FRC 95 The Grasshoppers 2024 Build Thread

[scottyoung]

@JamesCH95

The augers worked shockingly well (inspired by another team in 2023, but I can’t recall their name or number)

might that be us?

While we haven’t tested augers in 2024 at all, they were definitely a finicky system for us last year with centering cones.

Small, seemingly inconsequential changes to the setup produced significantly different results, including shaft stiffness, auger wear (even small amounts), and specific auger parameters (pitch/stiffness/major&minor diameters). We ended up printing a ton of different augers and arriving at a system that worked through a lot of brute force experimentation. Sometimes inexplicably it would just stop working mid-event and a few of us who had gotten good at knowing “what the intake wanted” would have to tweak things until it became happy again. You can pretty much track how the intake was behaving last year with our rise and fall of game pieces scored/match.

All that to say, I’m hoping that the auger exploration ends up being less involved for you than it was for us, and if you want to pick our brains about anything we tested just let us know. It’s definitely a cool system and can work, but it required a pretty deep dive into all of the system parameters to get and keep it running well. Of course the cone/TPU auger interaction may be totally different than the soft note/harder auger interaction, so this may not apply to what you end up seeing.

PS: Thanks for an awesome thread this year as always, 4909 has definitely been watching closely and has been exploring some ideas that you have put out here. Looking forward to competing together week 1 at GSD!

[Schmools]

Not sure how useful this is especially since your printed augers seem to be working well already but we made a similiar augur out of blue roughtop back in 2016 (it worked ok, probably required more tuning of compression than we really gave it make it work better before we turned to other centering mechanisms) - I guess the only advantage I could see is that maybe the roughtop augur is more grippy than the printed one? Again not sure that actually would be helpful in this case but I remember thinking it was a clever system at the time.

[JamesCH95]

Unfortunately not yet.

[JamesCH95]

@scottyoung

might that be us?

Yes! Thanks for the advice and insight on your experiences last year.

Our aim is to use augers and simple deflectors like Spectrum and the Zebracorns have demonstrated. Hopefully that doesn’t require serious tweaking on a regular basis.

[nicbamb12]

@mrnoble

This is very similar to what we’ve done when using the tube nuts (spider nuts, we call them) in the past. thread it onto a particularly long bolt, and then tap the bolt end to set the nut in the tube at the desired depth. Use the length (like, 3") as a lever arm to adjust the set if necessary. Back out the bolt when finished.

for someone without a lathe, how might you suggest inserting the tube nuts, any idea? We have a pretty basic arbor press, but no collets or anything. Any ideas would be appreciated (from anyone), thank you so much!

[chadr03]

We have historically done it by just partially threading a bolt in the tube nut then tapping on it with a mallet/hammer. It will often start to go sideways then you have to force it back to center by prying or hammer in a certain direction and tap again. Not the easiest thing to do without some sort of jig, but not impossible.

[nicbamb12]

That makes sense, I’ll have my assembly team try something along those lines (we are no strangers to janky setups)

[AllenGregoryIV]

@nicbamb12

for someone without a lathe, how might you suggest inserting the tube nuts, any idea?

When we have to do it without a lathe my preferred setup is. A 1" or longer fully threaded bolt, a 1/4" - 1/2" spacer (depending on how deep you want to set the nut into the tube), and a flange washer bigger than the OD of the tube you are inserting the nut into.

Put the washer on the bolt, then the spacer, then the tube nut. With a mallet hammer the nut into the tube until the washer sits flush with the tube. This should give you a consistent tube nut depth and a relatively straight install as long as you don’t bend the washer. Using a thicker piece of plate with a hole in it would work as well instead of a washer.

[nicbamb12]

We tried this method today, and it worked like a charm! Thank you so so much!

[JamesCH95]

Let’s talk about statistics Part III!

@StephenWitwick, @Wesley, @JWC_95 , myself, and a few other Grasshoppers did a big experiment today to evaluate different shooter wheel configuration. It includes high speed video, pictures, and lots of data. This is pretty firmly the biggest and most useful experiment we have ever run and published.

The intent for this experiment was to capture the approximate performance differences between major shooter archetypes, NOT to provide an optimal design for one particular choice. There are SO MANY details that I cannot capture in this post, but if you see something say something and post about it! A lot of this experimental setup is not ideal, we know that, and we would expect all designs to improve consistency if they were executed more cleanly. We also didn’t try to control every single variable independently. So, yeah, we change wheels and the inertia changes in the system. Use your best judgement when interpreting or applying our results.

