@rose_bandrowski has joined the channel
@Kenneth Wiersema set the channel purpose: Make an intake
@julian_bonifaci has joined the channel
@declan_freemangleason has joined the channel
@bill_bandrowski has joined the channel
@Ethan Rininger has joined the channel
If you get time tonight, perhaps looking into the grabbers and arms that Robot in Three days used will be useful for us tomorrow. I'll see about posting some videos here tonight
We have around 12" we're going to have to extend with Paul's current scissor design to be able to place a block on the second level of cubes. So keep that in mind. Also that number isn't final.
A link to a helpful Chief Delphi threac from another team. https://www.chiefdelphi.com/forums/showthread.php?t=161165&highlight=belts+intake
Great example of what we've been focusing on
Update: the mounting place is 1.5" from the top of the second layer of cubes when an opponent is winning
Working out how far the grabber can extend and be inside the 16" perimeter
currently its in contact with the cube as much as possible and still within 16"
39.5 is the angle the grabber is at (from horizontal)
Not necessarily final, however a start
Redrawing in autocad since you can't see anything I'm doing in fusion sketch
Here's an alternative sketch with a cut out of the bumpers
That degree angle from the horizontal is 54 degrees
The degree angle there is about 38 degrees
as you can see, if we have a cutout from the bumper that is relatively cube shape we get more surface area to grip the cube within that 16" restriction
@kaedric_holt Please be sure to get Rose’s drawings/models and model this up. We’ll add it to the full robot model tomorrow and see how it all fits together with the top of the scissor.
I can send you a file that you can dimension from
It's in autocad, can you work with that or do you want me to send you a dimensioned picture?
*Thread Reply:* The PDF I sent you is annoying, but it's basically just the length of the plate and the arm that's holding it. I can give you a better photo if necessary.
It's fine, I'm going to trust ya math and finish building it then.
Is the length of the arm the same w/ full bumpers?
I think it makes more sense to have a bumper cut out so design it with that
We can confirm with the team tomorrow if we can have space cut out of the bumpers but if you can, design it that way now
I posted this video in the engineering channel last week, but it looks like it may be useful to you here. I have a couple of drawer slides you can have (8 inches long, I think), if they would be useful.
Posting this question in #programming and #intake_grabber ...Besides vision capability, has anyone seen other teams provide the driver with a "successful capture" indication when the robot clutches a cube?
@Terry what do you mean? Basically a green light or some sort of indicator that says the cube is successfully captured?
We haven't considered an indicator, no
I'd like to consider it though. I don't have an extensive knowledge of electronics so I'll need guidance regarding the sort of system that will allow for such an indication
@rosebandrowski I was just going through a list of robot input/outputs in my head for the #electronicpneumatics team. Was wondering if the driver, working on the far side of the field, would need some positive feedback that the cube was successfully in our grasp so they can drive away. If we can do it all with vision, great. If not, then perhaps a pressure contact/switch of some sort should be considered. The driver is going to handle what? 6 to 8 cubes in a match? We want to eliminate as much guess work as possible in those precious few opportunities.
@Terry I agree that sounds like a good idea
@Terry - i would wager that a feedback switch would be a more guaranteed method to ensure drivers that all is good.
there should ample space for a positive feedback system as well
Either on the backplate or maybe the arms
There will likely be pressure from both though even if the cube isn't grabbed successfully so I say it needs some set amount of pressure
Also, we have a lower intake system (harvester) that is meant for alignment so there is less work from the driver
and will increase the success rate of pick up from the scissor lift grabber/arm
@paul_vibrans I how do you plan on being able to pivot the grabber arm? using the winch system? how should I spec out a motor for that?
The grabber pivots just the way the one on the prototype does, with wires that go around pulleys to the bottom. At the bottom there are two pulleys that don't move up and down, spaced 12 inches apart. The wires that go between these pulleys are attached to a slide that is moved by a lead screw, a custom made linear actuator, that is turned by a CIM motor. The parts except for the CIM are on order.
Okay, are we looking at a normal Andymark CIM that we already have in stock? And do you think that's going to take a gearbox and if so have we speced one out and do we need to order one?
I'm not sure if that actually falls under scissor lift's to-do or my group's to do. They're integrated
The screw has direct drive without any gears. I don't know which group will own it.
I want to start a conversation about how the software team needs to control the intake grabber. Right now I don't know how many actuators (motors, pneumatic solenoids, servos) are being used, and what they do. Also, we need to communicate the sensors needed in order to perform operations. Are motors being run with precise speed control (meaning there are encoders attached to them)? Are there limit switches that we need to sense in order to turn off motors before we burn them out? Basically I'd like to know all the states the intake grabber can be in, what motors are running while in those states, and what actuators do work in order for the physical state to change.
Saw this grabber here just released by West Coast Products - sharing not because we'd buy it but because it's an interesting design & there might be some ideas from it worth borrowing - it's below the elevator. https://www.chiefdelphi.com/forums/showthread.php?threadid=161646
Here's a list of updates from yesterday: We had designed the grabber arm with a different cube from the actual FIRST one, we need to make some adjustments for this -Longer grabber arms -Alignment with the handles (this will take thought on what angle or placement of the grabbers will get us that) -10" pneumatic (we used 8") -Grabber arm CAD is done? -We need to finish the harvester ideas and finalize in cad
-need to look for hallow blue tubing on mcmaster(?)
