View Full Version : paper: [Ri3D] The GreenHorns 2016 Present: Rivvet
Ginger Power
14-01-2016, 00:41
Thread created automatically to discuss a document in CD-Media.
[Ri3D] The GreenHorns 2016 Present: Rivvet (http://www.chiefdelphi.com/media/papers/3216?) by Ginger Power
Post any questions about The GreenHorns' 2016 Ri3D Robot here. Our reveal video will be posted here tomorrow before 3PM. The perfectionists on our team were so broken by Ri3D that they needed to make up for it by taking the time to make a perfect reveal video.
Ginger Power
14-01-2016, 16:24
Sorry we're a little late, but here is our 2016
Reveal Video (https://m.youtube.com/watch?feature=youtu.be&v=jRWDFRGVPMU)
Caleb Sykes
14-01-2016, 16:36
Sorry we're a little late, but here is our 2016
Reveal Video (https://m.youtube.com/watch?feature=youtu.be&v=jRWDFRGVPMU)
How difficult was turning for you guys with a 4WD? Did you notice that the batteries were draining quickly?
Ginger Power
14-01-2016, 16:42
How difficult was turning for you guys with a 4WD? Did you notice that the batteries were draining quickly?
I wouldn't want a 4 wheel drive robot. Turning on that concrete wasn't terrible, but carpet would've been a challenge. I've said before, if we had access to 6 pneumatic wheels that's what we would've done.
Tom Line
14-01-2016, 18:04
Ginger,
Did you find with longer shots that the ball would sometimes travel left or right, and not always straight? We've looked at a shooter design like yours and one of the considerations is spinning both sides at the same speed.
Nice job on the robot. In 3 days you made something that will do well at many districts!
Ginger Power
14-01-2016, 18:36
Ginger,
Did you find with longer shots that the ball would sometimes travel left or right, and not always straight? We've looked at a shooter design like yours and one of the considerations is spinning both sides at the same speed.
Nice job on the robot. In 3 days you made something that will do well at many districts!
The ball flew very straight in our testing. The real problem with shooting from far away will be the size of the goal. As others have noted it's the smallest goal relative to game object in FRC history. Given how safe shooting from up against the tower seems to be, I think that's the way to go. I think it would be capable of scoring from the outer works, but I haven't tested it enough to give you a guarantee.
Would a pair of mini-CIM's be likely to be sufficient to get similar results for the shooter?
Ginger Power
16-01-2016, 12:44
Reveal Video (https://m.youtube.com/watch?v=jRWDFRGVPMU)
**Fixed the link**
bbbiggers
17-01-2016, 16:26
Does the shooter use two Mini - CIM's ?
Does the shooter use two Mini - CIM's ?
Have you read their whitepaper linked to in the initial post?
Thread created automatically to discuss a document in CD-Media.
[Ri3D] The GreenHorns 2016 Present: Rivvet (http://www.chiefdelphi.com/media/papers/3216?) by Ginger Power
Ginger Power
17-01-2016, 16:37
Does the shooter use two Mini - CIM's ?
As Hallry said, I've written a white paper about our design that details Rivvet. We did use mini CIMs at 1:1, but if I had any motors/gearboxes available to me I'd use a 775pro at a 2:1 ratio driving 4" wheels. This will give you more torque as well as more speed. Then you would have the ability to limit the speed via code and make it more robust to changes in the battery voltage.
trumpthero786
17-01-2016, 16:59
As Hallry said, I've written a white paper about our design that details Rivvet. We did use mini CIMs at 1:1, but if I had any motors/gearboxes available to me I'd use a 775pro at a 2:1 ratio driving 4" wheels. This will give you more torque as well as more speed. Then you would have the ability to limit the speed via code and make it more robust to changes in the battery voltage.
I apologize if this is a silly question, as I am not a gearbox expert. You say that you would use the 775pro at a 2:1 ratio, but the versa planetaries that VEX sells start at 3:1. How would you get 2:1?
