doing some research for 2.008 instead of freaking out over 6.131 lab
Edit 10/17/2011: Found it! Thanks to Charles.
See explanation of “tilt sensors,” which include both mercury and rolling-ball type ones, here: http://www.ladyada.net/learn/sensors/tilt.html
(octopart search: http://octopart.com/partsearch/#search/requestData&q=tilt+switch&rangedfilters%5Bavg_price%5D%5Bmin%5D=0.54&rangedfilters%5Bavg_price%5D%5Bmax%5D=2 can maybe get it for $1)
http://smparts.com/-Switches-Rolling-ball-switches-%28AT%29/c0_2_561/index.html — but no prices listed, hah.
Hmm. So I know there was at least one previous design in 2.008 that used these switches. They put two on each side (top and bottom), facing outwards. Then, when the yo-yo spun, both switches would contact (and presumably switch on LEDs). Apparently they had issues with the thermoform shrinking and squeezing the ball switch and preventing it from switching though. All in all, I’m a bit iffy about these rolling ball switches. There’s no way to adjust their sensitivity (although I guess regardless PoV yoyo requires a microcontroller anyway, but the less pin-count the better), so what if it switches on when the the switches are rotated to (left right) and the yo-yo is slightly tilted? We’d be stuck with 100 useless switches.
I like Paulina Mustafa’s design of a flexure switch, she was prototyping this at MITERS (but I can’t find a blog for her). Looks sort of a like a ying-yang design, and the tips are flung outward and make contact with the sides when the yo-yo is spun. She’s using copper-tape, but I think something like conductive glue would be easier to apply.
http://www.helifreak.com/showthread.php?t=33567 “DIY Night Blades”
(spring inside a brass tube)
The batteries are wired in series with the LED and the centrifugal switch. The centrifugal switch is simply a tiny spring inside a brass tube. When you accelerate, the spring swings to the outside and makes contact with the brass tube turning on the lights. It is really quite simple and very reliable. This also makes the batteries last a long time since they are only on while you are driving and you will never forget to turn them off.
types of switches and design of switch contacts (including mercury switch aka “tilt switch”): http://www.allaboutcircuits.com/vol_4/chpt_4/1.html
still failing to find suppliers o.o
It’s official. CNC mills are sexy. Even decades-old bridgeport model ones that take giant floppies.
Oh yea, my partner for this lab is the awesome Cappie Pomeroy. He did all the lathe work o.o so don’t ask me about that.
So I got around to milling it.
|look, nyancat got fatter so that you can fit a 1/16” end mill between it’s crevices!
(The way I did it is I did a Tools>Sketch Tools>Sketch Picture, then resized that reference nyancat picture until it looked appropriate. I then put a Tools>Options>Grid>Major Grid 0.07 / minor grid off. System Snaps>check Grid. Also, make sure Units>IPS, InchPoundSomething. Then just trace it with lines).
However, my resizing and retracing ended up not mattering, since I decided to engrave it. This was because we had to make a 15 minute version, and no matter how I tweaked Mastercam (using 0.02” depth instead of 0.1”, using a wide endmill to clear it and then remachining with the 1/16”), there was no way the full 3D paperweight was milling out in 15 minutes.
|Yea so there are really dumb ways to make a path in Mastercam.
See the X’d out paths to the left — I was tracing all the paths by hand, essentially. Then Pat, the shop instructor, came by and showed me the magic of the the “polygon” tool which will automatically apply a function (in this case, engraving) to everything within that polgyon. It created some weird paths, like you get on a lasercutter which doesn’t recognize your vector as one shape but as individual lines*, but it gets the job done. In other words, he finished in 10 minutes what was taking me hours to do… hah.
*aka it doesn’t trace them continuously but rather jumps back and forth between different lines in the drawing
See link below for g-code / mastercam / solidworks files.
The jig was made by the instructors for the class, along with the “offset” used in Mastercam to make sure the blank is positioned correctly.
|gahhh technical difficulties. Yes, I used my phone to take a picture of picture on my camera. You really don’t want to know.
where “base/lab1” docs are for the CNC lathe and “engraving” docs are for the CNC mill.
Software used: solidworks (.sldprt, exported to .dwg), mastercam
Todo next: Mill out the actual thing.
I’m embarrassed to say how many hours I’ve wasted CADing nyancat paperweight for 2.008 now.
Off to go spend even more time producing CNC mill G-code using Mastercam for a machining appointment at 8am o__o
Aha, so I realized (thanks to Steve, shop instructor extraordinaire) that there are many things wrong with my CAD. Sadly, I chose 5 mils spacing for the grid I used to trace nyancat. Turns out our smallest drill bit is 1/16” ( over 6 mils). Eheh. Also, I was planning on going 0.3” deep, which would take forever to mill out using 1/16”, as it can only go ~0.02” down per z-axis pass.
|yea, milling fail.
|that, dear friends, is the look of a lot of milling time.
I was halfway through the infuriating MASTERCAM too. Lesson learned: avoid branching as much as possible. See that right-most sprinkle that touches the cat’s head? Caused me no end of trouble!
Next step? Infinite Mastercam’ing. Instead of being a nice cavity with extruded nyancat, it’ll probably end up being closer to pedestal with a nyancat on top done in thin layers. I’ll also use a larger diameter tool to pocket out each shape and then contour with the 1/16”. So, first up, redo solidworks using 0.7” grid spacing (resize reference picture accordingly — btw, found under insert > sketch tools > sketch picture).
I’ll stick the files up on github sometime soon.
Todo this weekend:
Saturday: Get ISP and PoV working, 6.131 lab writeup, MITERS misc. (project grants / safety / update website), PoV with partner 10am, Theater scene practice 2pm. Apply to summer jobs.
Sunday: 6.131 lab (1pm to midnight), 2.008 group meeting 7pm, watch a play (3 pm), go visit Sprout? (2pm)
Monday morning: lasercut cookies, document cookies and ISP.
Oh yea, speaking of ISP programmers, this week’s assignment was to mill ISPs using these desktop Roland Modela 3d mills:
So someone made an even tinier one that doesn’t need a USB connector (fits right into the slot): http://bardagjy.com/?p=628
Then someone riffed off of that and made one that has a break-off part for bootloading the chip: