Team Update #18
 

http://usfirst.org/uploadedFiles/Rob...pdate%2018.pdf

It's a real game-changer...

Re: Team Update #18
 

Can you imagine how this will effect the Finals on Einstein? You know there are going to be minibots that are almost equal speed...

Re: Team Update #18
 

no!!! thanks FIRST that you make it clear i like updates and seeing this from what you learned in previous week but cant this wait till championship

Re: Team Update #18
 

Not a fan. Witnessed MANY minibot climbs this weekend at San Diego that most certainly imparted 3-4 N of force yet were not auto triggered. It's going to suck when the field fails to work correctly and teams lose points they should have earned.

Re: Team Update #18
 

I saw far too many missed triggers (at least they appeared to be) on webcast this past weekend to be comfortable with this update going into our event this week.

Re: Team Update #18
 

I agree, Cory. More than once a 1 second minibot with significant momentum failed to trigger. Is it possible that the triggers still need too much contact time to register? BTW, we had a great time partnering with you guys and 3704!

Re: Team Update #18
 

When that happened at the WI regional, the refs manually triggered it.

Re: Team Update #18
 

Not a fan as well, we ran our minibot at around 2 seconds at san diego and it impacted with well over 2 newtons of force. It would certainly hurt to get hit by it.

We've since increased the speed a lot, more along the lines of 254, and the impact is pretty violent.

Re: Team Update #18
 

The triggers are designed with multiple sensors, and the FMS only looks for 1 of those to be triggered for any amount of time. Thus making the minimal amount of force still trip the tower. The towers have been tested and proved to be much more accurate since week 1, so I think this should be a non-issue now at week 3.

Re: Team Update #18
 

Someone in another thread on Chief commented that the towers are now looking for a triggered sensor for a certain amount of time to be sure it wasn't just a jostle from a robot.

Whatever the circuitry I am 100% confident that there were multiple (greater than 10) instances at San Diego where minibots compressed the platform upwards into the sensors without the tower actually triggering.

Re: Team Update #18
 

Seconded!

Re: Team Update #18
 

Do you have proof that enough force was exerted? or is this just your 'gut feeling'?

Re: Team Update #18
 

Do you know what thread that was? I'm curious about how that would work without limit switches.

The sensitivity of the towers now is highly dependent on how the field crew that implemented the fix did it. Really, what the minibots are doing now are pushing the bottom plate up about a quarter of an inch. However, that distance may vary depending on how meticulously the field crew assembled the plates. I was told to get up on the ladder on Saturday morning and adjust the distance a little bit more and make sure that the distances were equal, but it was only by an eighth of an inch by so. FIRST's solution works, as long as the instructions are followed to the T.

Re: Team Update #18
 

I could do the math for a 2.5 lb object moving 10 ft/s hitting that plate and you'd see it's more than 3-4 N, but if the plate is being violently slammed into the top plate such that you can hear it from the stands, I'd say it's a pretty safe bet they were triggered. Additionally we have 4 official towers and I can confirm that every single time ours goes up the pole it depresses the limit switches, but twice this weekend it did not show the tower as triggered at San Diego.

It seems pretty clear there are still problems with the field and I don't just say this because the tower failed to trigger with our minibot. It was a very common occurrence.

Re: Team Update #18
 

Assuming a .5 kg minibot can climb 10 feet (about 3 m) in 2 seconds, and that the plate depresses at most 2 cm (0.02 m), this means that its average (not top) speed is about 1.5 m/s and:

Using the work-energy theorem,

W = KE
F*d = .5(m)v^2
F = .5 (m)(v^2)/d
F = [.5 (.5 kg) (1.5 m/s)^2]/.02 m
F = 28 N

...is (actually less than) the average force needed to stop said minibot (because 1.5 m/s is the average speed, and not the top speed). If the compression distance is less, the average force will be higher. If the minibot is faster than 2 s, then the average force will be higher. If the minibot is more massive than 0.5 kg, then the average force will be higher...

...which means that, if the towers are doing their jobs properly, no one with what I would consider a "competitive" minibot has anything to worry about -- and if they DO NOT trigger the towers, then we have direct evidence that the specifications given (2-4 N of force to trigger) are simply wrong.

I'm quite tired, and didn't put much thought into this. Anyone want to poke holes in my analysis?

Re: Team Update #18
 

The only issue in Wisconsin that I am aware of occurred when a minibot just touched the plate but did not move it. The Head Ref did a valiant effort to make the determination and double and triple check the rules. At first it was called a climb but was eventually called as a non contact when they removed the minibot and the plate did not move.

Re: Team Update #18
 

And the naked king had beautiful clothing..."Week 2 results indicated that this change was successful."

Re: Team Update #18
 

The reason I ask is because one must takes into account that some (in some cases a lot) of that force is deflected and not transferred towards moving the plate. Many of the minibots strike the plate such that they change orientation rather than push the plate.

Hence my question if you knew that the force was being applied (and applied properly) or if it was just a 'gut feeling'. Sound of the impact does not equate into actual applied force.

