Welding experiment success!

A few weeks ago, I posted to the mailing list for a Welding / Fabrication class, and we had a few people show up to try their hand in making a new addition to the shop, a custom-built shelf for our welding bench.  The old shelf is pretty sad, if you've ever had a chance to meet it. 

It's falling out of the wall
I mean, just look at the thing.  While it served its purpose without complaint for many years, it's Ikea roots definitely show through.  The shelf has perpetually had a 5° slant as long as I've known it, so it was certainly never confidence inspiring enough to do pullups on. 

We're always in a state of flux at the shop, making small improvements here and there as we see necessary.  It all adds up after a while, and people who haven't visited the shop in six months are usually stunned to see how much things have de-Seussified in the interim (in fact, that happened just now when RJ walked in, haha).  Little upgrades like this make all the difference in the world, when you were used to staring at the eyesores like this one.

Looking into the Tested videos shown on the prior posts, there is a neat video on Youtube of Jamie Hyneman's (of Mythbusters fame) workshop, specifically on the custom racks they have holding their boxes of equipment in M5 Industries.  He says, "We buy tubing by the ton".  Well, we just so happened to have a bunch of tubing scraps at the shop just begging to be put to use.  Off to Solidworks!

I quickly drew a few 3d sketches, and then used the Weldments tool to turn my sketch into a solid model.  After a grand total of maybe five minutes from concept to finished model,  then I could use to generate engineering prints, Bill of Materials, mass information like weight and volume, even simulations for deformation and drop testing if we really needed that.  While having all that strength is useful if you want it, what is most important to me is ease-of-use, and how intuitively I can learn new concepts in software.  I went through the weldments tutorial once, beyond that, this is actually my first real-life Solidworks weldments project.  

This was entirely modeled up in a virtual environment before I needed to make a single cut, so when I made a few revisions to the size of the tubes in the sketch, the rest of the model updates instantaneously (along with any other associated information, like the prints and BOM).  Not having massive experience with welding, ANSI-complete prints, nor manufacturing management, what I found this tool was most useful for, was the ability it gave me to convey the important information about this structure to my students taking the welding class.  I simply handed them a set of blueprints with a cut list, and told them "All the relevant information is on this paper." 

They delivered.  I made the first two cuts and welds, the rest was up to the new guys. 

The Noobs followed what few instructions there were, the hardest part being getting a "feel" for welding.  It's something that can't be taught out of a book, rather, it's an art that needs to be practiced.  The biggest problem we encountered was heat management, understanding what's changing in the system when you start welding, and how the molten puddle of steel needs to be manipulated through the welding process. 

Welding steel is not much different than a hot glue gun and popsicle sticks, except it's much much hotter, and it'll melt the popsicle sticks away beneath the glue. With a few basic concepts like that in mind, as long as you're considering what's happening to the heat in the weld, then your results will show a little bit of insight to the process.  Mild steel is a moderate heat conductor, it uses a mid-range heat value (between stainless and aluminum), so it's relatively simple to work with.  Stainless, while it conducts heat much less than mild steel (requiring less heat overall), also has a higher coefficient of thermal expansion, so if you don't carefully tack down your weld in several spots, a whole piece of stainless will bend and bow as you weld it along the entire length, ending up distorted and warped.  Aluminum is another beast altogether, and not recommended for beginners to fabricate with.

Tack welds holding everything together
This whole project was built from scrap and leftovers laying around the shop.  If you had to go to the metal supply store and source all of this material on your own, I'd be surprised if you'd be $20 deep into it.  The worst part of this whole project is dealing with that expanded metal grating.  While it's wonderful for filling in open areas like the tops and bottoms of this shelf, it's pretty nasty stuff to handle.  Even the "flattened" grades of expanded metal are covered in lots of tiny sharp edges that will cut the shit out of you the second you turn your attention away from it.  I have to find out these things the hard way.  That's why I order $100 of material at a time, so Benner Metals will deliver the order to me instead. Twenty foot lengths of steel aren't too easy to negotiate, let the flatbed deal with it. 

The mounts had to be reoriented 90° from my initial drawing, mostly due to my lack of considering how much space the grating was going to take up.  Since this part is being mounted to a block wall, I had to get some 3" sleeve anchors and a carbide tipped masonry bit to drill the pilot holes.  The mounts started life as a small leftover piece of rectangular tube from Flea's jeep bumper project, which I quartered into nice flanges and drilled a .400" hole through them.  Once tacked on, I laid a weld bead on the butt joint between the mount and the frame.  I could have filled it in better, but I think a weld of that size would probably exceed the design limits of this part.

Dykem is also known as layout fluid.  I use this all the time to mark parts based on a measurement, to see where I need to make my cuts or holes.  After measuring the parts using calipers, then I gently scribed the intersections of the horizontal and vertical midlines to find a rough center for the mount holes.  Dykem is easily washed away once you're done, using acetone.  In a pinch, Sharpie marker works fairly well, just remember that Sharpies are ruined once you put some oil on the tip, and you can pretty much count on these steel parts being slathered in a light coat of oil to prevent oxidation.  Acetone, you'll come to learn, will be your best friend when working with metal, except when you have even the slightest cut in your skin, which the acetone will seep into and light up any exposed nerve ending with a nice bright searing pain, not much different than squeezing lemon juice all over your broken cuticles. 

The end result - total amateurs (including myself) successfully made a beautiful, custom shelf for the shop.  It's way overbuilt (the way I like it), cheap, and user-servicable.  Want to add some hooks?  Weld them on.  Want more shelves?  Weld them on.  Speaking of hooks, remind me sometime to tell you the story about why this wall is red... 