Yes but we can design the grabber attachments so it's not too big
After the meeting we talked about what pneumatics to order and I told him to order more of the 8 in and a 9in
I think you decided against the 9in because we didn’t have one
Yeah and 10 is slightly too big but we decided we could work around that
Please confirm and let peter know before he places the order
@brianbonifaci @julianbonifaci I thought we decided on the 10" pneumatic?
I don’t remember off the top of my head. We placed an order using the voucher last night and I’m pretty sure they are placing another order Sunday
Okay. I'll see about confirming that
We planned on using the 10" we already have.
If that’s the plan we will probably need to order more
Okay I can email peter to order... how many more do you think?
We want at least three of the same. 1 for the primary robot 1 for the secondary robot and 1 as a spare
also @Mark Tarlton we didn't finalize plans on the harvester how it'll retract and extend with pneumatics, but we're placing an order so do you have an idea on what sizes we're going to want?
We're working around having to rotate the harvester so it fits in the frame perimeter and we were discussing a design powered by pneumatics
that hasn't been tested or thought through though so I'm not sure what to tell peter or order
I saw someone entering an inventory of cylinders into a google doc.
Based on our pneumatics inventory we only have 2 10in pneumatics
Brian, any thoughts on the size pneumatic we'd use for that harvester system?
If peters putting in an order I should tell him to order something so it gets here even if we're not sure
So I'm assuming if we're going to be ordering a total of 4 10" pneumatics, you're going to want equal number of the 04 bore or am I assuming incorrectly.
You only need a total of 3 and you already have 2
Okay one ten inch pneumatic and a 04 bore
For that we're going to need one pneumatic per side (probably?) so we'd need two of the same size
Do we have enough air tanks, might need more reserve with all this.
I'm going to wait and see if Mark has any input on the harvester pneumatic before I send that order.
Make sure @peter_hall follows through with this order request ASAP. Air tanks and pneumatics cylinders.
Okay. Should I send him a request for 4 more 5" pneumatics? i'm unsure that's the one we want to use so I'm hesitant to request it
What supplier do we use for pneumatic rams? I looked at Andymark and don't see 10" stroke so I assume we're getting them elsewhere. I
@rose_bandrowski.. I think we will need a 6" stroke to lift the harvester. I wouldn't want to go shorter.
It looks like the 10:1 gear box kit from Vex pro is out of stock, we already have enough for the first and second robot though so I'm going to call off the order for now
Anybody have a picture of the prototype grabber
@Mark Tarlton commented on @Micah_Glasby’s file intake arm doodle: That's nice and clean and it looks like it would do the job. We may need to fuss with the position of the L-bracket to get the pivot to work just right. I think we need a pulley with a diameter of 2 to 3 inches to make sure the belts don't rub against the frame or to give room to add a slippery surface on the inside of the arm.
Interesting insight from team 3847 that having more grab on one side of the intake than the other helps with acquiring cubes that are not squared up relative to the robot... https://www.chiefdelphi.com/forums/showthread.php?threadid=162013
Question regarding our Switch game: Are we dropping/placing (which requires contact with the fence around the switch) or are we launching using the flipper arm (which just requires being in the right zones)? Relevant conversation: https://www.chiefdelphi.com/forums/showthread.php?threadid=162017
We are not likely to be launching (flinging?) with the flipper arm, as the mechanism timing (motor and pneumatics) will be hard to synchronize without a great deal of programming effort.
Rose: I saw that design and I think it showed the importance of being able to pull a cube away from the wall and 'gather it' in one motion, which I think we can do if we refine our design.
I'm going to be there about 10 minutes late today, I'll be there as fast as I can
Is there a CAD drawing for the pulleys being used for the intake?
Let me know if his doesn’t work and I can send a different link
I'll be about an hour late, but I'll get there as fast as I can.
In the programming thread, Dana thinks that the IR Sensor that Randy Groves has been working with will be able to sense the presence of a cube and can replace the switch we had planned on.
I have a version of the complete intake arms and mounts in CAD. I'll attach some screen shots showing how it fits in the robot
Thank you, I’ll see about transfering it to the main file. This sort of helps one of my problems
Do we have room for larger diameter confirming green wheels ?
Some of the stuff in the background is very wrong. What CAD background is being used?
*Thread Reply:* The images were based on the "New Final Robot ..." model as it was yesterday afternoon.
it's very tight. There will be trade-offs to go larger and still fit.
I took a closer look at using larger wheels on the intake arms. We would need to make the arms longer so that it would fold inside the chassis. Longer arms may interfere with the "Climb with friends" stabilizers.
*Thread Reply:* I recommend prioritizing optimizing the cycle time of the harvest/grab/aim/place sequence above "Climb with friends", which seems very likely to be a casualty of limited remaining time. If fast harvesting requires longer arms, bigger wheels, etc. then go for it because switch/scale cycling is our top priority. If we DO get to CwF, I think we will find a way (use space above harvester possibly). Please do enable the arms to fold in at climbing time so they don't interfere.
If you feel confident that the smaller wheels are effective then no need for larger wheel.
I will not be there tomorrow. It would be great if tomorrow the pneumatic cylinder could be attached to the second harvester. After that we need to test the second harvester, make sure everything works, and attach to robot. Holes are already drilled on the robot, the harvester just needs to be bolted on. Other than that, before bag day all we really need to do is make spare parts, which shouldn't take too long. Let me know if I'm missing anything, otherwise I'll see you all on Sunday.
The second intake was completed and tested today. Paul added lightening holes on both arms overnight so it's looking pretty good. Also Brian added a needle valve flow-restrictor to the main robot with flow set at 3.5 turns from fully closed -- it works really nicely!