Re: Team Update #18
 

Although I wasn't there, I am certain he has more than a "gut feeling" of assurance that more than 3-4 newtons were applied... HS level physics state that:

mass*velocity(change) = Force*time

Our minibot weighs 2.3 pounds --> ~1 kg
Our minibot travels up the pole in about 1.3 seconds --> ~2 m/s

By just saying that the minibot stops (although ours certainly is repelled backwards, too!) you get a change in velocity of ~2 m/s.

If the minibot is being slowed by the tower for less than half a second (I can't imagine one taking that long... most stop suddenly!), then the Force is greater than 4 Newtons.

(1kg*2m/s) / .5s = 4 N


I certainly hope that refs are allowed to add judgment to obvious false negatives!

Re: Team Update #18
 

If your minibot hits the plate and triggers: Score

If your minibot hits and doesn't trigger: no score even if visually witnessed by refs.

If the tower is bumped and triggered by a robot: no score/score corrected

I was a ref in Pittsburgh when this rules was changed/reinforced.

Re: Team Update #18
 

One of the biggest issues I see with this whole thing is that most teams don't have ready access to a real tower. They'll just have to hope that their minibot triggers it. They can't really be proactive about a solution despite the warning in the update. I predict that despite this update, there will be almost just as many tower failures next week as in the past two.*

*Provided the towers are not fixed. Fingers crossed!

Re: Team Update #18
 

What is your basis for this statement?

Re: Team Update #18
 

Team Update 18 and listed in the manual from a team updated around 15 or 16 because of how at week zero events robots acidentally triggered the plates.

Re: Team Update #18
 

Instance 1: common knowledge
Instance 3: been proven by previous events
Instance 2: proven by Update 18

Re: Team Update #18
 

Note the second paragraph of the Update. If you hit the trigger plate directly on a bolt, it will not register. Fix your Minibot so it hits inside the bolt-pattern circle.

It should work - it did work in the vast majority of instances in Week 2. If it doesn't work, the ref and FTA can determine if the trigger is working properly.

Re: Team Update #18
 

It says, "we will be relying on the automated scoring of the minibots". Does this mean that there will be no referee calls at all?

Re: Team Update #18
 

Or at the very least, less than previous weeks. Technically, the manual does say that if the plate does not trigger it, it goes not count.

Re: Team Update #18
 

Being at both a Week 1 event (Kettering) and a Week 2 event (Waterford) I can agree the changes had a night and day difference. At Kettering, hardly any of the towers registered. The slower minibots that creeped up the pole would generally register, but teams such as 27, 33, 67, and 2337's minibots would register rarely, if at all. But in Week 2, comparatively faster and slower minibots were triggering the towers with no issue at all (the only tower-related issue I saw was lights not coming on at all, but it still registered with the FMS).

The update definitely keeps a layer of uncertainty with "What if there's an error on Einstein?", but the changes seem to be working so far. That worry can be revived if we continue to see errors into weeks 3 or 4, but for now, awesome solution. From what I've seen, it works like a charm.

Re: Team Update #18
 

The rules stated minibots need to exert 2-4 N of force to register. Teams should have designed their minibots to exert at a minimum 4 N force.

All this update is saying is if teams have not designed their minibot to this spec, they will not trigger the tower in week 3, and if they have, the minibot will trigger the tower.

Perhaps FIRST has figured out a way to make the towers accurately sensitive to 2-4 N force. If after week 3, your minibot registered half the time and didn't register the other half, AND you have physically measured the force with which it hits the top plate to be >4N, then there is a case to be made.

As for right now, they are just rectifying a lapse in the rules that has been bypassed (in good faith and fairness) by the referees in weeks 1 and 2 for what they thought to be a field error.

All we can do is play by the rules.

Last Note: I don't think the vertical force problem is as simple as some of you are making it out to be. Remember 2 years ago in lunacy, there were hundreds of posts about how "it won't matter how many wheels are touching the floor because frictional force is not area dependent." In fact, it certainly was area dependent because of surface roughness- a fact that many teams overlooked. I remember team 2753 (a rookie) realized they had better traction with more wheels despite the simple physics, and made it all the way to Einstein because they had better acceleration and speed than anyone else.

In the minibot case, there are motors which stop turning when the light switch is triggered, there is rolling friction, some other parts dragging on the pole, the light switches absorb some of the force in being depressed, etc. All of these things could add up to a difference maker. And no, I have not attempted to calculate it either, but that's why if you think something is fishy about the system, you need to physically measure the force out on the practice field.

A rare 2 cents from my short arms and deep pockets...

Re: Team Update #18
 

I can see this rule being needed as we approach the Championship and many teams will have minibots, but I'd love to see the computer's results only being used when it's too close to call.

Assuming they can get the sensors to work properly this is a great change. However, from experience with Breakaway's automated scoring system, there will be many bugs which will cause teams to lose matches.

The field techs and refs have there work cut out...

