring attached to another short piece of box tube to deploy our hook onto the bar which was released with a spring loaded piston.

A video of it deploying can be found here.

The cable ran from the hook down through a delrin block in the top of the first box tube stage. To put tension on the arm and keep it from being destroyed as the robot lifted up we used a cord running from the top of that first stage down to the middle of the robot. You can see the cable here.

The winch was a 27:1 banebots with a CIM motor with a 1/2" wrench on the output shaft to keep it from back driving. The hook us just held in with velcro so it comes loose when we start winching in.

During our first few events we had to add a lot to get this system to work just right. Pre-Reading we added some surgical tubing to assist the 8020 sliding across the back.

After a few successful climbs at Reading the hook started missing the bar as it rotated around. We adjust the curvature of the hook but the big change was adding a long piston that pushed the assembly towards the bar and put pressure on the support cord. This kept the assembly upright while we deployed our hook and prevented it from swaying.

Going into Rhode Island we added a small piston that kept the climber from deploying early. During our first match of the year the climber wasn't locked down all the way so it popped up as we crossed a defense in autonomous. We survived Reading by placing a tiny strip of painters tape between the two stages to keep them from moving and the piston would rip it during the extension.

Before Boston we had a few failed climbs because our wrench flipped directions during the match even with some tape holding it. We decided to get rid of convenience and put a wrench on without the switch. Our consistency peaked during Boston with only one or two failed climbs from a bad lineup on the side.

The rest of the year we saw some wear and tear on few parts but it still works well.