We need to put together a list of spare parts. Off the top of my head, we need: a 7" pneumatic cylinder, more of the green conforming wheels, some 1/2" hex spacers (particularly the small ones for use on the wheel end of the arms. small 1/2" hex collars. We also need to make sure we have spares for all the various sizes nuts and bolts.
We need to take the extra gearbox/shaft components in case we loose the circle clips or locking screws. We also need to document the spacing and layout of the various components when assembling the complete unit. The spacing is particularly critical and is hard to recreate under pressure.
I think we could consider replacing the "churro" shafts with solid hex aluminum shafts -- particularly at the motor -- to tighten the assembly up. Also on the second robot, we're experimenting with putting the green wheel on the lower side of the upper arm to get a better grip on cubes. Finally, first chance we get, we should try to swap the intake assemblies on the main robot so that we're sure they are interchangeable and to put the lighter weight parts on the main robot.
We also want to think about adding a limit switch that indicates when the arms are closed. The programming team could use this to make sure the intake motors aren't running when the arms are closed.
Probably a crazy thought... there's no way to make the harvester rotate back once it's holding a cube in order to shoot into the switch is there? I noticed how nicely it pulls cubes in and then shoots them forward when motors are reversed - had to ask, because that would surely be a fast way to cycle on the switch
That would not be possible with the space we have available
even with it as it is, we should be fairly fast at cycling to the exchange at least
Students, Do we have a spare air cylinder for the harvester and grabber ? If not ,give me the specs to order.
We do not have a spare cylinder for the harvester. We need one that is 7" long. I believe the diameter is 3/4" but I want to double check that before anything gets ordered.
But jack should have already placed the order
7" x 3/4" is correct. I set down with Jack yesterday to verify the specs. I asked him to order a pair.
Track starts tomorrow so I can't make it to the meeting
How heavy is the current grabber and intake system together
I don't know. We didn't weigh them when they were off.
I'd guess around 7lbs for intake system with 1/3 of that weight being motors and gearboxes. I'd also guess around 3 lbs for grabber.
I have a good start on a combined harvester/grabber design. The motor will be close to the flipper hinge. I hope for a CAD image tomorrow evening.
There is also a camera mount concept for the top of the scissor lift.
@paul_vibrans some of the mechanics will be in the robot room from 3:15 till 5 pm on 3/7 and 3/8.
@paul_vibrans where did you say you found a plastic hex shaft? All I found on Vex / West Coast / Andymark was ThunderHex (7075 -T6). On other sites I found Nylon hex and other weaker plastics.
I did not find plastic hex shaft. I found plastic 1x2x0.1 rectangular polycarbonate tube at VEX
7075-T6 aluminum is stronger than ordinary steel, less ductile, not as stiff and 1/3 the weight.
Do you want me to order ? Exact item # ? @paul_vibrans
@Harrison Gilmore has joined the channel
Regarding potential design changes to the harvester (as indicated in James S. email to the mentors), I would appreciate it if the team could post some sketches to show how the device will be actuated, pointing out how it differs from the original design. What caught my eye was "replacing the pneumatic with a pulley belt system and a spring..." That implies changes to plumbing, wiring, and programming. The earlier we can review and comment on this plan, the more likely we can all arrive at success at the same time.
I hope to have a proposed design together today. I'm working through the details of the various options now. In the minimum, no changes to the control will be needed but we won't be able to change intake wheel size. If we go to the belt system, we might need to add one more pneumatic cylinder and the control sequence would change as well. If we go this route, we may be able to try larger wheels, place the arms below the grabber and explore other configuration options. For sure, we are staying with pneumatics and the same general components.
Not necessarily, I thought we would still have one pneumatic, that has the belts connected with it. The set up is still relatively unchanged. The hope with electronics is that the setup is unchanged. We will not be adding a more controller.
The big changes for the harvester is where the arms start at the beginning of the match and how pneumatics drive the position of the arms. The idea is to have the arms point towards the back of the robot when the event starts, swing towards the front to get into grabbing position and then swing the rest of the way in to grab cubes. By starting in back, we aren't as space constrained as we are now. Larger wheels can be used and we will have more flexibility in how the arms are positioned.
The biggest change in the hardware is with the pneumatics that control the arm positions. They need to change to support the extended range of motion (up to 270 degrees of rotation) and three stopping positions instead of two. We'll need to add a second pneumatic cylinder and solenoid. The way it would work is that springs pull the arm so that its pointing straight back. When one pneumatic is activated, it will pull 1/2 of the belt which rotates the arm about 130 degrees . This is about the same position that we use now for the "arm open" position. When the 2nd pneumatic is activated, the arm rotates another 130 degrees. If a cube is in position, it will grab the cube. If no cube is there, then it will rotate towards the front of the robot but probably won't close completely. The linkage between the belts and the pneumatics is such that if only one arm hits the cube, the other arm will remain open as well.
Moving the pneumatics will fix the problem where the arms get caught in the closed position. It will also fix the problem where one arm closes by itself when the cube is off center. Larger wheels in front will give better grip on the cubes and will act as a bumper when the arms run into things.
We'll also have flexibility in setting the height of the arms. We might be able to set them low enough that the grabber can be used while the arms are holding the cube.
We have enough materials on hand to make one Combined Harvester Grabber except for the following items: 2 each VEX 217-3463, 2 each VEX 3460, 13 each VEX 217-2737, and 16 each VEX 217-3875.