Re: Team Update #18
 

Our minibot is not particularly fast but is getting better. During Week 1 at Alamo, it tripped the sensor every time. But at the next event it will be lighter and faster (a design which we built before we shipped but had no time to test due to weather-related delays). After some complaints on Thursday, the few teams with really light and fast minibots got the referees to manually judge the minibot race. It seemed fair to us, those light minibots are great machines.

I want to point out a common engineering practice. If my company had to build a minibot that triggered the sensor EVERY time in order to get PAID, we would build something with a 2:1 margin for error - something that would under nominal conditions exert 8N of force.

And I think that minibots reversing direction too quickly may play a bigger role that we think. We may add a spring in the lever that depresses the switch in our newer design.

Re: Team Update #18
 

That's a great point. Our minibot just turns off when it hits the top. (And usually comes back down because of the impact with the plate -- meaning that the plate is providing the impulse to reverse the momentum). If some teams are actually reversing their wheel direction before the impact is finished, this could have a dramatic effect on the overall force (and the maximum force) with which they are hitting the plate.

Re: Team Update #18
 

"Week 2 results indicated that this change was successful."

Hmmm. Like many others who have commented, our 1.2 sec. minibot smacked the target with a substantial impact in Week 2 (my calculations say in excess of 100 N), and yet did not result in a trigger light a couple of times. Clearly with all the antecdotal evidence there is still a phenomenon occuring which can miss an impact of adequate force. Any mechanical action has both a time and force component, and perhaps the rules should have specified both. Like any control system the field system has a controller with a finite scan rate, and any control system is capable of missing a signal if it is short enough. I am a little nervous about a statement that essentially says "we have determined the system is now perfect" if that means no provision will ever be made for correcting a result when it is clearly wrong. I'm grateful we are not playing again until Week 6 to give some time for this issue to get sorted out. Let's hope no Week 3 teams experience any problems which effect their results.

Re: Team Update #18
 

2 observations:

1.) Design the mini-bot so that it can't hit the bolt first. Honestly, I think the bolt pattern was the cause of most of the times it didn't trigger at Pittsburgh. We personally checked every single contact point on the towers before each day and they all triggered with minimal force.

2.) Make sure there is a delay between when you hit the top and when you start reversing the wheels.

I think this is a necessary rule for the Championships. Mini bot races are very soon going to get too close to call and we need to perfect the tower designs and also for the teams to perfect their mini-bot designs so they do not hit the bolts and provide the necessary amount of force. I think thats the only path to reliable calls at the championships.

Re: Team Update #18
 

Forgive my tired brain, but is the bolt pattern in the field drawings?

Re: Team Update #18
 

Where in the rules does it say that you have to provide a given impulse? All it says is 4 newtons, not 4 newtons for 2 seconds, half a second, etc.

Teams made their minibots to FIRST's rules. They should fix their (still) not fully functional system, or put measures in place to ensure that minibots conforming to their published rules actually get scored.

Instead they just say "our system is perfect, so if you don't trigger it, even if you blast the darn plate off the top of the pole entirely, you still don't get the points". If the system is proven to have any flaws then FIRST ought to recognize that it is possible for mistakes to be made.

Anyone seriously arguing whether a 2.5 lb minibot going 6-10 ft/s will impart 4 N of force or not is just being a FIRST apologist. This is simple physics and some efficiency losses due to friction, etc aren't going to change much.

Perhaps we should take some video of our official competition towers being repeatedly triggered by our minibot to silence the doubters?

Re: Team Update #18
 

I agree! With all the minibot rules the time to change how the race is drawn and towers triggered is not week 3! I would expect FIRST to do the opposite and put refs in place of the electronic or at least watch.

Grrrr.

Re: Team Update #18
 

One obvious omission is that you did not account for the mass of the plate. The plate must be violently accelerated in the short amount of contact time it has with the minibot in order to move it enough to trip the sensor(s).

How much does a 12" diameter 0.25" thick polycarbonate plate weigh? My quick calculation says it's more than 4 Newtons (please check my math). So obviously 4 Newtons is not enough force to move the plate at all, let alone violently accelerate it. So where did the 2-4 Newton number in the manual come from?

The specification of what exactly is required of the minibot in order to trip the sensors is inadequate to allow accurate analysis. What exactly is being sensed? Proximity? Rate of change? Dwell time? Some combination? Does anyone know?

Re: Team Update #18
 

I am not apologizing for FIRST, I am just advocating pragmatism from the teams. The bolt pattern was definitely not obvious and I don't remember seeing it anywhere in the official rules. They are in the official field drawings in sections 37, 38 and 69 of the Game Field Elements but I know most teams (including my own) don't check those.

Believe me I sympathize with any team who's bot successfully climbs the pole but doesn't get the points for it. I know that it must be immensely frustrating. I want to eliminate that from ever happening. Thats why I think that with a few small minibot modifications, this will become simply a non-issue and the tower results will match everyones expectations.

I know there were cases where the minibots reached the top of the tower but it didn't trigger. I am just being honest and did not see any of those cases where it was completely obvious that the bolt was not hit and that it delivered enough force and still did not trigger.

