June 20, 2018, 08:41:29 pm
Welcome, Guest. Please login or register.

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 - zacodonnell

Pages: 1 2 [3] 4 5 ... 33
31
Progress Reports / Re: Team Brain Damage
« on: January 23, 2018, 08:37:58 pm »
I decided to try calibrating the drive speed controllers with my transmitter stick to see if that would help before I rip them out of the bot and redo the settings. I'll still probably have to do that because I didn't change the low voltage cutoff and I'm running 5S A123, which might look like a really tired 5S lipo or a healthy 4S lipo to the controller. I don't want it "saving me" and costing me the fight.

The calibration made a huge difference. Before I did it I could easily get the left side drive to stop working. It was like it lost the signal from the rx and wouldn't pick it back up. Even turning the bot off and back on didn't always bring it back. I think the neutral signal from the tx was on the very limit of the arming line for the controller, so it didn't spend much time "in the window" to re-arm it. Playing with trims and the like didn't help much, though I could sometimes get it to start going and then stop.

Once I redid the calibration (one side at a time, still with mixing turned on in the TX) it seemed to fix the problem. I did a 2-3 minute test with the weapon pod out of the bot and it performed fine. My little wimpy test box isn't big enough to actually test it, but I could drive back and forth trying to not hit the walls or the end. I'll need to practice more to get used to it. I turned the steering channel down to 50% to avoid oversteering and that helped a lot. Spinning in place with the flywheel going and the steering channel all the way up will be some kind of silly gyro dance I'll bet.

video:
https://youtu.be/UE2u8fXp_nU

-Zac

32
Progress Reports / Re: Team Brain Damage
« on: January 21, 2018, 08:45:51 pm »
Ok, initial flip and drive testing is done. The flips are a little weaker than I want, but I think that's because of the cleats and frame being messed up. I have to replace the side frame rails to fix the damage done by the Balhogs saw bot last year.  The sensors work well and turn the magnet off ~11ms before the arm strikes the rubber stoppers on the frame. I've done a little more tuning since I made the video to make that window as small as possible. I put a white stripe on the flywheel so I could see how fast it's going. Mostly for academic and efficiency reasons - no practical implications yet. Someday I'll actually use the info probably.

The robot also self rights reliably, even at very low speeds, which is great. The sensor should reduce the over-throw on the self right attempts and prevent wasting the entire flywheel when I do it.

Now the bad news:
The drive system is flaky. The left drive motor seems to lose track of the signal and stop working intermittently. Sometimes a power reset fixes it, sometimes not. Sometimes it never connects, and other times it is working for a while and then stops. Could be a loose wire, but I'm betting it's an issue with the radio signal and the controller being picky about arming. I also have witnessed a few "runaway jitter" scenarios where the motors start jerking forward and back semi-randomly. They seem to come when spinning up the flywheel most often, but not always. It could be some kind of electrical noise but it is really bugging me. I'll play with the signal leads and extend the deadbands on the controllers to try to fix it. If I can't in pretty short order, out they come and back in go the old 550 motors. When the drive is working it is WAY FAST though. My little 4x8 test box with a dusty floor doesn't let me really see though.

I'm also about 5 ounces overweight right now. I knew it would be over, but I was thinking 3 or 4. I think I can still find the weight, but yuck. Worst case I could go to lipo batteries for moto, but I want to make sure the new drive motors work out before I do anything drastic.

Walkthrough video of some of my testing:
https://www.youtube.com/watch?v=Izz1zTKJPUk&feature=youtu.be

And self righting:
https://www.youtube.com/watch?v=At8CjyNKJIw

-Zac

33
Progress Reports / Re: Twisted Sick Robotics
« on: January 21, 2018, 03:09:51 pm »
I'm not sure I understand. Remember, I'm a computer geek, not a machinist :). Were you using a mill or a lathe? I have a single point boring bar for my lathe, but I can't see how the three point version would work in a mill without a rotary table. 