If we build a second Combined Harvester Grabber, we need to acquire the following additional items: 2 each VEX 217-4100, 8 each VEX 3432, 2 each VEX 217-3200, 6 each VEX 217-4009, 4 each VEX 217-3199, 1 each VEX 217-4973, 1 each VEX 217-2818, 2 each VEX 217-2816, 2 each VEX 217-3463, 2 each VEX 217-3460, 16 each VEX 217-2737, 16 each VEX 217-3875, 3 each WCP 0097, 4 each WCP 0096, and 4 each generic 1603-ZZ.
@kaedricholt or @jackchapman please put items in vex cart so we can purchase if we decide this route .
There will not be enough time between Glacier Peak and Portland to build and test a Harvester Grabber, let alone building one for the competition robot. Also I will be traveling for 4 of the days in that period and will not be much help with construction.
I will get everything in the cart as soon as I have access to a computer.
@paul_vibrans Not clear. Are we building a combined harvester ? or not ?
I can't make that decision. If a revised harvester and the existing grabber work well enough to win in Portland as determined by our Glacier Peak performance then we won't need a Harvester Grabber. If they don't work then we need a Harvester Grabber ready to test the day we get back from Glacier Peak.
Not right now, possibly later as a back up plan, if we need something better from glacier peak
I think the question Paul raises is whether we want to have a solution in our pocket before the need actually arises or do we bet everything on what we're doing now. It's a risk/reward question for the team.
There is a cost (time, materials, and team attention) in creating the integrated H/G solution. On the other hand, if we find we do want it, the development lead time is such that we have to start now.
The 4 single polycord pulleys from WCP are backordered.
What else do we need gearbox wise? I know about the extra ring gear needed, and I got spares for the two harvester motors, and one for the main climber motor, but is there anything else?
We have everything we need for the first one except 4 belts, 13 collars and 16 bearings. They are on the first shopping list from this morning. The second shopping list is for the second robot and is in addition to the first shopping list. The second shopping list replaces what we took out of inventory on Thursday.
One observation from watching fast cycling teams is the ability to softly “shoot” the cube a few feet helps a lot with cycling fast and not getting hung up on the scale, so I would hope any complete redesign we do has that requirement on the list.
@paul_vibrans @coachchee I didn't see an order for the polycarbonate box beams. Do we still want to get that?
I did not get that order. @Mark Tarlton FYI. Students are available on Mon around 3 pm and Tues after 3:10 pm.
The new harvester will use aluminum. I thought Paul was considering the plastic tubing for his design. We saw it being used on another team's integrated grabber that's very similar to Paul's design.
The cross structure is aluminum. We have aluminum in inventory for grabber arms though polycarbonate 1x2 tubing is preferred. It is VEX 217-4819. I think we can survive without it at this time.
We need to replenish our stock of 1"x 1-1/2"x 1/8" rectangular aluminum 6061-T6 tube.
decent looking polycarb intake - ideas to borrow possibly: https://www.youtube.com/watch?v=PqV7966H8aQ&feature=youtu.be
*Thread Reply:* They should work better than what we planned. They have 1/2" hex shaft but they're 1/2" high compared to the 1" height of the big green wheels. This would allow us to space the arms closer together.
*Thread Reply:* Regarding the height difference, I just noticed in the photo posted, that they are using two wheels stacked one on top of the other so it's back to 1" total thickness. That would be another option to try. The spikes might also reduce the change of arms getting wedged together
“Best intakes” thread on Chief Delphi: https://www.chiefdelphi.com/forums/showthread.php?threadid=163910
Tell me if the link doesn't work, I've been having issue
I wish I had more energy to keep up with the new intake. That looks cool! How effective is it? I’m assuming since the wheels are larger it flips up now?
It rotates so the arms face the back of the robot at the beginning of the match
Has there been a discussion with programming yet?
*Thread Reply:* There has not been a discussion with programming, as far as I'm aware (If someone talked about this while I was gone, please post a message filling us in.)
We need an overview of the different states that the harvester needs to get into, and roughly how they get there. This can come from someone who has a good understanding of the whole design. This is the most important thing you can answer.
We also need an electrical overview, probably from @peter_hall. How is this wired, what new solenoids are there?
*Thread Reply:* Currently the plan is to have two separately actuated cylinders. For the intake we will have two solenoids each controlling one cylinder. The solenoid ID's are 5 and 6.
*Thread Reply:* One of the cylinders is longer than the other. So we will have to work on which is moved and when.
*Thread Reply:* Without feedback you're not going to get those moving in unison. We could use a hall effect sensor to tell when we should begin to actuate the shorter one.
@Mark Tarlton will have the best answer for you
FYI, @Mark Tarlton Some of us are meeting on Sun. 3/18 from 12 to 5 pm to finish new harvester.
@declan_freemangleason: here's the first part. poke me if it's not clear or you need more info
The big changes for the harvester is where the arms start at the beginning of the match and how pneumatics drive the position of the arms. The idea is to have the arms point towards the back of the robot when the event starts and then swing towards the front to get into grabbing position and then swing the rest of the way in to grab cubes. The new mechanical design has the arms always moving in complement -- opposite direction but rotated the same amount. If only one arm engages the cube, the other arm will not rotate in and block intake as the old design did.