Re: Team Update #18
 

Cory, I agree with what you're saying. However, I've heard that, due the false triggering seen early on in week 1, the field hardware now looks to see that the towers remain triggered for 75 milliseconds before a hit is registered. I can imagine that really fast minibots hit so hard that they bounce away and aren't making contact for 75ms.

What's annoying is that this means that the rules have changed, as you suggested. The de-facto rule now appears to be 4 newtons for 75 milliseconds.

Re: Team Update #18
 

Please see my earlier post about "4 Newtons".

Re: Team Update #18
 

my math agrees with yours. And I know the off switch on our minibot takes much less force than that to actuate.

Interesting.....the rules have been that the tower sensor must be triggered for it to count. Just because teams got away without triggering the first week or two, doesn't mean the rule doesn't apply.

I wonder how many slow heavy minibots will win races now?

Re: Team Update #18
 

Assuming the minibot stays on the pole after impact, the trick is to make the cut-off switches either slow (not a quick flip) or require more than 4N of force to push. Either will increase contact time with the top plate. Squishy surgical tubing comes to mind.

Design is iterative, right?

Re: Team Update #18
 

I would appreciate a more explicit mathematical write up as to why you are so convinced.

Most importantly, I want you to account for the drift in the motors due to momentum after the limit switch shuts down the motors at the top of the pole. Do you know, for a fact, the robot has enough speed/momentum to continue to carry itself upwards despite non-powered motors and gravity contributing to rapid deceleration? As pointed out in this thread, the non-powered motors might even stop the robot on a dime...

(For the record, I think there should be enough force too, but I think is not good enough in engineering, and FIRST hasn't given me good enough reason not to trust them)

Re: Team Update #18
 

I don't know about the towers at other regionals, but at Florida, for our 3 lb/1.6 second minibot, the tower triggered every time. Granted we didn't get our minibot and deployment working until our last qualifier, so it only had 4 goes at it, but it triggered it four out of four times. And for the most part, the towers there were registering minibots, both slow and fast.

However, being on the drive team leaves me with little time to watch matches, so I am not sure if every other team had this success rate, feel free to correct me if you minibot experienced a false negative at Florida.

Re: Team Update #18
 

I have tried several times this evening to access the tower diagrams and get timed out waiting for the download...does anyone have a stored image of the top plates and the bolt pattern they can share?

Re: Team Update #18
 

This is what it's all about:

You gotta TRIGGER the TARGET. If your minibot doesn't TRIGGER the TARGET reliably, you might need to redesign your minibot.

The description of the ARENA suggests that it will take a minimum force of 2-4 Newtons to TRIGGER the TARGET.

Re: Team Update #18
 

I don't want to attach/upload/etc the full field elements document, but I took a screenshot of the drawing of the bottom plate, which is where it seems that the bolts are an issue. Page 69 of the 2011 Game Field Elements_RevA, if you manage to download it and view it yourself.


Edit: You can see the bolts in some of Dan Ernst's pictures: http://www.flickr.com/photos/daniele...7626138552781/, http://www.flickr.com/photos/daniele...7626138552781/, http://www.flickr.com/photos/daniele...7626138552781/

Re: Team Update #18
 

Simple solution: remove off switch, replace components as needed. :)

This is an interesting point that never crossed my mind. Quick math suggests about 4-5 newtons of force just to cancel out weight of polycarb plate. Someone should bring a scale to one of this weeks competitions and see how much force is actually needed to trigger the plate (ignoring impulse). Might lead to some exciting discussions with the refs!


This week is going to be fun.

Re: Team Update #18
 

The rules don't say you need to hit the plate with 4 N force to win the race. They say you need to trigger the tower.

Isn't that the way it's been all along?

Re: Team Update #18
 

Good luck. Ever try asking to measure the overpass height in 08 to confirm it was within the tolerance listed?

And to trigger the tower you need to hit it with 4N of force. Your argument seems sort of circular? If you hit it with 4N of force and it doesn't trigger, then FIRST failed to implement their triggering solution correctly.

Re: Team Update #18
 

Exactly: by definition, if the sensors aren't tripped, you haven't triggered it, and are entitled to zero points. It doesn't matter how hard you hit it.

Fundamentally, the problem is that the rules define scoring in terms of a process that is hard to directly observe (were the switches actually tripped, or did it hang up on the bolts?), and which is inherently impractical to error-proof (did the sensors get tripped because a robot shook the tower, or because a minibot ascended it properly?).

When the refs were scoring it manually, there was really no way for them to systematically and conclusively distinguish false positives, false negatives, true positives or true negatives. They were just guessing. (And the timing aspect being based on triggering, and not mere contact made it all the more impractical to observe from floor level.)

I think this update makes the best of a game design choice that was, in retrospect, not so good. An alternative might have been a rule change, to change the criteria for triggering, but I can certainly see that that introduces other problems. I can only hope that FIRST did some testing and established that the best balance of true and false outcomes is achieved through the changes they've implemented.

