Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - cjbruce

Pages: 1 2 3 4 5 6 7 [8] 9 10 11 12 13 14 15 ... 49
141
I think that your number 1 priority would be to fix the A40 motor, it has been an issue for a couple of builds now

Could you send a screenshot?  I'm not seeing the issue in the inspector.

142
I'm hammering out bugs now.  I want things to be at least as stable as they were for the 16 January build.

143
What changes were made to physics? And what was the reason behind it? Feb 16 had it pretty spot on, so I guess I'm just confused as to why that needed adjustments...

Prior to the most 28 February build, there were two scaling factors in robot reconstruction:

1. massScaleFactor - Introduced in the very beginnings of the robot workshop.  This was originally set to 0.1.  We had some difficulties very early on trying to get driving to feel good for heavyweights, so this scale factor was introduced to bring the mass of the chassis down to approximately 10 kg.  This just happened to work well with wheel masses of 1 kg.
2. sizeScaleFactor - Introduced about a year ago when we started focusing on beetleweight driving for Bugglebots.  At the time, I couldn't understand why heavyweights drove fine, but beetleweights would randomly float and flip on their sides with no user input.  The only thing I could do that seemed to have any effect was to increase the scale (but not the mass) of the beetleweights so that they were similar in size to heavyweights.  This created all sorts of problems, so I settled on a scale factor that depended on the weight of the robot, but made typical beetleweights be approximately 1.5 times their original size.

In the 29 February build, I set both mass and size scale factors to 1.  We had found and eliminated the root cause of the problems that the scale factors were designed to address. 

The massScaleFactor created a part management problem where we had to manually adjust component torque and force values.  This was prone to error, and causing all sorts of random problems and inconsistencies now that Wham is creating a bunch of components.

The sizeScaleFactor created a problem where what you saw in the robot workshop didn't match what you saw in the arena.  In general, robots were way too big for the arena.  Moreover, scaling was inconsistent between robots.  One robot's A40-300 motor would be a different size than another robot's A40-300 motor.  This made it very difficult to create arenas at the correct scale when every single robot was scaled differently.

I am confident that eliminating the sizeScaleFactor was the right thing to do.  I am less confident that eliminating the massScaleFactor was the right thing to do, as it has exposed a few more things that I didn't anticipate.  One of them is an interaction with another bit of code that causes vertical spinners to more easily flip backward (Circumvolution and S3).  One of them is the fact that the increased weight is pushing down harder on the wheel joints, causing robots with low clearance to scrape against the floor (Nuclear Crisis).  I think these things are fixable, but it is a work in progress.

144
It seems like there's definitely something up with my vertical dual-bar spinner.  A combined 400 kJ of energy doesn't seem to do any damage to the opposing bot, yet any collisions made against the opposing bot cause my robot to flip over.  It's almost like no energy is being transferred to the target with the vertical spinners.

It makes me wonder now what would happen if I made it a diagonal spinner...

That might be due to bite.  How many RPM are your bars spinning, and what is their radius?

145
I have an idea of what might possibly have changed.  I’m not convinced anything has, but I’m not convinced it hasn’t either.  We are, yet again, in firmly undocumented Unity Physics Land.  I changed the masses of all the rigibodies.  My assumption is that this automatically changed all of the moments of inertia.  It is possible that Unity did not do that.  I will need to do some testing to find out.

146
Spinners in Feb28 have something weird going on with their mass or something. My verts start to wheely when their weapons are spinning.

It's by far at its most noticeable with S3. Drive forward with the weapon off, it's fine. Drive forward with the weapon on, and it's more overhead thwackbot than vert.
  [ Quoting of attachment images from other messages is not allowed ]

Is it the reaction torque that occurs when the weapon starts up?

Circumvolution is hard to drive too.  It tends to flip on its back a lot.

147
Gotcha!  Sorry.  I’m a little dense.  We can put it on the list.

So you were thinking of building a really big robot in its entirety, then scaling the whole robot down to make weight?

148
Maybe a picture would help?

Why do you want to change the scale of the parts?  Aren’t they scaled correctly now?

149
I guess I still don’t understand what you are trying to accomplish.  What exactly is the problem you are trying to solve?  Things are now scaled correctly.  Whatever looks right in the workshop now matches what you see in the arena.

150
A scale factor of 1.5 (typical for many of the robots in the 110 kg range) would make the robots mass increase by a factor of 1.5^3 = 3.375X.  This would put the robots at 371 kg.

151
Also, please understand that if you scale your 110 kg robot up you won’t make weight.

152
Or another case:

1. I make a drivetrain with motors, wheels, and gearboxes.
2. I scale the robot up.
3. Now the wheels don’t reach the ground, and they are completely inside the chassis.

There is all sorts of weirdness that would occur with a scale slider because some components are not scalable.

153
Hmmm.  That would be kind of weird to implement.  Some things can be scaled while others can’t.  Things would definitely come out looking strange.  Here’s an example:

1. I create a robot that has batteries arranged in a brick.  Each battery is immediately adjacent to the next battery.
2. I scale the robot down.  Batteries can’t scale down.  Now the batteries are overlapping each other.

154
There are some issues with the new custom wheels. When I replace my old custom rubber wheels or premade wheels with the new ones, some of my bots which used to drive well starts drifting, oversteering and bouncing.

