Never Do This, and Other Good Ideas Gone Bad

[Ian Curtis]

Originally Posted by IKE

So the scissor lift reminded me that there are a lot of designs experienced teams have tried once, and forever sworn off. This thread is meant to be a compilation of "Ideas gone bad" especially the ones you see year after year. Add your Gone Bad designs, and please explain details:

I will let someone else cover scissor lift. My favorite is the big hopper of balls with a single slide open door. These always work in your head, and they always fail in practice. Typically FIRST picks a compliant ball that wedge together and jam at the exit. This leads to frantic ramming in order to unjam the hopper.

I think anytime you have balls in an open hopper you are asking for trouble. Look at 254/968 in 2006. While they had a big hopper, they several belts at the bottom to get them moving towards the shooter.

We had fair success keeping them in line (i.e. a tube with a width equal to the diameter of the ball so they couldn't bunch up), but even this clogged on us a few times in 2006. So in 2009, we always kept a grip on them.

Moral of the story: Keep something you can power on gamepieces at all times. Leaving them "to do their own thing" will only lead to heartbreak.

[Tom Line]

Ah, the number of spectacular failures we've had.

In 2006, the Pi's rookie year, the team had a belting company sponsor us. The team decided on treads, and the company supplied us the softest durometer natural rubber they had. At 30 lbs the robot was pushing people across the carpet. At 120, when it turned, the treads went straight and the robot stepped right out of them. They pulled the treads mid-competition, refit wheels, and never went back. However, they had a single match where they hadn't finished yet, so they put the bot out on the field without any wheels. No one said they had to move, right?

In '07, we had a spring counterbalancing the arm. The spring was strong enough, but the 1/4 bolt holding it wasn't really. I wasn't there for the catastrophic failure of approximately 400 lbs. of spring tension. Thank goodness. Gas shocks are your friend.

'07 banebots. 'Nuff said. Never went back - but we're giving them another shot this year. We have Andy Mark planetaries already here as a backup, because when you buy from Andy Mark you just know it works.

In '08, we went for a 6 wheel omni wheel setup. 4 oriented front to back, and two sideways. Unfortunately, we didn't have the foresight to make any spring loaded. Ever tried to make every leg on a 6 legged stool touch the ground at once? Our drivetrain was quickly rebuilt and frankly was never where it should have been that year.

Also in '08, we built a forklift. We ordered Bishop Wisecarver ball bearing linear rails. They strung us along for 4 weeks, telling us the parts were coming, before finally telling us we'd never see them. We built the entire 3 stage lift system out of igus slides, and used them in a manner I'm pretty sure that Igus never intended. To this date, we don't talk about Bishop Wisecarver without nasty glares.

In '10, we tried a linear kicker. Linear bearings from Mcmaster car on precision ground steel rods. Total stroke of 6 inches. 3 inch wideup with a 3 inch slowdown after hitting the ball. It worked great on paper. In reality, we shot 1/4-20 bolts across the room into a couple 4x8 plate glass windows.

[1086VEX]

from 2006, 1086 has sworn off making another turret. just not our strong suit. and since 2007 we said no more elevators after making ours that year. but then we decided buying a premade system would still be possible.

[dmitch]

The main problem with the scissor lift is that this year you have to lift 9 and a half feet to get to the top middle peg. To make this scissor lift, you would most likely have to have 2 connected at the joints to eliminate sway. Then when the robot raises this 9 feet, you have an amazingly top-heavy and unwieldy robot.

Also, as our coach likes to point out, more joints=more places for failure.

Archimedes screw is fantastic if implented correctly our team used 1 in 2009 to raise the balls from lunacy from the base of our robot to the shooter. We used a 2.5" piece of PVC that had a garden hose nailed to it in a spiral to provide traction.

[Andrew Schreiber]

On 397 the running joke for NEVER DO THIS was giving the Schreiber brothers access to gas shocks. I would like it noted that no one was ever seriously hurt despite dealing with 1000lbs of force...