To do this we added a second pneumatic actuator. The two cylinders' motions are added together to rotate the arms but since they are of different lengths, it's important to know which one is which. For now, let's assume P1 is the longer cylinder (an 8" cylinder) and that it's attached to the existing solenoid used by the harvester. P2 is the shorter cylinder (5") that's attached to a new solenoid that the Pneumatics team needs to install. Assuming there's nothing blocking the arm rotation, this configuration gives us four rotation positions. One of the partially open positions isn't particularly useful and if a cube :
P1 P2 Arm angle Description State Name 0 0 0 Arms pointing straight back FULL OPEN 0 1 103 degrees Arms pointing out to the sides NOT USED 1 0 166 degrees Arms forward and slightly open. PARTIALLY CLOSED 1 1 185 degrees Arms trying to close but cube being held CUBE HELD 1 1 270 Arms folded in front FULL CLOSED I don't see much value in the 103 degree position so I'm ignoring it for now
The states then are:
• "FULL OPEN" : the arms are all the way back alongside the lift. This is the starting position. Let's call this 0 degrees. • "PARTIALLY CLOSED" : the arms are partially forward. This is the common position when trying to grab a cube. This is the case when one of the two pneumatics are closed. In theory, there can be two different positions for "PARTIALLY CLOSED" but I expect that one of the positions will not be exposed by the programmers to simplify driver control.
• "CUBE HELD" : the arms are approximately straight forward. This is the common position when holding a cube. Both pneumatic cylinders are powered in the close state but the cube is preventing full closure.
"FULL CLOSED" : the arms are folded in front of the robot. Both pneumatics are in the full closed state. This can happen if the arms fail to grab a cube or if the driver chooses to close the arm without a cube. Depending on how the arm heights are configured, full closed may only be to the point where the arms hit each other along the centerline of the robot. This position may be useful for knocking cubes out of the pile.
*Thread Reply:* Thanks @Mark Tarlton! The pnuematics are connected to each other then? Does that mean that the system will make itself go to those angles as a result of the solenoid states that you have in the table? (I'm assuming that the numbers under P1 and P2 are solenoid states, correct me if I'm wrong.)
*Thread Reply:* Yes you got it right. The pneumatics are coupled together so that their actions kinda get added together. The table formatting got messed up but you're right ..
I was in a rush, but I intended 0 to imply not actuated and. 1 to be actuated. You don't need to worry about the actual angles, just when to actuate each of the solonoids. Essentially P1 brings the arm from the back position to the ready position, and P2 closes the arms on the cube.
I won't know for sure if the grabber can grab while the harvester is holding the cube. We'll find out once it gets installed.
The motors should be off when arms are stowed in back and otherwise should behave as before.
*Thread Reply:* Great explanation! Thanks again. This looks like it's going to be pretty easy to get into our code. Must see for @ronan_bennett @Darwin Clark
*Thread Reply:* @Ryan_Olney
@Mark Tarlton I thought her we were using a 7in and a 5 in. Not a 8in and 5 in
7 and 5 was the original plan but reuse of the pulleys changed the numbers.
Things are starting to get a little tight on the robot with the increasingly large pneumatics. Also I don’t know how much space we need for the entire cable set up, but things will be very tight now
actually the 7 does work better but the numbers I gave will change slightly. It won't close fully, but the intermediate positions are better with 7"
Do we have 8 inches ? I just ordered from AM four 7 inches.
We have 8 in pneumatics, I remember that we were distinctly missing 7in, but we were finding a lot of 8”.
Based on how we set up the arms they can not close fully without some changes
We had 8" laying around which is why I proposed that number. If 7's are in that's the way to go.
I thought we only had two 7”, not counting the one on the 1st robot, and one with a bent shaft.
We used 7 in pneumatics in the first harvester
with 7" and 5" cylinders the angles are: 0 degrees (straight back), 104 (not used), 145 degrees, and 250 degrees (crossed but not completely folded in front). We should fold the arms back when climbing in this configuration to avoid compressing the arms against the tower.
I just ordered four 7 inches that won't get here till the end of the week.
We just need one 7" per robot so reusing what's already installed will work
Also, in Jack's video it looked like cubes were getting hung up on the middle wheel. It could be that we just need to run the motors faster. If they're still a problem, I'd suggest trying larger diameter wheels there to see if that eases the transition.
I saw the eject is stronger & the arms rotate up... possible to shoot into switch now rather than use the 2-stage place? If so, is there a positional state needed for that?
There may be ways of using the cable tension when grabbing a cube to tilt the arms back up. But it could be we have to lock the tilt in to make the rotation work at all. One thing I'd like to try is to see if the robot can harvest and grab a cube without coming to a stop.
@paul_vibrans What is the length of the cylinder that is powering the integrated harvester grabber?
7, 8 or 9 inches. Whatever we have in inventory. I started working with a Bimba M-048-DP.
@paul_vibrans how are the arms attached to the back brace
@paul_vibrans Yes that size will work well we have 4 M-048-DP it would be great if we could stick with that size.
H-G arms are attached with hex shafts. There is an image of the assembly posted on March 15. It is not that great showing the detail. From the top down there is a double pulley (that could be a single), a tall spacer, the cross bar with bearings, a 15T cog belt pulley, and finally the grabber arm with bearings.
Actually 1.09375 inches. Small keyboard and fat fingers.
Not shown are the posts to which the air cylinder connects. Also needs rubber band springs to squeeze the arms when the air is off. Needs a stronger counterbalance than is on the existing flipper because of the extra weight. Motor power comes from one of the existing harvester controllers. Air comes from the existing grabber solenoid valve(s) with the same logic. Motor logic is the same as the existing harvester when the new H/G is down. It is reversed when the new H/G is up and the cube is on its back.
@whobbs1496 The pulley brackets that I dropped off this morning need to be trimmed with the band saw to match the drawing. Otherwise the vertical legs will interfere. You might get away with trimming just one but that means it only fits one location.