Ginger Power
17-01-2016, 17:16
Not with a versa planetary :p I haven't done a 2:1 ratio before, so I'm not sure about the optimal way to do it. One way I can think of would be to drive a 16 tooth sprocket/belt off the motor and run a chain/belt to a 32 tooth sprocket/pulley on the intake wheels.
You could also just run the 775pro at 3:1 in a versa, which would still be faster and more powerful than mini CIMs at 1:1
trumpthero786
17-01-2016, 17:22
Not with a versa planetary :p I haven't done a 2:1 ratio before, so I'm not sure about the optimal way to do it. One way I can think of would be to drive a 16 tooth sprocket/belt off the motor and run a chain/belt to a 32 tooth sprocket/pulley on the intake wheels.
You could also just run the 775pro at 3:1 in a versa, which would still be faster and more powerful than mini CIMs at 1:1
That makes sense. Thanks for the quick and insightful help!
Ginger Power
17-01-2016, 17:27
That makes sense. Thanks for the quick and insightful help!
No problem! If you have any questions just pm me here. I check Chief Delphi more often than I check Facebook and Twitter combined. It's probably unhealthy actually...
trumpthero786
18-01-2016, 15:27
No problem! If you have any questions just pm me here. I check Chief Delphi more often than I check Facebook and Twitter combined. It's probably unhealthy actually...
I actually do have another question :p For the articulating arm, why did you use a pair of gears to go from the gearbox to the arm itself? Is it necessary? Could you have gone from the Planetary directly into the arm?
popvideo
19-01-2016, 20:48
Could I get the measurements of your robot, please? Also, what defenses was this tested against?
Great design!
thanks,
popvideo
GRA-V mentor
Ginger Power
19-01-2016, 20:54
I actually do have another question :p For the articulating arm, why did you use a pair of gears to go from the gearbox to the arm itself? Is it necessary? Could you have gone from the Planetary directly into the arm?
We were worried about putting any weird torques on the versa planetary. There is definitely a way to direct drive using the shaft (a hex to hex coupler...). Actually, if we had access to a hex to hex coupler that's probably what we would've done. The geometry also worked out nicely with the gears.
Ginger Power
19-01-2016, 20:57
Could I get the measurements of your robot, please? Also, what defenses was this tested against?
Great design!
thanks,
popvideo
GRA-V mentor
I'm doing 18 credits right now and I'm also 200 miles away from the robot... if you could specify what dimensions you're looking for I'd be glad to get them. I was hoping to CAD the robot, but time is valuable when you're in college.
Our drivetrain was successfully tested against the rough terrain, low bar, and ramparts. The others might be possible, but we didn't want to damage our pathetically built drivetrain. I'd recommend 6 pneumatic wheels with a drop center if you have access to them.
Quick question regarding the shooter. Did you guys have an issue with the shooter back driving the cim that controlled the elevation of the shooter? Example would be you had shooter positioned at say 35* did the shooter want to drop ?
Ginger Power
19-01-2016, 21:58
The turret would begin to drop when the angle between the ground and the shooter approached ~45 degrees as an estimate. I think if you're debating whether or not to add a mechanical brake to the shooter, I would recommend it. The reason I say this other than the back driving, the shooter would rock while driving, and especially while traversing obstacles. Stabilizing that seems like a good idea.
On the other hand... you could just ignore the rocking and then stall the articulation motor. Up to you!
We came up with a design on kickoff day for the shooter that was almost identical to the one you made, only real difference is we are using 775 pros with a 3:1 versa planetary, and the original set up for the ball to roll in on and be supported was going to be aluminum tube, but once I saw the plastic tubes you used I started looking for similar material. I am thinking the plastic covers for fluorescent lights will be the ticket.
Ginger Power
19-01-2016, 22:30
We came up with a design on kickoff day for the shooter that was almost identical to the one you made, only real difference is we are using 775 pros with a 3:1 versa planetary, and the original set up for the ball to roll in on and be supported was going to be aluminum tube, but once I saw the plastic tubes you used I started looking for similar material. I am thinking the plastic covers for fluorescent lights will be the ticket.