The top crate weighs +60 lbs.  I'd say this is pretty solid.

My fourth attempt at 3G (vertical up) welds
the finished result - pullup tested. 


Building the LED Pixel Box Part 2: LED Selection

Because the original Tested video had no details of what was built besides what was shown/talked about in the video, I had a few questions about what exactly was used. 

One of the first questions I had when I saw the video was what size box was that and what LED strips did they use.  There are tons of different types of RGB LED strips out there of varying densities and control types, and I have played with a few of them.  So I went about figuring out what strips they used for the box 

The first step is to determine how big to make the box. Taking a look at the finished box in the video I can see the box is square, and it goes from about the armpit to mid/lower belly of the hosts.

Judging by this dimension, and estimating based on the humans I have available. I would say this about 12 inches. This sounds good to me if I am going to apply some DFM principles in this build- lots of raw materials come in multiples of 12 inches (wood, acrylic sheets, lighting diffusion paper etc) so making this box 12" square should make acquiring materials easier, and with less wasted scrap in the end.

Now, to determine what RGB LED strips to use. Adafruit sells many different options available, and sells LED strips at a decent price products are shipped from the US. They also provide very nice product pages with good information and links to good sample code. This isn't about recommending one vendor vs. another, but they are a good starting point with a good reputation, so they are worth consideration. 

The 60 LEDs per pixel and 30 LEDs per pixel "neopixel" strips look nice, but they are not something Ican use. I want to use a RaspberryPi to control these strips, and according to Adafruit's product page:
The controller chip ... protocol used is very very timing-specific and can only be controlled by microcontrollers with highly repeatable 100nS timing precision ... t it will not work with the Raspberry Pi

Well, there goes that option, so now I move on to LED strips which are controlled by external (not on LED) controllers, such as the HL1606/LPD8806/WS2801  as I see on the 32 LEDs per pixel strips. I used a strip like this on the Mobile Club tabletop and strips like these are very common and easy to find on hobbyist sites such as Sparkfun  and Adafruit, as well as ebay and alibaba/aliexpress. There are many different types of strips, but the main variations are number of LEDs per meter, and the level of addressablilty for the LEDs on the strip (each LED can be a different color, a certain amount of LEDs have to the same color, or all LEDs have to be the same color).  For this project, I will need to make sure I can address each single pixel and make it a different color.  You will see LEDs that are "3528" and "5050" The 5050's are brighter. Look for those.

I need to determine the proper LED density.  I know I want the Pixel box to be 12 inches square and be 16 LEDs by 16 LEDs. So 12 inches per 16 LEDs leads us to 12/16 = 0.75, so I know I will want a LED every 0.75 inches.

A strip with 32 LEDs per meter would be an LED every 0.03125 meters, this translates to every 1.23" this is too far apart.

A strip with 60 LEDs per meterwould be an LED every 0.01667 meters, this translates to every 0.65" this is close, but not perfect, and I would need to find a source for individually addressable

But, in looking around online I found LED strips at 52 LEDs per meter- that is an LED ever 0.0192 meters, which is about 0.75" - Perfect! These strips use the LPD8806 controller which will allow us to address each LED individually. I found these strips by search for LPD8806 LED strip and found good sources for these on both alibaba/aliexpress and ebay. In the end, I chose to purchase from an ebay seller based in the US, so I could get my LEDs delivered fast and get a start on the project by the weekend.

In comparing the ebay product shot with the strips shown in the video, they look identical (note the shot in the video is missing the silicone jacket used for waterproofing.

So, it looks like I found the strips they used in the video, and the numbers/sizing adds up, so I feel confident going forward with these parts.

The pixel box will have 256 pixels (16x16) so at 52 LEDs per meter, so we need 5 meters of this strip.

The RaspberryPi and LED strips have been ordered and I am waiting for them to be deliveryed. The next steps are to look into how deep the box should be and how to best diffuse the light.

Building the LED Pixel Box Part 1: Inspiration

Recently, tested.com posted a video of a LED Pixel box to their web page

This seemed pretty neat and up my alley in terms of skills. It uses RGB LED strips diffused trough a translucent layer of film much like we did for the glowing LED tabletop of the Mobile Club.

It also is similar to a project that uses of RGB LEDs as individual pixels - the Rave Rover:

I know there have been many different projects like this, but I think the time has come to make our own, and seeing the Tested video was the inspiration to start to build.

So, my goal over the next couple weeks is to build a Pixel Box like the one seen in the tested video.

Stay tuned for more posts as I document the build process!

First step.. I'm not sure if I like the name "Pixel Box" for this.... Does anyone have a suggestion for a better name?


Shop tip for the day - threading hard to reach bolts

Howdy y'all,

Playing around at the shop today, I had to utilize a trick which has come handy many times, but it took me many years to even come across this trick.  This may be elementary for some of us, but I still think it's pretty nifty.

 Mounting the new control box for the CNC mill today, I had a few hard-to-reach bolts on the case which needed a nut on the other side.  I didn't have room to hold it with my fingertips, and the nut would fall out of the bottom of the wrench if I couldn't get my hand in there.

A simple piece of electrical tape can help retain the nut on the wrench just enough to position it, and then let go once it's started to thread.  Take a small piece of tape and place it STICKY SIDE DOWN on the wrench.  The tape is the perfect size to take up the gaps in the flats on the wrench and gently grab hold onto the nut.

Press the nut into the tape, down into the hex flats, and your nut will be retained nicely without falling out.