Semi-analogue CNC machinery

>>1128160
a thread
some words
a doodle
no explanation
essentially one of those
fuck you op

File: 20170208_164754.jpg (1MB, 2560x1920px) Image search: [Google]

1MB, 2560x1920px

>>1128163
If i wasn't clear enough:

There are gears A, B, C, and D.

Gear D is actually linear, so it will move up and down

Gear A is attached to a hand crank, as well as an electronic thing that record the amount that gear A rotated (idk what they're called)

Gear B is very large, is tangentially connected to gear A, and coaxially connected to gear C

Gear D is tangentially connected to gear C, and also harbors a drill bit.

By cranking the hand crank on gear A, you will gradually move the drill up and down. The degree of rotation of gear A will be recorded by the electronic piece, and can be used to calculate the lateral movement of the drill bit.

Hence, a CNC drill press that shouldn't cost thousands of dollars.

>>1128169
up OR down*

It's not cyclical, sorry

>>1128160
>Essentially a drill press with almost arbitrarily high tolerances for depth (just get a bigger gear)

You will be limited by the tolerances of the gears themselves. You'd be better off hacking up a cheap set of digital calipers and using those to detect depth. Actually, that's not a half-bad idea...might have to try that at some point.

And, honestly, I don't know why you'd want a drill press with depth accuracy greater than even a few thousandths. I can't think of a single application where you'd need an extremely accurate hole depth, but NOT an accurately located/bored hole (which is what you get with a drill press).

File: CX604_zoom.jpg (425KB, 1200x1200px) Image search: [Google]

425KB, 1200x1200px

>>1128160
See that handwheel at the front and the one on the side?

Can you guess what it does?

Also for the record you don't understand anything about backlash, tolerances or how machinery works.

>>1128226
>round column

do you have one of those? is it an ok mill?

>>1128226
I have no idea what backlash is, ill admit that

But i AM a civ eng major, so i know my way around tolerances. The idea is rough as is, but i could easily calculate the standard distribution of error with any given setup, assuming i know or can guess at the tolerances of its components.

>>1128222
An interesting thought actually. My plan was to use this idea to make all sorts of CNC bs, since a drill press obviously isn't the only thing you need.

You could apply this concept horizontally to measure location as well. You could even reuse the same system, lock the measurement in place, then reconfigure to do another thing.

>>1128281
>>1128222
Of course, if you lock the pieces in place anyway you could always just use a caliper to measure most distances, assuming the caliper is large enough. Then use the gear system just for consistency

I think i like that idea better, actually. And the cost would be about the same, with less time

>>1128281
>civ eng major

nigger, what the fuck are you doing? any engineering department worth its salt is going to have a machine shop. get your ass down there and start making shit.

holy fuck, its like you don't even know what your tuition is paying for.

File: dil_press.jpg (2MB, 2000x3008px) Image search: [Google]

2MB, 2000x3008px

with this 6mm 1mmpitch ballscrew, i can hold at least +/-0.001 newtonmeters
cast iron base, full cnc control, ER8 collets, 4 angular contact bearings
sub $1000

what does /diy/ think

>>1128281
You know maybe you should look up how existing cnc machines keep track of their position before coming up with NEW ideas.

I mean you wouldn't want to come on here and basically describe how people have been building NC systems since the beginning of time and then try to convince us that this was something NEW that you just came up with.

I mean that would be embarrassing.

>civ eng major
>tolerances
Does the resolution on your layout software even go below an 1/8th of an inch?

>>1128286
he said fucking civ eng, thats like roads and shit

could smell the engineering on him a kilometer away though, my personal guess was an electrical engineer


op, ditch the gears
gears are out, regular gears at least

File: sn-blackpowder.jpg (92KB, 800x600px) Image search: [Google]

92KB, 800x600px

>>1128286
Lol let's just say the things i want to make might not sit so well with a college campus

>inb4 read the sticky

>>1128317
you would probably have more support from the faculty than you think. actually use your university as an intellectual playground. be bold. get your fucking money's worth.

>>1128284
>Then use the gear system just for consistency

...but that's exactly what the gear system throws out.

Look, there's a big difference between resolution and accuracy. It's trivially easy to get an extremely high resolution, via gears, pulleys, signal interpolation, or what-have-you. But high resolution does NOT imply high-accuracy. Even if you have a system that can generate a million counts per inch, it's not particularly useful if it can't do that both REPEATABLY and PRECISELY. Any sort of mechanical separation (and I mean ANY kind) will reduce your accuracy. Period. A direct measurement will always be the most accurate. Gears introduce both lash, meaning there is play in the mechanism that can throw things off, and will have minute irregularities in their construction that will increase the margin of error on your readings. Toothed belts and pulleys work better, but still aren't perfect.

There's a reason commercially-made DROs always read directly from the encoder, and it's because it's the best way to do it if you need to be as precise and accurate as possible.

>>1128307

Ghetto/10.

>>1128337
When i said "for consistency", i mean moving the drill bit up and down. Yknow, so I don't try to hand drill things as part of a drill press?

The actual measurements would be found directly, via caliper, and locked in place with blocks before drilling

>>1128487
The only consistent part of the design is that the spindle will eventually move down when you tell it to and eventually move up when you tell it to. Big gears are entirely unfeasible and incredibly stupid to actual implementation. And gears have way too much backlash to be accurate. Having an encoder 2 gears away from the tool is idiotic.

And what's the point of making manual measurements if it's supposed to be CNC?