We desperately wanted to throw some LEDs in the tubes (http://www.mcmaster.com/#standard-polycarbonate-hollow-tubing/=10ra5x7), but time is a killer. Going back, if we had access to 775pro's that's exactly what I would've done. I'm not sure if I'd do 3:1 or 2:1 though... would've needed to do some testing!
On the tubes do you know what od you went with off hand?
On the 2:1 vs 3:1 we figured that a mini cim is spinning around 6000 rpm so a 775 at 3:1 is also about 6000 rpm so should work out pretty much the same as you guys ended up with. We are going to use the versa planetary encoder to measure and set the rpm using a talon srx.
And leds in the tubes are a given!
Ginger Power
19-01-2016, 23:02
I bought 1.5" OD with 1.125" ID. I actually bought the tubes because I thought they might have utility as a roller intake, plus we would be able to press fit a bearing into the tube. The rigidity was more than satisfactory and they weigh barely anything.
Awesome. Thanks for the tips.
Did you work with 4" wheels on the shooter right from the start? If so, why did you go with this size? If you tried other sizes what were your findings?
mitchklong
25-01-2016, 11:19
I can tell in the video that there is a device that induces the ball into the fly wheels. It looks like it might be that little gearmotor with the microcontroler, but its covered in tape and I cant tell. What is it and how did you get it to work ?
Ginger Power
25-01-2016, 11:35
Did you work with 4" wheels on the shooter right from the start? If so, why did you go with this size? If you tried other sizes what were your findings?
Sorry for the delayed reply! I must've missed your post. We used 6" wheels right away and didn't like the profile of the shooter, it was much bigger than we wanted. So we tried 4" wheels and it got us a shot that we liked with a profile that would work. I could see a team shrinking the profile even more and running 3" or smaller wheels at a higher RPM.
Ginger Power
25-01-2016, 11:58
I can tell in the video that there is a device that induces the ball into the fly wheels. It looks like it might be that little gearmotor with the microcontroler, but its covered in tape and I cant tell. What is it and how did you get it to work ?
We actually used a rotary servo (http://www.andymark.com/product-p/am-3217.htm) like this one from AndyMark. We mounted a 4" piece of aluminum to it that acted as the lever arm. Our setup was slightly underpowered, but that could easily be remedied by using a stronger servo, and/or a shorter lever arm. Just be sure to do the math. Your trigger should be able to reliably push the ball into your shooter when your shooter is at any angle.
waterbott
25-01-2016, 12:14
The turret would begin to drop when the angle between the ground and the shooter approached ~45 degrees as an estimate. I think if you're debating whether or not to add a mechanical brake to the shooter, I would recommend it. The reason I say this other than the back driving, the shooter would rock while driving, and especially while traversing obstacles. Stabilizing that seems like a good idea.
On the other hand... you could just ignore the rocking and then stall the articulation motor. Up to you!
What would you recommend as a means of dampening the rocking? Our team was contemplating using gas springs, an air cylinder, or even just elastic cord.
Ginger Power
25-01-2016, 12:21
What would you recommend as a means of dampening the rocking? Our team was contemplating using gas springs, an air cylinder, or even just elastic cord.
All of the above would be great for that purpose. It all depends on the geometry of your robot and where you can fit everything. Any of the options you listed can be as effective as the others if implemented correctly.
We actually used a rotary servo (http://www.andymark.com/product-p/am-3217.htm) like this one from AndyMark. We mounted a 4" piece of aluminum to it that acted as the lever arm. Our setup was slightly underpowered, but that could easily be remedied by using a stronger servo, and/or a shorter lever arm. Just be sure to do the math. Your trigger should be able to reliably push the ball into your shooter when your shooter is at any angle.
We have one of the REV Smart Servo's and the torque is good, but the speed is very slow - much slower than what we see on the robot you built. Were you using this particular servo, or one like it? Sec/60 specs on what you used would be great if you have it. Thank you guys so much for answering all these questions! Very much appreciated!