@Mark Tarlton We also have a heating tool to fuse cords. Remind me to show you. I don't know why the team does not use that. We used it last year.
*Thread Reply:* We did use it earlier in the year but the lighters were easier and faster.
@paul_vibrans we're fighting the pulleys and risk not using the new harvester/grabber at GP. What do you think of going with two motors, one on each arm and getting rid of the pulleys and bands altogether?
@Mark Tarlton Is that the only mechanical issue we risk with the new harvester? Is weight a problem?
Can you flare the pulley grooves? They are pretty tight.
Belts seem to be a show stopper at this point. I'm not aware of other problems at this point but we haven't been to run long enough to know. We are also out of time. I think if we had another day we could probably sort out the pulleys but as it stands now, running the old harvester grabber is the default choice
To me, trying two motors and then going all out with counterbalance tubes has the highest chance of success in the time we have. We're also seeing a lot of wear in the double pulley mounted on the motor
Mark, when you say 'counterbalance tubes' what do you mean?
Elastic tubing that acts as a spring so that when the HG is rotated forward, the tubing stretches and helps lift it back. Like we already use to help the lift up and help grabber up
It takes 2 BAG motors and the trick is to keep everything compact.
We can slot the cross tube to the holes for the vertical shafts so we can slide the motor shafts, couplings and shortened vertical shafts in from the side using the same VEX bearing blocks. Angles for resisting motor torque will have to be added to the cross tube.
When we are moving the grabber are they going to be in the same positions.
Before giving up entirely on the revised harvester/grabber we should consider that not being able to eject is caused by lack of motor torque. Changing arm design will not improve this, especially if the outer wheels are 4 inch diameter as are the ones on the H/G that has been built. Can we change to VEX 775PRO motors on the same gearboxes? They weigh only 0.09 pounds more each than BAG motors and have about twice the torque at the same RPM. We also could extend the arms by using smaller drive pulleys on the outer wheels.
No that is not what I am worried about is if we go for the revised version of the grabber for competition. Should we also change the angle of where the grabber grabs the cube.
If that doesn't make sense then I can explain what I mean for the next meeting.
I was thinking since you can't really pick up a cube with the grabber facing downward.
Sort of like trying to cut vegetables by aimlessly stabbing
Did the old harvester with the BAG motors really eject cubes into the vault?
That I know of, the only time we did the vault we had a bit of a challenge ejecting the cube. But I believe we were still able to get 5 or 6 cubes in the vault.
Before we scrap the new harvester/grabber in favor of old harvester arms added to the top of the scissor, we need to know why the old harvester worked and the new H/G does not.
Based on what I was seeing I think it has to do with the shorter arms
When you watched the new harvester attempt to grab cubes it would hit a corner or side and push the cube away rather than suck it in
My analysis is that part of the success of the old harvester was because the harvester attachment to the frame is below the line of action of the intake rollers such that the intake force tends lift the cube, reducing floor friction. This is true in both the intake and eject modes.
The line of action of the new H/G is well below the pivot at the end of the scissor lift. The intake and eject forces tend to rotate the cube into the ground, increasing friction like a self activating brake. The solution to this is to rotate the H/G down on intake so it has a significant lift function and then to rotate the H/G up on eject. Changing to old harvester arms will not change this.
Grabbing cubes off the wall may be a problem because there are only 2 positions for the H/G arm cylinder, open and grab. There needs to be a neutral position where only the spring (rubber tube) force clamps the arms together. This is how 4911's grabber works. 4911 told me they struggled to get just the right tension on their spring and they actually did not have a forced open position, just grab and neutral.
We can add a link easily to the new H/G arms that does not require pulleys and cables to make the arms work in direct opposition. Maximum open position may not be as wide as desired, but it will allow testing.
Large diameter rollers at the ends of the arms probably work against effective gathering because they will be more difficult to squeeze between cubes.
We have not done enough testing with what we have to know what is really going on with any of our systems.
My impression is that the arms and wheels were contacting the cube in front of the CG and so the cube spun away or was pushed back by the arm instead of being pulled in. Especially when the cube was twisted 45 degrees, the arms weren't getting behind corners of the cube. Cube placement had to be very close to square to get a grip on both sides to pull it in. I agree that the forward tilt of the old harvester arms helps lift the cubes some but I don't think that accounts for what we were seeing.
And the old harvester doesn't do the same thing when contacting cubes in the same place?
If that is true then there is an advantage for small diameter intake wheels, smaller than 4".
Seems like some iterative test with different options is needed.
If by "the same place" you mean the position of the cube relative to the robot, then the two arms give very different results. The V1 harvester arms will contact the cubes from the side while the HG arms will hit from the front and the cubes will bounce off. The HG arms need the cube to be very close to the target before they can be used effectively. Close enough that the old grabber could probably be used directly to grab the cube. This small working area of the HG intake is what motivated last night's discussion.
There V1 arm hinge also a mechanical stop that prevents the arms from being pushed to far open. That helps the arms guide the cube home. The distance between the two arm pivots is slightly larger than the cube width and was determined by the gap in the bumpers. I would have liked the pivots farther apart but the grabber /flipper needed the cube as close to the frame as we could get. Wider arm pivots would improve the contact angles between the wheels and the cube making the target volume even larger. The diameter of the wheels also factored into the pivot spacing. Since V2 harvester was using larger wheels, the pivot spacing was increased accordingly. A smaller internal wheel was added in the middle of the arm to guide the cube into the bumper gap. I didn't see the V2 arms run enough to get a sense of their performance relative to the V1 arms.