Experimental Setup
Shooter specs:

• 1x MiniCim per axle wired to 12v nominal
• Fresh battery every 1-2 configurations
• Shooter positioned 30° from horizontal
• 11.5ft range
• 4in Colson wheels, 3in 35A stealth wheels
• 1/2in nominal compression
• 1x worn out legal Note, 1x fresh legal note (fresh note used in all HSVs), 5x off-brand notes; all notes cycled ~3 times per configuration
• Shooter bolted to platform with 120lbs of sand for recoil absorption, wheels taped to the floor
• Window wire grid is 1/2in on center
• HSV shot at 1080P/1kfps, so if you watch if frame-by-frame each frame is 1ms
• Positional data collected from tape marks via pixel counting, each image calibrated by counting 12in near the center of shot spread from window wires
• Shooter plates modeled here

Configuration 1
2x Colson wheels offset to one side
This was tricky to feed well, the note wanted to pop out of the wheels, as seen in HSV 1 - flub clip.
[no direct picture taken, see setup pic above]

Shot spread:

Configuration 2
4x Colson wheels offset to one side
Easier to feed, note didn’t seem to ‘pop out of the side’ as much
Noticeably more inertia in shooter

Shot spread:

Configuration 3
4x Colson wheels spread out for a ‘no spin’ shot
Wheels spacing is admittedly not ideal for this configuration
Fed okay, felt easy-ish to slip left and right

Shot spread:

Configuration 4
2x Colsons, 2x Stealth, tapered axles (both sides 1/2in compression nominal)
Easier to feed
Noticably more power
A few flubbed shots traced to off-brand note (green one), eliminated as outliers
Image below is not an optical illusion

Shot spread:

Configuration 5
4x Colsons with 1/2in nominal compression, 4x Stealth with 1/2in nominal gap
Feeding seemed easy
Substantially more power vs 1 and 2

Shot spread:

Configuration 6
A combination of 4 and 5
4X Colsons with 1/2in nominal compression, 2x Stealth wheels with 1/2in nominal compression, 2x stealth wheels with 1/2in nominal gap
Tremendous power, easy to feed
Two shots maybe should be tossed as outliers, but maybe not… kept in analysis
Image below is not an illusion, inner axles are tilted

Shot spread:

High speed video
1-3x examples of each configuration with the same good-condition offical/legal Note
1kfps

FRC 95 The Grasshoppers 2024 Shooter Prototype HSV 1

Analysis
I used a Standard Deviations test to compare all six datasets.

Overall it shows at configurations 4, 5, and 6 compare similarly and are all superior to configurations 1, 2, and 3. I suspect that configuration 6 could be tuned a tiny bit more to be an absolutely devasting sniper shot, and 4 and 5 could be improved substantially as well. Additional flywheel inertia and careful application of slippery materials may be helpful for all of these designs.

We are going to continue to plan on using a design derived from Configuration 4, but with independently driven 3in stealth wheels instead of tilted axles. The reduced number of axles improves packaging and total weight without an detectable sacrifice to accuracy or range.

[AmbroseWiering]

Wow this is a really fascinating write up and analysis of shooter performance. Thanks for sharing.

[JamesCH95]

Related to the above testing, and all of our prototypes, I always had an intuition that ‘full width rollers’ were a bad idea. My mental model said that full-width drivers would distort or collapse the note* but I couldn’t explain why in a technical sense. And like my momma always said: “if you can’t prove it and show your work why should anyone believe you?”**

So this morning, I did.

This free body diagram makes the assumption that the note is a loop of rope or chain, i.e. only a tensile member and incapable of supporting any bending load. Obviously this is not completely true, but it’s probably closer to the truth than assuming the note is rigid like a frisbee.

This FBD indicates that as the infinitely floppy note is driven closer to the outside the more force from the shooter is converted into force accelerating it down-range. We can make some logical leaps from these results:

• Single-sided shooters (like configs 1, 2, and 5 from earlier) should have a contacting solid backer on the non-drive side so the note doesn’t pop out of the drive wheels (see configuration 1 flub in HSV)
• Full-width rollers are not super efficient and commit significant energy to distorting the note (see configuration 7 in the HSV)
• Driving from the narrowest wheels, situated as wide apart as possible, will likely result in the best shooter performance overall (assuming contact time and traction are sufficient)
• Configurations like 4 and 6 will add power AND support the note from lateral deflection AND add spin. This is evident in configuration 6’s increased power.