Re: Team Update #18
 

My argument isn't circular. You're hung up on the 4N number. The game rules don't mention it. The ARENA description does, but it does not say that it's an upper limit. It suggests it's a lower limit.

You've been in FIRST long enough to know that the ARENA doesn't always behave the way the manual suggests that it might.

I guess we'll find out in a few days how our minibot does. I expect we'll probably have to play with the off switch, if we can get it to deploy legally. Should be interesting.

Re: Team Update #18
 

So, do you really need to hit it for 75 milliseconds!? That seems like an awfully long time. And what about issues such as switch bounce. Is that dealt with?

Re: Team Update #18
 

I think you're reading it wrong. I read it as "the minimum force is 4n" as in "any force 4n or above will trigger".

I think first saying "lalala our sensors are perfect" and ignoring teams is just stupid myself.

Re: Team Update #18
 

Well, switch bounce wouldn't matter, because you can't score twice on the same tower. Switch bounce just makes it look like there was a bunch of hits when there was actually only one.

Re: Team Update #18
 

The manual says that "[a] minimum contact force of approximately 2-4 Newtons, depending on contact location, is required to ensure the contact sensors in the TARGET trip reliably."

I don't read that as a definition; to me, that looks like an observation for the benefit of the teams. If it were a definition, it would be the worst definition in the world. (Approximately? A range of force? Depending on contact location? And how do you quantify "reliably"?)

Re: Team Update #18
 

Here's my issue: if they only specify a minimum, then how to you guarantee you trigger the sensor? If my minibot applies 1000 N and it doesn't trigger in a match, then that is perfectly acceptable field behavior by the definition in the rules as you interpret it (however ludicrous that is).

As an engineer, I expect specifications that are guarantees of proper performance - not a specification of "below X it's guaranteed to not work, and above X may or may not work but we're not really sure". With no specification of "above XX N the tower WILL trigger" how do we design minibots?

Can't the silly force triggers be replaced with light sensors or proximity switches? Those won't be fooled by robots bumping into the bases, won't have different properties depending on where you hit the plate, and won't bind on the pole like I've seen the plate do.

Re: Team Update #18
 

Easy way to know if your minibot will be ok? Have a FIRST certified tower during minibot inspection with a force gauge on it. Then you can have a document from an inspector stating that your minibot meets the minimum required force needed to trigger the target. So if it doesnt happen on the field you have some sort of justification for the refs.

This would eliminate everyone on here saying "show me the calculations to prove it!" or "you didnt account for this drag force or weight of the plate!" etc...

Re: Team Update #18
 

I guess we were all hoping that the 4N spec was an attempt to provide a specification that teams could test or at least calculate and know where they stood.

Even if you build a competition tower exactly according to the Game Drawings you cannot test TRIGGERING at home because you don't have the FMS setup. That's not an engineering challenge, it's a lottery.

Re: Team Update #18
 

That's exactly what I was thinking, but unfortunately it's too late now. Teams have already designed minibots with the pressure plate as a constraint, and played that way for two weeks.

If I'd designed the thing, I probably would have used a series of three or more sensors at the desired level, and used some logic to have them "vote". When a majority say you've reached a discrete position, you'd score. (And it would fail gracefully if it didn't work: a human could still judge the position visually, rather than have to estimate whether a switch was or was not tripped.)

According to the logic embodied in the rules, you can't.

Re: Team Update #18
 

Guys,
Lets not get crazy. The towers work. There is no complicated issues here. If your minibot can climb the pipe in under five seconds it will trigger the top. If it struggles to get up, then the plate may be able to push back with enough force to stop the minibot and not trigger. I saw more than one minibot over the weekend climb the tower with the drive wheels spinning faster than the climb rate. If the minibot made it to the top, only one was not able to trigger the tower in ten seconds. That one was spinning the tire on the pipe at the top. It was simply a matter of equalized forces. The plate did not move.

Re: Team Update #18
 

Good call on the math, Ether.

A lesson I seem to have to re-learn every single season is that every measurement FIRST gives is nominal, whether it's said to be nominal or not.

Re: Team Update #18
 

Al-
What your describing is in complete contradiction with some other long time FIRST vets. I've taken both yours and their advice many times, so I'm at a sort of impasse. We haven't competed yet (this week at Bayou we will) but I'm moderately concerned our minibot may not trigger the tower every single time based on its similarity to minibots these other veterans have. While I don't doubt that most minibots will trigger the tower, how can we make sure the towers are triggered EVERY time?

-Brando

Re: Team Update #18
 

Brandon,
Have faith. Participants, FTAs and Refs are not going to allow teams to miss out should there be field issues. If there is reasonable doubt, I am sure the staff will take action and/or resort to manual scoring. There are contingency plans. We know how hard you work to get to the point of competition.

Re: Team Update #18
 

Thanks Al.

-Brando

Re: Team Update #18
 

Yes.

If an issue ends up complicating a regional, the head ref and FTA has the right to make the call and go to scoring the races manually. In fact, I'm sure many head refs will ask for the refereeing volunteers to take mental notes as to the finishes, so if a problem arises mid-match, that match's scores will still end up counting.

