More Heroica

Playing Heroica we set up a board where the beginning branched to one way to a long hall of high level monsters and keyed doors, and to the other way with keys and potions. The plan was that they'd go gather the keys and stuff, then go face the boss but my most driven son decided to forego the loot gathering and charge straight for the boss, relying on perfect rolls to get through the doors and beat the monsters, and darned if he didn't claim victory before the rest of us gathered half the things. So we decided we needed a new mechanic. A door to stop people from charging the boss. The key eventually became a switch, and I give you the Lego Heroica Boss Door.

Tiki Chess

Whystler's Tiki chess set was one of my favorites in the chess contest. Each of the pieces just had so much character and detail and it wasn't too much of a challenge to figure out which one was which, unlike some other sets. Every piece just works so well with the theme. It's a real treat to have this one. Click on the below image and just bask in the detail.

PS. notice the King's side knight with the broken staff. That's the one I was trying to reprint when my HPB burned out. I'll fix you... one day! (Edit: It's fixed now.)

Shells Vs Time

TheNewHobbiest asked me on twitter:

@CymonsGames I was reading your blog post on printing with additional shells. Do more shells shorten or lengthen print times?

That's an excellent question. Let's find out.

Fortunately I can find out without actually doing prints. The estimate function in ReplicatorG will give me something to compare.

The first subject I tested was Cinderella's Castle by SpaceCowboy850. I picked this one because I wanted something reasonably complex, but after I finished the tests I realized that they very small towers might actually fill in 100% with the loops. When that happens it won't put any more loops in that part adding no additional time. The second test subject was the Maltese Falcon by colinfizgig scaled down to be about the size of a chess piece. I picked this one because it's mostly round shape left lots of room for adding loops.

The procedure was simple. Slice the model with 0 extra shells and estimate the time, then slice the model for 1, then 2 then jump to 5 extra shells, estimating the time at each step and plot the results. Each pieces was made at 0.15mm layer height, 10% infill and 100mm/s feed rate with the altshell plug in turned on so the outside layer is printed slower than the rest of the layers. (That's more or less my default.) The results are as follows:

At first glance results seem pretty linear. The castle actually looks like it's making a slight drop, which can be explained by the towers not being able to take any more shells. But strangely enough the slope in the Maltese Falcon's line is also slightly less than linear. This of course makes sense after some thought since each shell is smaller than the last one. Numerically this breaks down to, with the Maltese Falcon, an increase of 5 minutes for the first additional shell, 4 minutes of the next, and decreasing as more shells are added, with the Castle it was an additional 14 minutes for the first additional shell, 13 for the next and decreasing.

While 5 minutes doesn't seem like much the thing to keep in mind that if you're printing a whole chess set those additional shells are adding 5 minutes per piece.

After all this I decided I wanted to how much time it would take to print with 0 extra shells but 100% infill, which produces a solid chunk of plastic, and the results were surprising. Estimate said the 100% filled Falcon would take 53 minutes and the 100% filled castle would take 3 hours 11 minutes. I had to check twice but for the Falcon after 2 or 3 shells there's no difference, time wise, to just fill the falcon in 100% with 0 extra shells. Clearly this isn't a universal rule because the castle did not have a similar result. So perhaps it's only small, solid, mostly round shapes? Either way if you're thinking you need to add more than 3 shells estimate it against 0 shells and 100% infill and determine if you can spare the plastic.

This was an interesting experiment, so thank you TheNewHobbiest for starting me on this journey. Knowing this will definitely change the way I print in the future. Probably. Maybe.

(PS. Updated the resolution comparison with a graph because I had excel open. Yup, it's a curve. You're welcome.)

3D Scanning Pick 2 list

Seems to be a lot of talk lately about 3D scanning on various forums I'm on so I thought I'd collect and condense my thoughts in one place where I can drive traffic [winky emoticon]. It is the next logical step after 3D printing. My take on 3D scanning, with the footwork I've already done, is that you have 3 options and you can pick 2:

• Fast
• Accurate/High Resolution
• Cheap

Fast and Accurate: ScannerKiller offers an amazing package that can do a very high resolution full body 3D scan in an instant. It may even be possible to capture video at the speeds they're doing. They use a method called Stereo Photogrammetry and pre-calibrating their cameras. Despite the precalibration the setup still need lots of cameras (16 by my count) and projecting a randomized texture on the subject with a flash to add texture to smooth objects, some of which may be applied to a DIY option. But if you're going to buy one from them the asking price is $17000. There's also NextEngine, endorsed by Jay Leno. $3000. And really it's not that fast.

Fast and Cheap: ReconstructMe and a Microsoft Kinect and you're good to go. Sure the result looks like it's covered in mud, but it only took a second and only cost you about $100 for the Kinect.

Accurate and Cheap: There are lots of options in this category.  123DCatch or Agisoft StereoScan use photogrammetry but can be tricky. For 123DCatch like apps you need a lot of photos to make it work (more than ScannerKiller). Plus if your camera isn't super accurate or if the lighting changes it could mess up the whole process. 123DCatch is free and works by uploading your pictures to their server where they process it, Agisoft has a similar app for $179 but you download it and processes on your computer. You could also download a free copy of the David Scanner software with a decently high resolution webcam and either a laser line level or projector. David Scanner has a great tool chain to stitch multiple models together accurately. But with all these options the subject being scanned needs to hold very still through the scanning process which can take minutes to complete. Not good if you're trying to capture a child for posterity. There are some other DIY options as well but their tool chains are less complete.

"Did you see Big Bang Theory?!"

For the record the answer is "No". I do not regularly watch the Big Bang Theory. "But it's all about nerds!" No, it's all about laughing at nerds. Just pay attention to the laugh track, which is another thing that bugs me. But despite me starting out negative on BBT, it does do a good job of honestly representing nerd culture so you can be sure that the nerdities will at least be accurate.

Their take on 3D printers was honest. They're big, expensive, slow, and it's pretty exciting when you first get it to cough up a whistle. They didn't go the home 3D printer route because that would have made the color prints they were making impossible. Then they spun the whole thing into a laugh about finances in a relationship and the clueless-ness of the nerd about the same. I wasn't happy with how easy they made 3D scanning look. A kinect does not take good scans and it's not easy to throw on a six pack. But over all it was a good day for 3D printing. I'm sure a ton of people after the show googled "3D printer" and learned alot.

If you haven't seen the show and can tolerate a laugh track you can watch the episode on CBS' website.

Makin' Hay With My Mad Skillz!

Can you 3D model? Then put down that controller, turn off the TV, open your Here's a couple of sites you can register for to turn your free time into money.

Desall.com has the idea of holding regular design contests. If you don't think you're a very good designer but want to get better these sort of contests are the perfect place to build and perfect your skills. That's how MCFrontalot got started, and look at him now.

Or you can go on freelancer.com and list yourself as a discount freelancer. Another great place to improve your skills while you build your portfollio. Just sign up and browse the most recent relevant projects, then bid for the jobs and make yourself some money.

If you know any other sites where you can turn your burgeoning modeling skills and spare time into income be sure to post in the comments.