For another, when my 4WD or 8WD bots only have 2 or 4 of the motors running while the other motors shut down, if they are attaching to the old custom rubber wheels or premade wheels, they work fine. But if attaching to new custom wheels, the bot will fly to the air when driving.

Here is one of the examples.

  [ Quoting of attachment images from other messages is not allowed ]  
  [ Quoting of attachment images from other messages is not allowed ]
I've talked about this issue at P93 before, and ya, it still exists. And here's a more obvious example.
  [ Quoting of attachment images from other messages is not allowed ]  
  [ Quoting of attachment images from other messages is not allowed ]  

Btw, we kinda have demands of the size scaling factor cuz we've made numbers of bots that are normal-sized in previous builds but now they are slightly a bot small. Also some replicas has many tiny detailings that we cannot make the bot small. It might take us a long as time to rescale them, but with the scale factor we can save a great amount of time rescaling them. It's like, it can rescale all the components that are able to resize(like chassis and all customed stuff) together with other components like motors, batteries and default wheels etc stay unchanged. Would it be something that can be realized?

I'll take a look tomorrow, but I'm not quite sure I understand the problem.

1. Scaling has been broken for a while.  This build fixes that.
2. What you make in the robot workshop now matches the scale of the arena.  What you see in the workshop is what you get.  This wasn't true before.  The "Normal" you saw before was wrong.  Robots were scaled inconsistently in the arena based on their weight.  That is fixed now.
3. You shouldn't need to rescale anything.  Everything should be as it was before, just now correctly sized in the arena.

155
I noticed with the latest build that the blade-to-gear axle connection on my custom bar spinner is very weak.  Then again, it is also capable of spinning up to 200 kJ in less than 20 seconds.  Is there something I'm missing durability-wise?  It's actually a 67 kg bar  :confused:

Actually in general, I noticed that all components seem much weaker, even the internal ones like the motors and batteries.  They seem like they break off a lot more easily, which is awesome!  But I'd think the weapons would be a little more robust.

I agree.  I’m looking at the damage one of our robot’s wheels took from their last fight from a drum spinner.  Both wheels cracked (two separate hits) rather than breaking outright.  The in-game model would have them ripping off from a hit from an A28-400-powered 20-pound drum at 1:1 gear ratio and 24 volts.

156
The new arenas are awesome! Also the sliders inside the workshop makes it much more convenient to build high-customed-part-bots, really appreciate this one. And there're some issues imo of this new build.

The bots are not back to original size, but slightly smaller than those in previous builds. Maybe it's for other purpose but Idk.

Those new weapons look nice, but the weight might not be that suitable. 10kg for a flywheel, 0.5kg for an axe is a bit too fragile and lacking of energy. Resizing is still not enough. If we could tweak the material thickness of them that would be better.

The damage is unrealistically high in the new build, bots seems to be fragile as plastic, which makes me have to tweak the damage slider to 5%, but even so the small weapon parts like spinner teeth still easily fall off.

Also the bouncing problem of new custom wheels is getting more severe. And bots keep give out noises while driving.

Glad you like the sliders!  Those are Whammet and tashic’s work.

The bots are actually back to their original, realistic sizes.  They haven’t been correctly sized for about 12 months now.  I introduced an upscaling factor a looong time ago to attempt to solve beetleweight driving problems. It turns out the root cause of the problem had nothing to do with scale (but it was exacerbated in tiny robots). The root cause has been fixed, so the size scaling factor has been removed.  What you see now is to the correct scale.

I’ll let Whammet tackle the new components.

Noted about damage.  I’m at our school competition now.  There are about 30 spinners here.  Things definitely aren’t breaking as readily as my damage model would suggest.  Granted, these robots are only 85 pounds and not as powerful, but still.

I didn’t realize there was a bouncing problem with the wheels.  Would you mind sending an .RR2Bot file?

Thank you for all of the great feedback!  I feel like we are so close on this!

157
I just enabled downloads on the 16February build.

158
kix,

I’m at our real-life robotics competition today and tomorrow, so I won’t be able to do much fixing.  I can re-enable an older build on itch.io though. Is there a particular build that I should enable for your tournament?

159
New build is up! 28February2020:

http://robot-rumble.itch.io/builds

This build mostly just fixes the scaling bugs introduced in the 27February2020 build and pointed out by kix.  The scaling factor was based on weight of the robot and was designed to address some physics issues we were seeing that were more severe at the lower weight classes.  The physics issues were caused by improper COM calculation.  COM was fixed, but the scaling factor was not completely removed until this build.

​[Added] Added sorting options for the list of components: - A-Z and Z-A : should be self explanatory, it uses natural sorting (numbers are compared by their value not alphabetically). -Weight: also self explanatory, for components that can change their material the default material is used. -Size: it calculates the bounding box of the entire component (based on colliders), then the dimensions of the bounding box are added and that is the value that is used to sort the list.​

[Bug Fix] Restored the flipper piston to its previous torque value.

[Bug Fix] Fixed the problem with the incorrect mass of spinners in the telemetry screen.

[Bug Fix] Fixed the robot size/scaling problem.​

160
Fixing the following:

1. Restoring the flipper piston to its previous torque value.
2. Fixing the problem with the incorrect mass of spinners in the telemetry screen.
3. Fixing the robot size/scaling problem.

I'm heading to bed, but should be able to push a new build in the morning.

Pages: 1 2 3 4 5 6 7 [8] 9 10 11 12 13 14 15 ... 49