Re: Team Update #18
 

re: the force required to move the plate, vs. it's weight. If your minibot must lift the entire plate evenly, then it might take more than 4N to trigger the tower. If your robot contacts the plate on one side, then it only needs to lift that side of the plate. The plate should still rest on the two bolt heads on the other side of the pole. The force required to trigger the FMS should be significantly less than the full weight of the plate.

So Ether's math seems correct, but his lack of a free body diagram might lead one to an erroneous conclusion.

Interesting how so many people get upset about a game challenge that is not quite what they anticipated.

Re: Team Update #18
 

I guess a description of the new tower triggering mechanism would be useful for some here who did not get to see it at a week 2 regional. There are NO force sensors up there. There are simply 3 metal contact points that act as simple switches (note, this is as of week 2. week 1 had 3 limit switches). As long as 1 of those contact points is hit for enough time for FMS to read the signal it will trigger. The setup itself is dead simple.

One problem is that any changes made to the tower need to be so cheap, simple, and easy to assemble that FIRST can ship it to all the regionals and they can put put together by the field staff there. I think FIRST did a reasonable job with those constraints in mind. I would love for the bolts to not be there at all because I still think thats the cause of most robots that reach the top and don't trigger. Maybe the bolts could be reversed so that they are anchored to the bottom plate? (this would increase the effective mass of the bottom plate, so more force would be required to push it up)

Regardless, these are the current constraints of the system and I think with a few small mini bot modifications you will be able to trigger that setup every time. And if it still doesn't, then you have a better case for there being a major flaw in the triggering mechanism.

I believe that FIRST and the teams need to work on getting the automated triggering to work with very very high reliability. I believe the reason for that becomes clear if you examine the following case. For this case, lets make 2 assumptions. First, that teams and FIRST do not continue to refine the triggering. Second, that we allow the refs to manually override the triggering results.

With those assumptions, lets say we are in the final match of Archimedes deciding who will go to Einstein. All 4 minibots go up and all hit the trigger with imperceptible differences in arrival time. However, 1 minibot does not trigger. Now what do you do? A precedent was set in previous matches that the refs can make a judgement call based on what it looked like. Even if in that case the minibot was not as close of a call, but you know that will not stop the team from flooding the field with shaky video footage and their first hand accounts to try and show when it hit. I think these 2 assumptions lead to a messy future.

Now, if teams and FIRST work to refine the automated triggering mechanism then we will probably never even have this problem to deal with. And even if there are a very very small number of cases where it seems like it should have legitimately triggered but did not, then at least there is a consistent precedent of not counting it unless the robot triggers the sensors and the playing field will still remain even for all teams.

Re: Team Update #18
 

Yes...playing by the rules that have been there since the beginning.

Re: Team Update #18
 

Yeah, we get it.. That's the rule. I don't think anyone is arguing that you need to trip the sensor. The issue at hand is that there is some obfuscation in the method of doing so. Sure, there is plate you need to lift, and it takes some amount of force. But we as teams have no way to actually test what we believe will trip the sensor short of going to the regional and trying it. Sure we can guess all we want at what will and will not trigger the sensor, but in the end it turns out the process for tuning these things is not as black and white as you would hope for. Sure, things don't work perfectly, that's engineering. But there might be a more deterministic way to see who won the race. Beam breakers? We dump way too much time and money into these regionals for "just trust us" to cut it. Real time scoring has never been very reliable, we all know that. It's not an issue if there is an accurate backup. While I'm certain (now more than before with Al's post) that this issue can be resolved at the regional level, the wording of the update was a bit disconcerting. I also look forward Einstein Finals Match 2-2 when everyone's minibot is more or less the same speed.

TL;DR: We all just got told "you're holding it wrong".

Re: Team Update #18
 

Do you know if any teams built a TARGET using the field drawings, designed a minibot to make it work every time, and it's not working on the actual field?

If so, then I'd say we have a problem. If not, some teams just didn't do their homework.

Re: Team Update #18
 

You can build poles until the cows come home. It's the triggering settings that are the issue.

Re: Team Update #18
 

Cory of 254 said exactly that in post 15 in this thread.

Re: Team Update #18
 

As an engineer, I work to engineering specs. If I meet those specs, I expect to be paid. You can imagine that I might be a wee bit upset if I were to design a product that meets the customer's specs, deliver that product, and be told "I'm not paying you because the product didn't do what I expected it to do". If your specs are not written clearly enough, that is your fault, not mine.

In FIRST, the manuals and rules are the specs. Your argument that the 4N force was not in the rules is not valid. The peg heights are not in the rules either. The entirety of the game documentation is the specification document to which all teams work. Our minibot cannot hit the bolt heads, and I can guarantee that it imparts far more than 4N of force. Therefore, by any definition, it meets the specs. If we get to competition this weekend, and our minibot doesn't trigger the tower, am I supposed to just accept that and go on? Is that proffessional? In the real world, it would not be tolerated.