NEVER make absolutes in an engineering discussion.

(ignoring the first one)

Never rely on gut feeling, check the numbers or don't build it.

Never trust an engineer to do the numbers right the first time, have a couple people look at the numbers.

"If you need slip rings in FRC you are probably over complicating it" - said by 397's old lead mentor. Totally true for our team.

[theprgramerdude]

Never build a permanently top-heavy robot. In '08 we ended up having to drive backwards because the lifter/ball-hitter was too far back, and we'd flip if we drove forward.

[cire]

2003... Never use #25 chain for your drivetrain. It may seem like you will never hit that several hundred lb limit, but when you throw in missalignment and shock loading, well lets just say we went through a lot of chain that year.

2004 - Never use 4wd in the 4 corners with high friction wheels... as somone else mentioned it is very bouncy when you try to turn - you DO drive straight very well though! We ended up wraping zip ties arround one set of tires which sort of helped..

2005 - Ropes don't belong on robots. We had at least 4 ropes on our robot holding things in place, although we managed to get it down to just the winch by competition, which worked pretty good.

2006 - Trust your numbers calculations (after you have checked them 5 times).. We made a ball shooter, calculated the theoretical speeds using 6" wheels, which fit well within our limits. Somone decided it would be better with 8" wheels - Our shooter ended up shooting too fast and we couldn't use it at the competition.. Another lesson is to make your design adjustable - There was literaly no way to take apart the shooter without taking apart half the robot.

[JackS]

Here is a list of 2791's don'ts with some explanations beneath it.

1. No Set screws
2. No Jags
3. Don't put non graded bolts in drive train.
4. Always account for force of pre-charged pistons
5. Don't put taps where they aren't 200% unavoidable

1. Set screws. We learned the hard way at FLR in '09, tightening the set screws before and after each match. Needless to say, by the time we hit Friday in Atlanta, every last set screw had been replaced with a cotter pin.

2. Jaguars. In '09 we blew at least 4 Jaguars, maybe even more if you count off-season work. Until we are convinced Jags are as tough as Vics, we won't put another one on our bot.

3. Never put any non graded bolts in a drive train. Every last bolt head sheared in '10 and took 3 hours to replace Saturday morning.

4. Never put a piston on your robot until you are sure it can handle the load. Last year every time our piston pre-charged, it bent our entire chassis. 15 lbs. of 8020 later, it still bent our entire chassis.

5. Taps. '10 was our first and maybe only experience with 8020, and we overkilled on the taps. It took forever to fix anything on our chassis, and by the end of the season, the screws had been put in and taken back out so many times the taps were loose.

[boomergeek]

Originally Posted by JackS

Here is a list of 2791's don'ts with some explanations beneath it.

...
2. No Jags
...


2. Jaguars. In '09 we blew at least 4 Jaguars, maybe even more if you count off-season work. Until we are convinced Jags are as tough as Vics, we won't put another one on our bot.

No Jags... without using the CAN interface to limit the rate of voltage change allowed.

http://www.chiefdelphi.com/forums/sh...k#post9 47848


Team 241 had 3 Jaguars short out in 2010 and one in 2009:
All resulted in the circuit breaker on the associated Jaguar breaker trip continuously that made the rest of the robot uncontrollable- (caused variable lag time of one to three seconds on all other functions on the robot).
We guessed that the failure mode was caused by voltage spikes caused by attempts at wide voltage swings too quickly.

One failure occurred at BAE unveiling event ; (less than 5 minutes total use prior to failure)
One at the Granite State Regional ; (less than 10 total minutes use)
One at the Connecticut Regional ; (less than 40 minutes total use)

The first two were on our drivetrain (CIMs).

The last on was on our ball magnet roller (Fisher-Price).

At Manchester, we added a CAN control to limit the rate of voltage change so that the Jaguar would not be swung from +12V to -12V too quickly (and vice versa).
Since we make that change, no more the Jaguars blew on the drive train- Because we had 2 blow out of 40 minutes total use (10 minutes times 4 CIMs) and we have not had any more blow after 240 minutes total use- we think the workaround helped.

