Daniel's Work Thread: Toward a Hackerspace Scanner

[Fab52]

Hi Rob,

The problem with the press fit bearing is in the instruction we will give to the CNC owner.

If they try to cut it over 400 ips with a 1/4" bit at 0.250" depth(see G-Wizard), we will end it up with a swirl pattern, instead of having 22mm (0.8661417322844001") we ended up with something like mine 0.880", so no press fit at all but 0.014" of a play...

We have to understand, the CNC guys who do custom work are there to make money and they try to go as fast as possible, don't they ???

See you

Fab

[jck57]

Well, so far I've cut one scanner using the one bearing solution, the one where it's just a press fit and you use nuts as spacers. I'm going to be doing some more in the following weeks. If I hear that there's a problem, we can start looking at these other solutions. There's certainly many to choose from!

Rob, that set-up looks fine as long as you get a good press-fit. Only thing I would do is put a washer against the carrier before the nuts go on. Did it feel any different working the carrier with single bearings rather than double bearings?

[rob]

The press fit I'm getting is such that it nearly requires an arbor press to get the bearings in. Once they're in, they're not going anywhere. After I installed such an arm on my machine, the movement was as smooth as the two-bearing system.

This will work for baltic birch, but what about softer woods?

As for Fab's concerns... I'd like to put a tolerance in the files, maybe +/- 0.005". Otherwise you can get CNC operators who just speed through the cutting without any concern about quality or accuracy. The vast majority of parts will probably end up being -0.005"/+0.010", but for things like holes and bearing pockets, the tolerance needs to be tighter.

[jck57]

The press fit I'm getting is such that it nearly requires an arbor press to get the bearings in. Once they're in, they're not going anywhere. After I installed such an arm on my machine, the movement was as smooth as the two-bearing system.

This will work for baltic birch, but what about softer woods?

As for Fab's concerns... I'd like to put a tolerance in the files, maybe +/- 0.005". Otherwise you can get CNC operators who just speed through the cutting without any concern about quality or accuracy. The vast majority of parts will probably end up being -0.005"/+0.010", but for things like holes and bearing pockets, the tolerance needs to be tighter.

I took a 7/8" spade bit and ground the sides til I could make a hole with a good press fit. Tried in on OSB and the bearing was well secured after pressing with a vise. I suspect cabinet grade fir plywood would be OK too.

BTW, did you see my post in this thread about laying out pieces 1/4" apart?

[daniel_reetz]

I saw it- that's exactly what I'm working on right now, and I think you're right - it's the smart way to go.

There are two issues with that approach.

One, the bit is loaded around half its circumference, which is a bit hard on the hardware (tho I don't care).
Two, the pieces don't end up with much of a support matrix, so either you need a vacuum table to suck them down to the cutting bed or you need to "tab" the heck out of them (leave small connecting pieces between the pieces that later need to be removed).

Tabbing enough will solve the problem. It's definitely the least-waste solution; I'll share some artwork tomorrow when I get back home. One side benefit of the 1/4" approach is that the common-line cutting has the potential to divide our total line length/cutting distance by a third or more, so I am aggressively figuring out how to do it, including changing part shapes or not actually cutting outlines of parts, but rather cutting a line pattern that is continuously moving but results in parts close to our desired parts. Examples will make this blathering mess of a paragraph clear.

jck57, Rob, everyone, I don't want to come off the wrong way, or as dismissive or too delegate-y. A lot of great ideas and advice have been posted here, and Rob has spent significant time explaining to me (in email) a lot of the CNC stuff that I'm re-discovering while playing with tools like GWizard and Matlab. For whatever reason, working on it plus hearing the good ideas here (sometimes only hearing them in their fullness after I figure them out for myself) is what's working for me right now. I've been doing a lot of late night installs and working some crazy hours and it's sapped me and cramped my usual working style. Anyway cheers to all the great stuff here.

[rob]

You could take a break and make a vacuum table. At NextFab, they threw one together in a few days, and it's powered by a shopvac.

Although, yeah, you probably don't need a distraction right now!

[daniel_reetz]

new layout.jpg (687.72 KiB) Viewed 785 times

For some of these pieces, it is really hard to see how to reduce the waste. Cranking away here.

[daniel_reetz]

So, I'm now spending some time with the "back" of the scanner. In my early designs, the piece on the left was "notched out" to accommodate the user's hand turning book pages. However, ultimately the scanner was rotated around and the new "back" was left as it was originally.

To simplify machining and reduce wood usage, I'm going to suggest that we make the following change:

simplification.jpg (330.08 KiB) Viewed 783 times

Can anyone see any problems with this? I like the look of the new part less, but only slightly. I think aesthetics are very important, but again, this is the back of the scanner.

[rob]

Go for it. It's the back, nobody's going to see it. And the front supports are straight in the front anyway.

[jck57]

This program may be of interest: http://samtecsolutions.com/optimizer_for_anyshape_profile_nesting_software.asp

"Optimizer for Anyshape is Professional, Profile (rectangular/non-rectangular) Nesting Software specifically designed to produce optimized cutting layouts, minimize material wastage and maximize productivity."

Free download of demo version.

Many similar programs available. Search for sheet layout cutting optimization.