BTW, the use of belts along the length of the arms was an explicit request from the drive team at the beginning of the season. We added the small wheel simply to help get the cube moving in the right direction and keep them from hanging up on the front corner of the arm. That may still be a better choice.
I meant the position of the end of the arm and the center of gravity of the cube.
The issues of opening limits for the H/G have not been investigated fully because of time diverted by the belt drive.
We can easily add hollow polyurethane belting around the toothed belting to create a traction surface between the H/G wheels. It doesn't need to carry power, just be sticky.
The spacing of the V2 harvester pivots must allow the cube to be gripped securely when the robot gets bounced around. If it is going to be part of a grabber that is.
Ahh. By V2 harvester I was referring to the new version of harvester-only that we built. I used HG to mean the harvester/grabber.
I think constraining arm movement to both limit the amount the arms open and having them move in sync are important to performance. As you pointed out, the HG doesn't need a very wide range of motion so a simple linkage will suffice to keep them in sync.
The H-G cylinder attachment needs to be moved so it can be made stronger. Drill a 0.25 hole through the arm where the cylinder needs to be attached. Drill out the 0.25 hole in the top to 0.5 so a half-inch rod with 1/4-20 tapped holes in each end can be pushed in and attached with a screw in the bottom. Attach the cylinder with screws in the tapped holes in the top.
Two solenoid valves are required, one to fill and vent the cap end of the cylinder to open the arms and one to fill and vent the rod end to close the arms. There are three arm conditions: forced apart, floating free, and forced together.
Add a means to keep the arms from opening too wide. Add a means to keep the arms from closing too much. Add a spring to close the arms when both ends of the cylinder are vented. Adjust the spring tension to grab cubes most effectively.
@paul_vibrans.. good progress with the HG today. They removed the pneumatic cylinder and repositioned the elastic to make more room for the cube coming in. Now performing much better. Kenneth is working on the new pneumatics and linkage layout. The challenge with the linkage is to not block the cube when coming in. They're still working to get the lift counterbalance sorted. If they need to rebuild the arms or crossbar, should they use polycarb box or stick with Alum?
If you look in the grabber folder on fusion you can see my slightly edited version of the new design
Consider leaving room for a centered low (but not too low) camera for assisting acquisition of cubes not in direct sight lines. With HG rotated down for cube acquisition, the proposed camera on the HG arm will be pointed down & unusable for this purpose. The HG cam will really then be for high placing. Not the top top priority I know, but good to keep in mind.
@whobbs1496 @james_sovick Did the new elastics work ?
We were in the process of re stretching them when we left
Do you think the robot is ready for programmers by Friday afternoon ?
I think it will be. But I don’t know how complicated the pneumatic mount will be
The pneumatics are almost finished. We probably won't get the arm linkage working, there's not enough space but it isn't essential and can be added later if we figure it out. For me, the big question is with scissor lift and elastics. Finding the right balance so that it can go up with a cube and come down empty may be tricky.
As of today we’ve been working on this new variant of the harvester grabber for 5 days now, and a new harvester for the past 2 weeks. We have not had a single design that has been successful as of today, with issues continuing to plague us. This week we have been taking Paul’s design and modifying it with an end date of Friday to determine whether we will be using it or not at Portland.
On Thursday we got the pneumatic mounted and up and running. It looks promising with a quicker cube acquisition and securing system that seemed to be able to pick cubes up well. However we became plagued with the problem we still haven’t solved, the scissor lift will still not go up with the new combined Harvester Grabber System. It is roughly twice the weight of the previous grabber. Will has spent the last 4 days working with what seems to be every possible set of elastics, and below is a list of different things that we have tried.
The Donated Stress bands From Chris Rininger The New Therabands which are currently being used The black elastic tubing that we have used in the past, which is being used in conjunction with the Therabands
All have failed to resolve the issue of the scissor lift not going up or coming down if it does go up, or neither.
Today we have tried adding in a steel cable to the elastics to limit the use of some of the elastics similar to the loops we have used in the past. This may have worked a little as it seemed to get us close to something, but we’re still not there yet. Currently we have a plan to try and do two of these to get a stronger initial push, but then decrease once the scissor is moving up. I would like to try this in the first hour of Monday.
Other ideas that have been floated are
Using a pneumatic or a motor as a kicker to give the scissor lift a little initial boost. - Mechanics don’t think we have enough time to do this (at the meeting today it was decided we abandon the combined system when we had to go to this) Power Down --Tested, along with unseating the pneumatics it doesn't provide enough power (admittedly we used the pneumatics test rig, but almost nothing happened so it is no longer viable) Using a motor to pull the scissor lift down --Three different people have thought of this, but we don’t have the time to design the system like the pneumatic kicker. Removing weight from the Harvester Grabber - No, anything we remove right now is negligible in the mass of the grabber. We’re currently looking for a simple solution that doesn’t require a lot of mechanical construction time. So if anyone has an idea, or wants more details on what has been tried, please say something that might get us heading in a direction.
But if we cannot get any of these ideas working in the first two hours on Monday, I would say we need to run with what we have currently, as we have a reliable and tested solution for the robot. While the older design is less than ideal, the drive team has experience using it, and we know it works, rather than trusting in a new design that may or may not work once we put it on the primary robot. Even if we get this new combined system functional there is no guarantee that it will increase our performance, especially if we are still debugging. I would say that we have almost exhausted this design now, but there are still some simple things worth trying before we end it just yet. My current thought is that we need to give it up, and focus on improving our driving as much as possible.
