Daniel's Work Thread: Toward a Hackerspace Scanner

[rob]

Nice work! You've pretty much got it to where I have it, modulo tabs. More = better. Since you have an up-cutter, I was surprised to see that you got fuzz on the bottom side. Usually up-cutters give fuzz on the top side. Maybe your bit is dull -- it doesn't take much. Is it discolored on the bottom? Does it shave your fingernail easily? In any case, fuzz is expected, and including sandpaper is a nice idea.

Since you're using plywood with five big pine plies in the middle sandwiched by two nicer birch plies (I think, right?) you will probably want your tabs to be 0.35 inches so that you get about three plies of tabs. Otherwise they'll just tend to break off. You need to sand the tabs off anyway, preferably with a small belt sander, so the difference between 0.2" tabs and 0.35" tabs is moot.

A few more judiciously placed screws will help with plywood warpage. One in the middle, and then maybe a quick grid would work. Screw from the middle outwards.

Hope tomorrow's attempt works out much better!

--Rob

[daniel_reetz]

Man! I spent the last two hours going to different Home Depots looking for the same plywood again. I HATE HOME DEPOT and I NEVER WANT TO SHOP THERE AGAIN. And for $50 a sheet, this is some pretty crappy plywood. Anyway I found two sheets and put another hundred bones into the Home Despot empire.

I'm using a downcutting bit right now, Rob, so your intuition is right. My bit may be dull and I'm going to order more ASAP.

I don't have a belt sander and still am magically thinking I can somehow avoid it. One thing I like about everything now is that with the proper RPMs and stuff, I'm no longer creating crazy fine dust - just chips. Since I live in my workshop, I'd like to keep that fine dust to an absolute minimum. We'll see.

I'd like to think of a better wood-clamping system in general, I think maybe some T-nuts into the bed of the router could help, and then I'd just drill a hole pattern into each sheet or something. For the first ten or twenty scanners it's not really gonna matter, but as time goes on having an efficient and repeatable clamping system is going to matter more and more.

Blah blah I'm tired so I'm quitting for the day. Tomorrow I cut TWO SCANNERS. Then I'm going to work on some kind of system to get plywood in and out of my place. It's really clumsy and dangerous right now.

[daniel_reetz]

A nice PDF on ways to move plywood.
A sheet handling tool.
A "lazy lifter" for sheet goods.
A discussion about sheet good handling methods.
A relatively cheap drywall dolly.
A hydraulic table cart.
And finally, probably the tool I really need, a panel handler from JET. With a cart like this I could unload from my truck onto it, and then push it into the freight elevator, and then rotate it flat for quick loading onto the CNC router. If I set the front and back of the router right, I could push goods on and push goods off.

[Fab52]

Hi Daniel,

First don't wait for the girl with hot pant in Gorilla handler ad to help you moved plywood ... after 30 years I never saw her around when I have plywood, drywall sheet or cast iron radiator to move around.

The drywall lifter will not work for you... the minimum height is to high for unloading from your pick-up and the base of this lifter is wider than most standard door.

You pick the right one but G'wiz the jet one is expensive... Few years ago I have 24 cast iron radiators (+/- 300 pounds each) to get out from a two story house to get it restore, we use dolly like this one made of square tubing and steel flat bar on the top.

See you

Fab

[daniel_reetz]

Hi Daniel,

First don't wait for the girl with hot pant in Gorilla handler ad to help you moved plywood ... after 30 years I never saw her around when I have plywood, drywall sheet or cast iron radiator to move around.
Fab

Ha!!! Yeah, I've never seen her either. Thanks for the good humor and wise advice, Fab.

So, today, I found a host of new issues with the original artwork. In particular, after cutting another scanner I found that some parts just don't fit! In all honesty, they're really not that bad, and Rob's kits have obviously been working VERY well, but the situation bothered me anyway. Basically, at some point, the entire artwork was scaled a few thousandths off. Here's a pic:

You can see the dimension box in the top left showing that the part is oddly sized (it should be 13.5x7.75, not whatever fractionally-too-small dimension it is). My guess is that this all happened at some early stage DXF conversion, perhaps when GaryK started his work on the project. In any case, it had to be corrected. So I spent much of the day chasing down problems in the artwork and correcting them. I also corrected most or all of the problems listed in my earlier video. Unfortunately for me, that also means reprogramming most/all of the toolpaths to cut another copy. It also means spending another $50 on plywood.

