Re: [CAD_CAM_EDM_DRO] About to start building a CNC wood router

On Fri, 09 May 2003 22:51:05 -0500, you wrote:

>>
>>
>>>Especially since I'm
>>>homebuilding my own board. what can I say? I'm a nerd!
>>>
>>>
>>
>><grin>
>>
>
>Do I sense a little nerd solidarity forming here? :-) Would it help
>you to know that I still have a TS1000 with ZX81 motherboard? hehe

considering that I will probably do a g-code interpreter for the xyzr
table, possibly so... Multiple microprocessors to get the thing
running, mostly embedded types... design is still fluid.

>
>>>Anyway...like the subject line says, I am about to start building the
>>>machine itself. And, I'm trying to do it on the cheap, while
>>>maintaining some level of precision. Unlike metalworking setups, I
>>>don't require a level of precision measured out to 10 decimal
>>>places...the wood will move more than that with changes in humidity! A
>>>few thousandths of an inch would be nice, though. My intention is to
>>>build wooden (laminated material) clock parts, so reasonable precision
>>>is going to be a priority.
>>>
>>>
>>
>>And the size of this project is?
>>
>
>Sorry -- I'm still gathering surplus parts, and the final design won't
>be completely finished till I have all the parts. Since I'm salvaging
>things like motors from printers...my design has to be a little more
>flexible than I like.

OK, the design is going to be driven by the parts you can find and the
compromises you make. I asked because a 5 inch envelope machine is
quite different from a 5 foot envelope machine....

>
>Anyway, as to size, I've identified need, want, and acceptable
>compromise for my first attempt. Here're my thoughts, so far.
>
>Need = just over 4x4" table...my largest clock gear could be cut on a
>table this size.

Ok, doable, I think.

>
>Want = 49x97" table, the exact size of the largest sheet goods I can
>easily get in my area...MDF. It's 1" bigger in each dimension, than any
>other sheet good.

Yep, can you say "home despot?" I knew you could....

>
>Acceptable compromise = 12x18" table, which I think is managable for a
>first time effort. It gives me a large enough area to cut almost
>anything I actually need, and some room to grow, while not requiring
>creative solutions to flex over an 8' long active surface. Parts much
>larger than this, in the types of woodworking I actually do, do not
>usually require the precision available to me on a CNC machine. And,
>since I know of nowhere to scavenge steppers of a size that would make
>sense for a 4x8' table, the parts are, currently, unavailable to me.

Other than the mass of the table, with going slowly enough, the table
size should not be a problem. Since you wanted a gantry design, that
might be surprisingly mobile.

>
>Since this is a wood router, and my intention is to do primarily 2D
>parts, my Z axis travel will be quite limited. 2 or 3 inches, maximum.

No problem, but doable.

>
>Of course, as with most plans, changes will undoubtedly be made as I see
>creative solutions to my problems, and ways to integrate the scavenged
>parts into the design. However, this is as close as I've come to a
>final design. I've decided on a moving gantry design, which will span
>the 12" dimension of the table. I've also decided on screw drive for
>the motion...directly coupled to the stepper shafts.

Flexible couplings can be your friends here....

>For a small
>machine, this sounds managable...for something larger, the design will
>have to be refined, significantly. For example, whip on a long lead
>screw would be a killer...a rack would be a more appropriate solution.

Belt drives have been done.

>
>>>Here's my problem. I live out in a small country town...our primary
>>>sources of parts are WalMart, a small local hardware store, and some car
>>>parts stores. Thankfully I'm a nerd and had some old HP LASER printers
>>>that I could scavenge some steppers from -- 100 oz in. Not huge, but
>>>they should be sufficient for a woodworking lathe, and they're of very
>>>good quality.
>>>
>>>
>>
>>www.herbach.com for herbach and rademan, good quality steppers, but
>>low in-oz, little bigger than 100 oz/in, but not bigger than 200 right
>>now.
>>
>
>Well, I've already got the steppers I'm going to use. Like I said, they
>were scavenged from old Laser printers I had around -- 3 minebea
>astrosyn steppers, 6 pin. They're mismatched, but since I'm going to be
>building the boards anyway, this doesn't present a huge problem.
> Largest one is 100 oz-in...a little small, but it should suffice for a
>smallish proof-of-concept unit, and I'll build with an eye towards
>upgrading the motors later.

OK, unipolar they are.

>
>>look for electronics goldmine on the web, L298 drivers with heatsink
>>for 2 dollars each. Go look at the specs, you will probably want the
>>
>>
>
>I didn't think the L298 did step translation -- the Allegro 3967 does
>translation and drives small loads -- hook up to some power transistors
>of appropriate size, and I could drive anything I wanted. It does
>microstepping, too...and they give samples! :-)

The 298 is a driver, not a translator. 2 to 2.5 amps max, 40 to 50
volts max, needs external snubbing diodes. However, it is chopper
drive, which means that you can have the fast risetime to the motors
with current limiting and use a 30 volt or so supply. It will run
bipolar motors, so you could use just half of that winding on each
motor. Tried that with some 140 in/oz Fuji steppers and got better
performance running them that way, even off 15 volts.