On the tubes do you know what od you went with off hand?
On the 2:1 vs 3:1 we figured that a mini cim is spinning around 6000 rpm so a 775 at 3:1 is also about 6000 rpm so should work out pretty much the same as you guys ended up with. We are going to use the versa planetary encoder to measure and set the rpm using a talon srx.
And leds in the tubes are a given!
Isn't the Versa Planetary encoder using the CTRE magnet that needs to be bored into a CIM shaft? Wondering how that would work on a 775 shaft.
Ginger Power
27-01-2016, 17:08
We have one of the REV Smart Servo's and the torque is good, but the speed is very slow - much slower than what we see on the robot you built. Were you using this particular servo, or one like it? Sec/60 specs on what you used would be great if you have it. Thank you guys so much for answering all these questions! Very much appreciated!
It wasn't exactly the REV Smart Servo, but it was similar in nature. I don't have exact specs because I'm not sure of the model, but I'm going to make an educated guess that ours has a lower gear ratio. Ours was also underpowered so that makes sense.
If you've got plenty if power but a low speed, you can increase the length of the lever that you attach to the servo. Sorry I don't have more info! That's the disadvantage of being 200 miles away from the robot.
Ginger Power
28-01-2016, 00:57
I'd like to clarify something that has bothered me. To articulate our shooter mechanism we used two 10:1 versa planetary stages for a 100:1 overall reduction. This is way outside the recommendations from Vex's Load Ratings Guide (http://link.vex.com/vexpro/pdf/VersaPlanetary-LoadRatings) for a Cim motor. I apologize to any teams who may have gotten the idea to do this from us.
A much safer solution for articulating the shooting mechanism would be to use a Mini Cim motor in a versa planetary at a 108:1 ratio. This can be safely done by using a 3:1, 4:1, and 9:1 stacked up. It's still within Vex's recommendations and it provides a similar amount of torque.
Another solution would be to use the Gem 500 (http://www.andymark.com/product-p/am-0401.htm) gearbox from AndyMark. I believe the joint at which the shooter articulates should be the strongest part of the robot. I will say from my experience with the Gem 500, it will have no problems, it's a tank.
Again, sorry if anybody copied this aspect of our design. I'd like to blame the sleep deprivation, but it was a result of our lack of experience. We've learned from our mistake and will do better next time!
For the shooter, is it better to run a 3:1 gearbox with a 775pro then to do what you did with the mini-cims? Even though they will both be at around 6000 rpm?
Ginger Power
28-01-2016, 20:03
For the shooter, is it better to run a 3:1 gearbox with a 775pro then to do what you did with the mini-cims? Even though they will both be at around 6000 rpm?
The problem with what we did with the mini CIMs is that they were at their max speed. This means the speed of the wheels is completely dependent upon the battery voltage. 775pro's are lighter, and can be run faster than mini CIMs in this application. It is my belief that you will see better results with the 775pro's.
gpetilli
29-01-2016, 09:57
Not with a versa planetary :p I haven't done a 2:1 ratio before, so I'm not sure about the optimal way to do it. One way I can think of would be to drive a 16 tooth sprocket/belt off the motor and run a chain/belt to a 32 tooth sprocket/pulley on the intake wheels.
You could also just run the 775pro at 3:1 in a versa, which would still be faster and more powerful than mini CIMs at 1:1
I would recommend the Vex Cim-ile at 2.4:1 with the 775Pro
http://www.vexrobotics.com/vexpro/motion/gearboxes/cim-ile.html
Actually, this has the same exact mounting as the miniCIM so you could swap it out directly and do the comparison.
mustangs2647
29-01-2016, 10:09
So how tall was your robot when the shooter was down?
Ginger Power
29-01-2016, 10:25
So how tall was your robot when the shooter was down?
~14" We never had to worry about clearing the low bar. The top of the frame was 9" off the ground so we had high ground clearance with a low profile overall.
gpetilli
29-01-2016, 10:49
Isn't the Versa Planetary encoder using the CTRE magnet that needs to be bored into a CIM shaft? Wondering how that would work on a 775 shaft.