I'm going to stay up late tonight programming the toolpaths for this new revision. I'll try to cut a scanner tomorrow night after work.

[daniel_reetz]

Alright, well now I am REALLY REALLY ANNOYED.

I spent two straight nights correcting all the tiny errors that had crept up in these files. I re-programmed every single line and pocket. And then I went to cut these parts from the plywood I bought - ostensibly 18mm plywood. 18 millimeters = 0.708661417 inches. The cuts looked perfect! The only downside to my new cutting strategy, it seemed, is that it was taking longer than before because of some ramping/etc I was doing to be kinder to the tooling.

WELL After cutting out a whole scanner and finding out that NONE of the joints fit, I thought to measure the wood. Well, the wood is .74inches thick. No wonder I'm having fit issues. It's really unfortunate considering this wood is $50 a sheet. I can't keep throwing away scanners like this. I'm thinking I'm just going to have to spec-fit every joint by cutting them down.

AGGGGGGHHHHH

[daniel_reetz]

On the plus side, if you use "Raster" mode for pockets in VCarve, the pockets come out EXACT. But, the cost is time. Some of these pocketing operations cost 2x the time of others.

[daniel_reetz]

well, in the end i programmed a "finishing pass" for all mortised joints. basically removed .02" of material on all of them and things (at a first glance) seem like they might fit. This may be the "best" interim solution - it means it's not necessary to reprogram all the toolpaths for everything, just for a subset of parts.

so it looks like i've cut my first complete hackerspace scanner since the beta model. gonna give it a quick test-assemble tomorrow and then ship it off to Paper Upgrade for a for-real evaluation.

total cutting time 1:30 including finishing pass, at least 30 min too long.

good night all.

[rob]

Well, the wood is .74inches thick.

Whatwhatwhatwhatwha-a-a-a-at?! That is unbelievable! But also incredibly lucky that you found this out. Now you can mill down the joints like you said. The good thing is that all of our measurements are taken from the top of the bed, not the top of the material, which means that if you mill down to 0.72", you will end up with 0.72". You are setting z=0 to the bed, right?

You can definitely get more aggressive with the ramping. VCarve has options for spiral ramp (VERRRRY LONG) or, I forget what it's called, diagonal ramp? In any case you set the distance the ramp goes over. Try something like 2", since your depth pass is only 0.25" -- so you're going into the wood 1/32" for each 1/4" bit diameter travelled. But before you waste another sheet, try some practice squares to make sure it will work.

One other thing: set your material height to 0.75" or 0.8" and recalculate. This way you can accomodate anything from 0.72" to 0.8" and again, since your z=0 is the bed, you'll still end up with the right size.

Another bit of advice: set your pocket's tooling in VCarve to 80% stepover. That should save a little bit of time.

--Rob

[bonesgeorge]

Daniel, could you please share the new DXF files with the corrected sizes with us? I'd really appreciate it!

So, today, I found a host of new issues with the original artwork. In particular, after cutting another scanner I found that some parts just don't fit! In all honesty, they're really not that bad, and Rob's kits have obviously been working VERY well, but the situation bothered me anyway. Basically, at some point, the entire artwork was scaled a few thousandths off. Here's a pic:

dims_problem.jpg

You can see the dimension box in the top left showing that the part is oddly sized (it should be 13.5x7.75, not whatever fractionally-too-small dimension it is). My guess is that this all happened at some early stage DXF conversion, perhaps when GaryK started his work on the project. In any case, it had to be corrected. So I spent much of the day chasing down problems in the artwork and correcting them. I also corrected most or all of the problems listed in my earlier video. Unfortunately for me, that also means reprogramming most/all of the toolpaths to cut another copy. It also means spending another $50 on plywood.

I'm going to stay up late tonight programming the toolpaths for this new revision. I'll try to cut a scanner tomorrow night after work.