More importantly, this is supposed to be a race. The tower trigger is just a means of scoring that race. Imagine an Olympic sprinter that happens to be too short to reach the tape. If he crosses the finish line first, but runs under the tape, does that mean he didn't win the race? I think this whole argument is silly. The first minibot to reach the top should be declared the winner. All this argument about how to decide that winner is a side note. FIRST should come up with a foolproof method of determining the winer.

Re: Team Update #18
 

The spec is TRIGGERING the TOWER. The spec is not getting there first.

If 254's official towers use the design in the published field drawings, and the minibot does work on them, but not on the modified design tower switches, then I'd say we have a problem.

Re: Team Update #18
 

To lift the entire plate evenly it will take more than 4N.

It's a simple second-class lever. I thought that was obvious so I didn't stated it explicitly. My bad.

Re: Team Update #18
 

OK fine. Let's throw out the last paragraph of my post and concentrate on the rest. What is required to TRIGGER the TARGET? If I don't know the answer to that question, how do I design a minibot to reliably TRIGGER the TARGET? Does it take 4N of force? Does it take some force for a specified period of time? Do I need to hit the plate in a certain area? Do I need to tap out morse code on the plate when I get there? I don't care what the specs are, just let us know so we can design a device that meets those specs.

Re: Team Update #18
 

Where are these "published field drawings" I keep reading about?

2011 Game Field Elements_RevA.pdf mentioned in Team Update #18 does not contain any information about the sensor, or its location, or any details concerning how the signal is processed by the computer equipment to which it is attached.

Re: Team Update #18
 

The TARGET is described in the ARENA drawings. As far as I can tell, the only way to make sure your MINIBOT will TRIGGER it is to build it and try it--or do a detailed analysis of how the plate will behave when struck by your MINIBOT. The 2N to 4N thing in the ARENA description gives you an idea of what's required, but you still need to look at the design of the TARGET to be able to figure out exactly what's required.

As for the sensors...I guess we just have to guess at how it works, and make sure the plate will move a long ways when the MINIBOT hits it?

Re: Team Update #18
 

There's the rub. How does FMS read the signal. What if FMS is busy doing something else when the switch is hit?

What are the wires from the switches attached to?

Are they generating an interrupt and being time-stamped? Is that happening locally and the information stored so that FMS can read it when it gets a chance?

The engineer in me wants to know.

Re: Team Update #18
 

I honestly don't think this is as big of a problem as it is being made out to be. So far as I could tell at Pittsburgh, the sensors were pretty darn reliable. The only times they did not trigger it seemed like 1 of these 3 were the causes:

1.) Hit the bolt
2.) Did not hit with enough force (like, creeped up really really slowly with the wheels slipping)
3.) Turned off / reversed immediately after hitting the bottom plate without pushing much at all

If your minibot avoids all those 3 things and are still getting false negatives during the practice day at Week 3's: talk to a FTA. They will help you work out whats going on. I think the practice day will be plenty of time to sort through those 3 issues and any others that come up. Like I posted before, the triggers are really simple, there are only a few possible causes for them not to go, so I really don't think it will be necessary to hit in an exact spot, other than not bolt, or with a super specific force or time (or tap out morse code :P)

Now, I must say that I am never crazy about automated scoring systems. I wouldn't have designed a minibot race at all because of how hard it is to score without some automation. However, this is the game thats been given to us and you can see my earlier posts about why I think that the automated scoring is necessary given this game. And only because I believe the automated scoring system is necessary am I advocating for teams to take whatever steps necessary to make sure they can trigger the setup.

But still, my gut instincts say that if those 3 little things are avoided, this whole thing will just not be an issue.

Re: Team Update #18
 

You don't have to lift the entire bottom plate. You just need to tip one side of it up in order for the switch to make contact. A naive analysis says it would take between 1/2 and 3/4 the weight of the plate to trip a switch, or -- hey, lookie here -- about 2-4 newtons.

Re: Team Update #18
 

I haven't witnessed it, but others have posted here that not all target triggers were reliable, ie: targets that should have triggered in week two didn't.


I don't agree that reliable automated scoring is difficult. I was a Cub Scout leader several years ago, and our pack had a home-made Pinewood Derby track. The scoring was automated and home-made. A simple beam break trigger and a timer. It was 100% reliable (I never saw a failure to trigger in 4 years of races).

Re: Team Update #18
 

You're a little late to the party. That point has already been made.

Re: Team Update #18
 

I dunno, having only been at a Week 1 regional I haven't witnessed towers triggering properly.

But clearly there are some pitfalls to be aware of, just make sure they are addressed. IF someone witnesses a reasonable hit not being recorded in Week 3, please post a video here. Then we can argue the point.

Re: Team Update #18
 

Assuming the problem has to do with the time it takes to TRIGGER the field system, not the force exerted, it seems that some means of controlling the deceleration at the time of impact with the TRIGGER plate, might help. Perhaps add a springy finger to the top of your MINIBOT?