So please give us ideas on how to improve this design, or why we shouldn't leave it.
We’ve looked at them, mostly throwing them out because we didn’t have anything big enough in the right force range, most of our longest constant force springs require too much force to get them moving. I might want to play with the idea a bit more, but both Will and my first instinct was that we were uncertain how to use them in this situation. Also, we don’t exactly want a constant force all the way up, more force to start the scissor lift going at the bottom, but less at the top, in order for the scissor lift to come back down.
Have you already thought about doing lighting holes on the harvester(if there is even space on the harvester to do that)
Lightings holes as said above won’t make a meaningful difference.
I was just reading this about springs: https://www.sciencebuddies.org/science-fair-projects/references/linear-nonlinear-springs-tutorial#nonlinearsprings It seems like a more linear metal spring could work better than the latex if I’m thinking about it right. In the bag of stuff I donated I think I included an old school arm exerciser that uses metal springs instead of latex. If not, have it here.
I remember seeing something of the type in there, might we worth trying, but I think our current rational is that elastics are better, because they are more easily tuned than springs
Could use metal spring for most of the force and weaker elastics to tune - just brainstorming
Something to try certainly, but we need to have a point where we’re ending this, as I don’t think we can get this working by Portland.
Yes, at some point it’ll be better to seek gains elsewhere... driving practice, refining in-match strategy (we’ve lost several close matches - how might we have won them?), etc. Sounds like you all think we might be there
I think a spring installation alongside the scissor that pulls mostly upward is the solution. I will create a sketch and post it.
We haven’t tried that, I think. If it doesn’t get in the way with everything else that we have on the scissor.
My one worry with the spring you brought is it is very stiff and also very long and may not fit when the scissor lift is all the way up
Not saying that's the right spring to try to use - coach may have others (for demonstrating oscillation, for example). I wonder what a segment of last year's gear peg spring would do. As you and Kenneth have indicated though, there's limited time to keep trying different things, and driving practice with whatever system we end up keeping (original or new) would also help reduce cycle times.
This explains what we were trying yesterday. This is a graph showing the forces that the elastic bands need to supply to lift the scissor and cube. It gives an idea of how much force is required to get the lift started. I think we were on the right track with very short bands connected to a cable to provide forces only during the first portion of the lift. What this graph shows is that a LOT of force is required at the start. I think this second band assembly should be something like 10" of cable and then multiple loops of band with the length of each loop being about 3". When the bands are fully stretched, they will be about 9" long with 10" of cable giving a total length of 19" . They provide about 30 lbs of force per loop. In addition, multiple loops that run the full length of the lift base are still required to provide assistance over the entire distance of the lift. In this case, the loops are about 6" or 7" long so that when stretched, they will be about 20" in length.
I have "V's" and steel shafts to try with more vertical lift springs. There is one for each robot. I need to know where they should be delivered.
A band across the ends of a V lifts almost vertically when the lift is bottomed so is much more effective than a horizontal band under the scissor. They should work together to get the performance we need.
That looks great! What did you have in mind for "tabs" that the Vs are screwed to? Are they Located that sit on top of the rails on either side of the Vs?
*Thread Reply:* (stupid autocorrect). Are they Ls that sit on top of the rails on either side of the rails?
The tabs lap over the vertical outsides of the rails. There are removable plastic guards there now. They would have to come off.
The edge of the V is supposed to align with the end of the rail but in any case the shaft should be centered in the V with the scissor down. It is the middle shaft, not the lowest shaft.
BTW, we haven't experimented with the elastics that help the flipper. I expect that will need help with the extra weight as well. Any thoughts on how to help there?
More bands on the flipper should do the trick.
I need to get the V's to someone today. I am not able to attend any robotics meetings before Portland.
I can come by and pick them up now if you like. I'm going in Monday
@paul_vibrans On Mon 4/2 the team will be meeting from 10 am to 4 pm . You can drop off if this works for you . Otherwise maybe @whobbs1496 can pickup from your house . Let us know . Thanks
@Mark Tarlton Picking them up now is best. 9034 Springwood.
I forgot to mention that the V's will need to have some pockets drilled into the bottom to clear screw heads.
We'll evaluate the parts Monday on whether we want to keep on working with improving the scissor lift or not. We need to have a conversation before Monday on when we want to stop working on the new grabber harvester, or do we keep on trying right up until the last minute.
The new harvester/grabber is working well enough that we will try to use it at Portland (assuming the crooks that stole the robot yesterday return it by then). The elastics that assist the lift going up and down are still a bit tricky to get adjusted properly but the team was able to demonstrate most of the necessary behaviors. Hopefully we'll be able to dial it in at the competition.
Yes, it's still pretty rough but we can get better with time. We should be more competitive than at GP.
Kudos to the student team for putting in a lot of hours and having the guts to attempt a new intake-grabber design for Portland. It was a lot of coordinated effort between the sub-teams and continues to be plenty of trial-and-error to get the design to work smoothly. In the end I believe we have a faster and more robust cube delivery system. Even if the redesign of THEMIS does not work out perfectly I applaud the team for trying (and this would not have been possible without the second robot). Paul: the “V’s” are a clever solution. Mark & Chris: thanks for pursuing improvements.
@Terry +1 Congrats to the team for facing challenges head on & endeavoring to improve! And also great job building the second robot that has enabled these efforts. The team has taken itself to a new level this year in my opinion. I'm excited for Portland!
I'm super proud of you guys. Good job, lets make the most of what we've got!