I dont think it is legal (or safe) to bore into a motor shaft. I hope you meant the versa planetary 1/2 hex output shaft - which would be fine.
I would recommend the Vex Cim-ile at 2.4:1 with the 775Pro
http://www.vexrobotics.com/vexpro/motion/gearboxes/cim-ile.html
Actually, this has the same exact mounting as the miniCIM so you could swap it out directly and do the comparison.
In the link it says the ratio for a 775pro with this gearbox is 12:29. Is that correct ratio or would we have to change the gears, and if the latter, where would I get the gears and which ones?
(I am really sorry if this is a stupid question. We are a rookie team trying to figure out what we are doing! haha)
Coach Seb
01-02-2016, 09:44
Thank you so much for this robot design, we build a shooter very similar to yours and it shoot very well... we also are using 775pro with 3:1 versa.
My question is around the elevation, maybe our framing is too heavy, but what would be a good recommendation to hold the shooter at the proper angle position mechanically while trying to aim at the target?
Ginger Power
01-02-2016, 10:06
There are a number of ways to articulate the shooter, and a number if those are covered in this thread, and in the white paper. The way I would recommend would be to get a gearbox with a very high gear ratio >100:1 and use that to drive a chain and sprocket setup which is connected to your arm. Two gearboxes I would recommend are the Versa Planetary from Vexpro, and the Gem 500 from AndyMark. Both can give you enough reduction.
You could also use a dart linear actuator to pivot the shooter. There are tons of possibilities!
Coach Seb
01-02-2016, 10:27
Thank you!
How would the position be steady with the gearbox? last we tried, we could not hold it in position....
Rookie team here... learning all to put this together ... appreciate your help..
Ginger Power
01-02-2016, 11:07
Thank you!
How would the position be steady with the gearbox? last we tried, we could not hold it in position....
Rookie team here... learning all to put this together ... appreciate your help..
The gearbox will be difficult to back drive if it's at a high enough ratio. That probably won't be enough to keep it from wobbling completely. In order to do that I think you'll need to use a brake mechanism. Another solution could be to employ a hard stop for your shooter at an angle that you know works from a certain spot on the floor. This hard stop will help reduce wobble. An additional solution (which could be employed in combination with either of the others) would be to stall your motors in a way that holds your shooter at a desirable angle. This won't prevent wobble, but it should help.
In the link it says the ratio for a 775pro with this gearbox is 12:29. Is that correct ratio or would we have to change the gears, and if the latter, where would I get the gears and which ones?
(I am really sorry if this is a stupid question. We are a rookie team trying to figure out what we are doing! haha)
The 12 and 29 are the number of teeth on the gears
so, 29/12=2.41
gpetilli
08-02-2016, 09:24
In the link it says the ratio for a 775pro with this gearbox is 12:29. Is that correct ratio or would we have to change the gears, and if the latter, where would I get the gears and which ones?
(I am really sorry if this is a stupid question. We are a rookie team trying to figure out what we are doing! haha)
There are no stupid questions. 12:29 is correct, which if you do the math (29/12 ~ 2.4:1). I believe you were originally looking for something near 2:1. Others have suggested versaplanetary 3:1, which is also a fine option - possibly easier to mount. You need to look at both in CAD and see which works better for you. The 2.4:1 will spin your flywheel 3.0/2.4 or 25% faster.
greg7mdp
21-02-2016, 20:35
Hi Ginger Power,
Rivvet is awesome, and our design is quite inspired by it. Unfortunately we are running way behind, and are having great problems with the articulation to rotate the shooting mechanism upwards. Since we want to be able to have the shooter operated via vision software, we need to be able to set a precise (and repeatable) upwards angle. Our current motors just tilt the shooter all the way up, and drop it when the motors stop. Also we have a motor on each side, and I don't think it is the best setup.