Re: Team Update #18
 

Great idea. And simple to implement. Perhaps just a small loop of surgical tubing at the point of contact would do the trick.

Re: Team Update #18
 

I was thinking of a thin section of large diameter PVC pipe. We use a few of these on the bottom of our (heavy) MINIBOT to cushion the fall back to the BASE.

Re: Team Update #18
 

Or FIRST should fix their system. Why should the customer fix a part that came to them out of spec? We are paying for something and if FIRST doesn't deliver the onus should be on them to fix the problem, not on us to work around their broken implementation.

Re: Team Update #18
 

The spec is that your MINIBOT has to TRIGGER the TOWER. They didn't specify exactly what it takes to do that. But you still have to do it. If your MINIBOT does not do that, then modify your MINIBOT so it does. If that means slowing it down....oh well.....

Re: Team Update #18
 

Nominal measurements, nominal specs, last-minute changes to fix last-minute problems... These are common in FIRST, have been common in FIRST at least as far back as when I started 827 and got crushed by 254 at the Silicon Valley Regional.

Some level of overengineering, and not engineering things too close to your target, is required all the time, methinks.

Re: Team Update #18
 

The discussion seems to be divided into two groups: those who have a fast, high-impact minibot and have experienced first-hand a target not triggering when it clearly should have, and those who do not have this first-hand experience, but speculate on whether the system is or isn't reliable. I put more credence with those with first-hand experience.

Regarding "the bolts"....it seems that they should have been fixed to the lower plate, and moving through the upper plate, so that they move with the lower plate no matter where it is contacted, rather than being fixed to the top plate and protruding through the bottom plate, so that they create 4 areas which block the minibot from triggering. From my understanding of the drawings, this would be an easy change to make.

Re: Team Update #18
 

I put more credence with those who read the rules and understand what the rules require the TEAMS to do.

There is no "should have TRIGGERED the TARGET" in the rules. You do it, you get credit. You don't do it, you don't get credit for it.

Part of the design challenge is to make sure your MINIBOT TRIGGERS the TOWER. Some TEAMS seem to have missed that. Apparently they thought the challenge was to get a minibot up the TOWER as quickly as possible.

Re: Team Update #18
 

IF the triggering programs and times of triggering on the towers hadn't been changed by FIRST between week 1 and 2 then this wouldn't be a problem for me. What is a problem is that it is changed to a level of "it seems to be working" instead of "every minibot who goes up the pole and exerts the 2-4N of force will trigger the top".

And to top it off humans wont' be used as the backup when electronics fail... seems backwards to me.

Re: Team Update #18
 

That would be an example of speculating about what should happen, as opposed to experienceing what did happen. Surely you are not suggesting that there is no possibility of a malfunction.

Re: Team Update #18
 

I think this is an unfair categorization. I was standing on the side of the field at the end of almost every match at a week 2 regional and saw every minibot that did not trigger the post. There were not that many of them and of them and it could not be shown that they did not hit a bolt or did not turn and reverse before pushing up the pad.

If someone has a really good first hand account of a minibot that definitely did not hit the bolt and definitely did not bounce off too fast and still did not get triggered, please speak up so that the cause can be investigated.

Re: Team Update #18
 

If thats the case, then why did the first two weeks of the season use manual scoring to determine minibot race winners? There's no way for a human to know if a trigger was sensored from the ground. The GDC referred to the endgame as a race, we all refer to the endgame as a race, because it is a contest for teams to get their minibot up the tower as quickly as possible. Yes there are constraints, size, weight, motor usage, battery usage etc. One of those is the 2-4 N measurement the GDC added in to the game documentation.

The sticking point to me is that some teams clearly are exerting that force on the top plate. Maybe some of us feel more comfortable calculating the exact force required, but I'm comfortable enough with our minibot to say it hits the target hard enough to trip the switch. Now, if our minibot hits the target LONG enough to trip the switch is the big question because there is no requirement in the rules about length of time required to trip the switch.

If the GDCs intention was to make the entire challenge not just complying with the deployment rules, minibot part usage and the actual challenge of ascending the pole, but to also include the tripping of the sensor as a key factor, there should have been much more information provided besides "2-4 N". Whether this includes an approximate timeframe this force needs to be applied in or something else. To have every team interpret the field drawings of the tower sensor to determine if their minibot trips the sensor or not is insane to me.

Thats my only gripe. Now I don't know how big of a deal this all is. Al has told me to rest easy, and I trust Al, so I will do just that. I'll also be making sure our minibot runs are recorded on video just incase a sensor doesn't happen to trip. At least we'll be able to try to break down why it didn't that particular run.

-Brando

Re: Team Update #18
 

This observation received very little attention in this thread. Could it be a significant factor in the reliability of the triggering?

It seems it would be possible for it to bind on the pole, especially if the contact point between minibot and plate is far off-center causing the plate to tilt. Does anyone know what is the coefficient of friction between polycarbonate and galvanized(?) steel?

Re: Team Update #18
 

I believe speaking up about such an account is exactly what several posters have done.