-Zac

34
Progress Reports / Re: Team Brain Damage
« on: January 21, 2018, 02:20:01 pm »
Thanks. I got it all installed and have done just a couple of tests. I didn't put any retaining compound on it yet, but I added it to my next order. Might be a little tricky to get the gearbox grease off of the parts before I use it - any pointers there?

They are WAAAY faster than the old setup, but I think I need to increase the deadband in the middle of the controls because it is tricky to get them to zero out and arm. It also seems like they don't both do it at the same point so I have to mess with trims a bit. When I set a dual rate on the steering channel it made the problem worse, and I am also seeing some jitters from electrical noise when I push the weapon throttle up. I'm hoping a wider deadband and some calibration will help all the problems.

I'm going to mess with some settings and put the little noise filters back on the wires and see if that helps, but the way it is now isn't workable for the competition. Slow and steady is way better than fast and unreliable for me. I've spent most of the day testing the weapon sensors and timing. I have a lot of high speed video that I'm using for analysis. I'm going to keep messing with it for a bit but I'll publish out a quick rundown once I make some more progress.

-Zac

35
Progress Reports / Re: Twisted Sick Robotics
« on: January 20, 2018, 09:01:36 am »
How did you do the internal cut on the UHMW? Rotary table on the mill would make it a breeze, but on the lathe isn't it hard to keep the chips cleared out inside the bore? Did you just make a square HSS cutter? I really like the idea of that kind of wheel.

-Zac

36
Progress Reports / Re: Team Brain Damage
« on: January 20, 2018, 08:58:22 am »
I haven't tried the brushless drive yet. The torquecalc says it should slip the wheels around 36A, which should be right at the peak efficiency of the motors and right in the wheelhouse of the 70A red bricks (I hope). I haven't done any real tuning on the simonk firmware so we'll see what happens. Did you use any kind of retaining compound on the pinions on your motor? I'm not sure the press fit is going to be man enough for the bigger motors. I was thinking something like Loctite 641. I'll try without and see if slamming the sticks back and forth can throw the pinions.

Backdriving the assembly by hand sure feels smooth compared to the 550 motors, and the build quality on these is WAY higher, obviously.

-Zac

37
Progress Reports / Re: Team Brain Damage
« on: January 18, 2018, 10:33:15 pm »
Today involved a little work on my house instead of my robot. It is all squared away now, and I managed to get the keyway cut on the new weapon motor too. The glue set nicely on the drive motors so I'll be pulling the old ones out and replacing them, as well as installing the new weapon motor this weekend. I did break an endmill but I had a spare (thankfully)

Pictures:
https://plus.google.com/photos/107279152885550274671/album/6510901572747399649/6512596638728282226

-Zac

38
Progress Reports / Re: Team Brain Damage
« on: January 15, 2018, 10:10:08 pm »
Some slow progress this weekend. I decided to try to see if I could steel (heh) some weight out of the steel in the clutch. It turns out that I can't, without weakening it tremendously. Once I took the outer pulley ring completely off I got about 8.5 lbs on the scale (vs ~18 lbs on the raw pulley). My tests with the one in the bot (which is part way between them) is about 13 lbs on the scale for this test.

Tonight I spent some time hunting for weight reduction measures now that plan A failed. I found a couple, so I think it'll be ok. We'll see I guess. Since I have to redo the entire wiring harness to replace the drive motors and controllers I decided to just go for it and find the weight later. That's delaying the time to test a little bit, but will save time in the long run because I won't have to take it all apart

Pictures:
https://plus.google.com/photos/107279152885550274671/album/6510901572747399649/6511476831480104274

-Zac

39
Progress Reports / Re: Mutually Assured Destruction
« on: January 15, 2018, 09:35:47 pm »
How mobile are you if the floor has seams or is made of wood? I had a lot of trouble keeping a heading when I was using a similar setup

-Zac

40
Progress Reports / Re: Team Brain Damage
« on: January 14, 2018, 09:23:35 pm »
Only if the weight faires show up. I spent today cutting down and prepping an alternative clutch plate that didn't have the steel outer ring, and it provides a significantly reduced amount of torque at the clutch. Just another indication I don't really have a good sense of how the magnetic fields are set up / what is affecting the effectiveness of the clutch. I'm right up against it on the weight.