*I think @Bjorn @Justin_Foss and co. have a lovely freeze-frame of a note going full-hotdog, but I can’t for the life of me find it.
**Nearly a verbatim quote from my mom when explaining why I needed to show my work in Algebra.

Addition: here is a good image from Team Scream’s build thread that shows how the notes get laterally distorted during shooting.

Team SCREAM Open Alliance | Team 4522 and 4766 | Build Thread 2024

Your shooter provides a great view of the note shape as it is grabbed and spun: [image] It looks like you have a slight bit of compression. Did you find that improved consistency? Can you confirm compression, motor type, and wheel type and wheel speed? And is this in open loop or closed loop control?

https://www.chiefdelphi.com/uploads/default/original/3X/f/2/f205c1c4ff478239adc58f824c42d0d51d02b311.jpeg

[JamesCH95]

I missed these image from the big testing post. The notes in configuration 2 and 5 (this image from 2) were spinning so fast that they would peel over (and sometimes remove) gaff tape marks from earlier in the test run!

We wrapped a few sessions of 21 shots only to count 19 or 18 marks, and to find a couple coiled up bits of tape on the ground.

We also may or may not have peeled the recycling sign off the wall at one point…

[davepowers]

@JamesCH95

video

This video is stellar, super cool to see!

[Wesley]

@JamesCH95

We also may or may not have peeled the recycling sign off the wall at one point…

It’s fine, it was on the floor to start with.

[Travis_Hoffman]

@JamesCH95

We are going to continue to plan on using a design derived from Configuration 4, but with independently driven 3in stealth wheels instead of tilted axles. The reduced number of axles improves packaging and total weight without an detectable sacrifice to accuracy or range.

What kind of maximum range are you targeting when utilizing the 3" wheels? Also, are you planning for some upduction in the wheel speed? It looks like you’ve got some different-sized pulleys in your Onshape model. Thanks in advance, and thank you for sharing so much useful information.

[Wesley]

It’s time for our annual ChiefDelphi new battery name vote! This year’s theme is cheese.

• CHEEZBURGER!
• Pepper Jack
• Gouda
• Brie
• Mozz
• Cabot
• Cheddy
• Extra Sharp
• Muenster
• Bleu
• Feta
• Ricotta

We name all our batteries to keep track of them— names tend to be easier to remember (and more fun) than numbers.

[JamesCH95]

Let’s talk about tube nuts!

With a 5C collet closer and 19mm 5C collet I was able to press a tube nut into a Rev Ion dead axle tube with zero fuss. I threaded a 3/8in 10-32 screw into the tube nut to push from and to recess the tube nut slightly. It only took one hand, I used the other hand to record the video.

This will be our process for installing tube nuts. We will be able to install tube nuts safely in the pits at competition, which is a good situation to be in. It also works on nearly any length tube because we can hang the press off the edge of a table a little bit and clamp the back to our toolchest worktop.

We bought the 5C collet from McMaster, and I got the closer some years ago on amazon. You won’t need to buy particularly nice collets or closers to do this on your team, concentricity and parallelism are not important to this process like they are in machining.

[JamesCH95]

Robot update*

We’ve slapped together an undertaker to test. It uses printed augers on the outside and F4 tape in the center for a little more grip. Drive is 12:54 on 1x NEO 550 per roller via 2mm GT belt. We applied some PTFE film to the frame rail that the notes slide against, which helped immensely.

Details are all in our OnShape models in the OP of this thread.

Taste the rainbow!

We mounted the ‘shoulder rails’ of our chassis using Rev Ion tube caps. We opened up the outer frame rail holes enough to let a SHCS to pass through a bolt, lettings us clamp just one web of the tube. This gives us a strong and clean right-angle joint. It will also let us remove the shooter mechanically by pulling 12 bolts.

This is also my new favorite deburring tool. It works a treat for deburring smaller holes.

*forgive me for multiple posts. I wanted the tube nut post to stand on its own.

[chicenbokner01]

those augers are sick!