But due to an oversight, our CAN workaround to baby the Jaguars did not get put on the Jaguar driving the fisher-price until after we blew another Jaguar.
After we put the CAN workaround on the roller, we have not had another Jaguar blow. (16 minutes use without a failure).

---update---

In over 8 hours of off-season use, not a single Jaguar has blown since we put in the CAN workaround.

[IndySam]

In 2005 we used casters on the front of our robot. When we would go to the loading stations if they got on the raised part of the loading station we would get stuck.

I have seen many instances of caster problems on robot so "no casters is a fixed rule for us.

#1 No casters
#2 No scissor lifts.
#3 No elevators (I have seen many instances of very effective elevators but not from us)
#4 If balls are the game piece they must always be under the influence of at least one moving part.

[Brandon Holley]

Lots of good ones out here already, but my personal favorite:

NO SET SCREWS

I will never ever trust a set screw alone to retain something to a shaft on a FIRST robot.

-Brando

[artdutra04]

Originally Posted by Brandon Holley

Lots of good ones out here already, but my personal favorite:

NO SET SCREWS

I will never ever trust a set screw alone to retain something to a shaft on a FIRST robot.

-Brando

Retain, or transmit torque?

Relying on set screws to transmit torque on an FRC robot is almost always a bad idea (except for things like potentiometers), but we've used them dozens of times to hold something in place (laterally along shaft) without any problems. That being said, I prefer to design around them, and either use Delrin spacers, stepped shafts, tapped shaft ends, E-clips, or roll-pins to avoid them. If we need to use them, we almost always drill for two set screw holes 90 degrees apart, use Loctite threadlocker, and only use the cone-tip set screws.

[IKE]

Originally Posted by Brandon Holley

Lots of good ones out here already, but my personal favorite:

NO SET SCREWS

I will never ever trust a set screw alone to retain something to a shaft on a FIRST robot.

-Brando

What about trusting Set Screws on FRC Mini-Bots? The Mini-bot FTC kit was full of set screw based components.

[PAR_WIG1350]

Originally Posted by cire

2003... Never use #25 chain for your drivetrain. It may seem like you will never hit that several hundred lb limit, but when you throw in missalignment and shock loading, well lets just say we went through a lot of chain that year.

last year, in Atlanta, 1350 threw a chain in every match, it was almost as funny as it was disappointing (think of a car in a cartoon driving as parts fall out all over the road, but it just keeps driving [we could drive very straight with only 3 powered wheels on the ground]) . Every time it was a #35 chain, The # 25 chain that drove our slick wheels worked perfectly, (unless the #35 chain driving the #25 chain fell off, which didn't happen as often as the chain to the plaction wheels broke).

So, in summery, #25 chain can work well in a drive train.
----------------------------------------------------------------------------------------------------------------------

Never use too many* long* chain runs in critical** systems.







*these are not terms that I can define absolutely, use your best judgement.

**if you need it to play (well) and it isn't redundant to the point that if one breaks another/the rest can do the job almost as well without it [If its failure means you cannot continue to play the game] than it is a critical system.

[Dillon Carey]

2007: always use active accumulation devices (rollers)

Our first arm in 2007 used a claw to grab the tubes. It didn't grip very tight and required much more attention from the drivers.

2008: make sure you can replace any part of the robot in the length of a timeout

We lost Wisconsin that year due to a motor burning out and not having a spare assembly ready to replace it.

2010: always use capture rings on CIM shafts. This way, even if the set screw holding your key in comes out, the key can't come out

At Kansas we lost in the eliminations due to overheated CIM's. They overheated because a key came out of one, causing the other CIM on the same side to overheat.

2010: do not put a bearing on a hard anodized shaft.

At Midwest we lost due to our chain falling off. Don't cantilever a swerve module off of a hard anodized aluminum/needle bearing combo.

(This doesn't mean we will never try these types of systems again. These are just examples of poor implementation.)