We have a GEM500 gearbox (unused yet), and I think we want to put a strong servo on it for repeatable settings. Is that a good idea? If yes what kind of servo could we use? And also how to extend the gearbox shaft so that it can be attached to the other side (maybe with a bearing?)
I apologize for the basic questions - I know nearly nothing about motors since I just started helping my son's team about 4 weeks ago (I'm a software guy).
Ginger Power
21-02-2016, 20:57
Thanks for the kind words. My first and strongest recommendation would be to abandon vision tracking and auto aimimg. 99% of teams are better off picking a spot on the floor (hard stopped against the tower, at the edge of the batter, or from the out works) and using basic robot-centric sensors to aim. For example a gyro on the arm to tell you what angle it's at. When the shooter goes all the way down you could have it reset the gyro so as to avoid gyro drift. Have a known angle for each shooting position. You will be more accurate as a result of this.
If you implement the Gem 500 with 3 or 4 stages on it, and have an external reduction via chain, you really shouldn't have trouble with back driving the gearbox. I would highly recommend plugging a Cim or mini cim motor into the Gem rather than a servo.
I'm sorry but I'm not sure I know what you mean by "extending the gearbox shaft to reach the other side". If you could explain what you mean by that I'd be happy to other my opinion on a solution.
Thanks for your design we used it for our shooter, love it... I'm excited to see what we can do with vision and will see if we can build in redundancy without vision, we added to the front mini tomahawks from TeamRedracted . We trashed our shooter yeserday in practice but rebuilt it stronger today. If we get our scaler working we should be in good shape for San Diego week 1.
We are using bag and larger wheels.
greg7mdp
21-02-2016, 22:17
Thanks a lot for the fast answer, much appreciated. I wish I had a picture of our shooting mechanism to post, as I don't know how to explain it well. I'll try to get one tomorrow evening and post it
I understand what you say about vision tracking, but I think we could make it work if we had an accurate way to set the shooting arm vertical angle. I somewhat have it working on my PC, processing images (from a webcam) with OpenCV in one process and sending instructions using network tables to another process running a robot simulation with robotpy. I also run the same OpenCV image processing running on a raspberry pi mounted on the robot. However, without an accurate way to set the shooting arm vertical angle, this will never work.
Why isn't a servo a good solution? I thought it would allow for repeatable settings, but clearly I must be missing a big issue.
Ginger Power
21-02-2016, 23:08
Anything is possible with a dedicated software mentor and some time. I would never discourage a team from trying new things and growing their knowledge base. If vision doesn't work how you want, just be sure to have contigency plan!
I'd love to comment on a picture so I can give it a fair assessment. Here's what I can say: If your shooter is mounted on a live axle (for example a hex axle that rotates with the shooter) then you should easily be able to drive the shooter up and down via a hex sprocket on the live axle, and another sprocket on your Gem 500. I'd use #35 chain. You shouldn't need to direct drive the shooter axle, and doing so would be harder to implement I think.
There are benefits to being addicted to Chief Delphi. Quick replies being one of them.
cadandcookies
21-02-2016, 23:37
Just wanted to drop in and say thanks-- the Greenhorns robot was a key inspiration for 2667 this year, and I don't think we'd have built the same robot without yours existing.
We look forward to seeing 4607 at North Star Regional!
Ginger Power
22-02-2016, 01:02
Just wanted to drop in and say thanks-- the Greenhorns robot was a key inspiration for 2667 this year, and I don't think we'd have built the same robot without yours existing.
We look forward to seeing 4607 at North Star Regional!
Thanks Nick! That means a lot. Snow Problem's robot was amazing this year as well. I think we're really starting to figure out this whole 3 days thing. Too bad our teams can't do the same thing :p
greg7mdp
22-02-2016, 22:19
Thanks again for all the help. Here is a picture of our robot. It may look familiar... imitation is the sincerest form of flattery :-). We may have a solution for tilting our shooter, as we had some help today from a motor expert.
http://imgur.com/h1ZYeAR
vBulletin® v3.6.4, Copyright ©2000-2017, Jelsoft Enterprises Ltd.