I do have one additional pin available on the octopus though, so it could happen.

-Zac

41
Progress Reports / Re: Team Thrasher / Sunny Robotics
« on: January 14, 2018, 11:10:17 am »
Very cool. Has it fought any midcutters? The design is a little closer to the first undercutter I ever saw in person (http://www.lazytd.com/rc/rcg/hworf2.jpg) than the more common designs you see these days. Seemed nice and reliable in that video.

-Zac

42
Progress Reports / Re: Twisted Sick Robotics
« on: January 14, 2018, 11:05:28 am »
I'm curious to see if they will provide the benefits in this geometry that you need. With the flipper I'm throwing a bunch of weight forward so the bot tries to kick pretty much exclusively backwards, but with the hammer head rotating around the interaction will be much more complex. I think it will help a lot to prevent the bot from kicking backwards, but I'm curious to see how much it will help to keep the front down.

I like the gratuitous weight utilization in the shield on the front :)

-Zac

43
Progress Reports / Re: Twisted Sick Robotics
« on: January 14, 2018, 09:05:17 am »
I really like the counterweighted axe idea.  I see you borrowed my cleat idea for the outriggers on the back of the robot. I'm very curious to see how much kickback you get with this bot.

-Zac

44
Progress Reports / Re: Team Brain Damage
« on: January 14, 2018, 09:00:23 am »
So... I missed the Motorama registration window. You people and your facebook... I'm going to have to get used to this whole "more competitors than we can handle" thing.  Fortunately, I was able to get in on the wait list, and even though Magneato doesn't appear on the BDB, it will be there unless I blow it up on the bench or in my test box. Triggo is dead, with a couple of huge dents in the shell from last year and a warped baseplate / smashed front axle. Rather than rebuild it and continue to cause box breaches when I fight Big Ripto, I'm focusing on the Sportsman for now. It's a hard thing to say, since I've never gone to Motorama without a shell spinner since 2004, but life... happens. 

With that, I've started making updates for Moto this year. At Franklin I had to "kick it old school" and run without an onboard microcontroller. That was pretty rough on the guts of the bot, and eventually having the robot try to spin the flywheel up the second the clutch was turned off (but not necessarily disengaged) was too much for the weapon motor. Kyle pointed out I should just switch the motor to an inrunner to avoid the silly grub screw / press fit problem but the better solution is to coordinate the throttle better using a microcontroller.

I pulled the old control board out, popped the Arduino Nano off of the socket I made, and installed a new one I got here:
https://www.amazon.com/Arduino-Elegoo-ATmega328P-without-compatible/dp/B0713XK923/ref=sr_1_fkmr0_1?s=electronics&ie=UTF8&qid=1515938156&sr=1-1-fkmr0&keywords=elegoo+arduino+nano+3+pack

I skipped the ICSP headers this time and used a real soldering iron and popped three of them out pretty quick. After some bench testing I put the real software on there and went to work on the new breakbeam sensors. I'm using these: https://www.adafruit.com/product/2168  The sensors are used to detect when the arm is past the useful point in time for me to disable the clutch. Still need to test it at speed but I'm hopeful it will save my rear bar, rope, and flywheel energy when I miss.

I still need to wrap up the wiring, replace the damaged weapon motor, and do some testing. Stay tuned

Pictures:
https://plus.google.com/photos/107279152885550274671/albums/6510901572747399649

-Zac

45
Progress Reports / Re: Twisted Sick Robotics
« on: January 10, 2018, 06:36:42 am »
Yeah, I can bring an arbor press this year.

-Zac

Pages: 1 2 [3] 4 5 ... 33