>
>>them. For the board, if you can do printed circuit boards, you will
>>be *very* happy. www.mgchemicals.com has a good PC board kit, most of
>>this you can breadboard yourself. I'd recommend the most dense
>>negative you can get, period.
>>
>>
>
>Like I said, I took apart *some* of my LASER printers for motors...I
>have an HPLJ-4ML on this machine, and it prints beautifully. A little
>transfer paper, and I can easily etch. Of course for a proof-of-concept
>board I'll probably just wirewrap it...proof the circuit...and then etch
>(or mill?) some boards. Alternatively, I also still have an old pen
>plotter -- they do make an acid resist pen!

Right now, I prefer photo sensitized and then etching. I will give
the cnc boards a try, but have to build something to do that. Actual
production methods are still debatable, but I know what works best
(photo sensitized boards...).

>
>>also for some things, like the z axis, you may want to go to linear
>>bearings. www.mcmaster.com you can get 1/2 inch ceramic bearings
>>(need to be lubricated, but tolerate dust a bit better than linear
>>ball bearings) for under 10 dollars each.
>>
>
>My current intention is to use rollerblade bearings, some of which are
>built with a dust shield. I've salvaged some beautiful, if a bit short
>(just under 12") rails from the LASER printers which I thought would
>work very well indeed. HP designed good products...

Yes, however, if you can find linear bearing supported rails, they
would be absolutely best, IMHO. Linear bearings are in the 15-18
dollar range for ball bearings.

>
>Would these ceramic bearings be any better than a homebrew linear
>bearing? This is the basic design I'm intending to use...simple, and
>from the looks of it, quite robust.

Well, I can't say much about his metalworking skills, but it will
work. The skate bearings are 8 MM inside, typically, and if you got
some 8mm drill rod, it would simplify your life. Not too expensive.

The ceramic bearings are probably more accurate, and certainly, other
than the hole for the bearing assembly, a lot easier to do than the
approach I came up with. I plan to use the ceramic bearings and
ceramic rod for the z axis of the setup, along with a 1/2 inch acme
leadscrew and a thrust bearing. The disadvantage is that they are
more expensive, though.

consider a piece of channel, or the equivalent. take an angle and
mount two bearings on posts off that angle, so they would run on a
piece of square tubing running at a 45 degree diagonal. Surprisingly
good bearings for the effort. Bearings are inside the channel.

>http://www.rudijanke.de/Home/Projekte/CNC-Frase/CNC-Frase_Mechanik/hauptteil_cnc-frase_mechanik.html
>
>>>lead screw and the anti-backlash nut. The only threaded rod I can get
>>>in this town is allthread, which is 60 degree standard screw thread.
>>>
>>>
>>
>>Ick...
>>
>
>Tell me about it...but in its defense it is cheap, commonly available,
>and seems reasonably consistent. Being rolled, I would expect
>consistency between adjacent threads, but I may be wrong. It wouldn't
>be the first time.
>
>I'd still prefer ACME threads, and am pursuing this...hopefully I won't
>have to make the 100 mile trip to Dallas (TX) just for threaded rod, though.

Enco will UPS it to you... They might even have a sale on.... It's
surprisingly cheap, and probably less expensive than the allthread if
you want the 1/2 inch oil finish stuff....

Plus, it doesn't come pre-bent.

>
>>Acme Rod is surprisingly inexpensive. Enco has a sale on it, IIRC....
>>www.use-enco.com. Available in 3 and 6 foot lengths, but you're
>>likely to pay a bit more for shipping if you get the 6 foot length.
>>Nuts are more expensive than not, but the inexpensive plain variety
>>can be turned down. (Hint: turn it down on the shaft to make it
>>centered),
>>
>>
>>
>>>Not the best for transferring loads. Is this going to cause me trouble
>>>in the future? I mean, I know it'll work, but will it work well enough,
>>>and for a sufficiently long time? Should I just bite the bullet and
>>>mail order some acme threaded rod? Where can you get it? My background
>>>is not in machining, so I don't know the parts sources!
>>>
>>>
>>
>>You don't have a lathe? Access to one? Drill press I hope....
>>
>
>I have a very well equipped woodworking shop, with most power tools
>including more than a few special purpose ones. Almost any kind of saw
>you can think of, drill press, 36" bed lathe, etc... Tools aren't my
>problem...skill is my problem. Or, lack thereof. ;-) My interest in
>CNC is the level of precision that can be had...far better precision
>than I can manage by hand.

Well, the precision that you can put into it is the precision you can
get out of it....

>
>>>And, the anti-backlash nut. Obviously nothing like this is going to be
>>>
>>>
>>>Simply tapping "some" material should
>>>not provide greater precision than a storebought nut...or would it?
>>>
>>>
>>
>>Might, but an Acme thread is an expensive tap.
>>
>
>There's a local place, believe it or not, that carries all sorts of
>taps. They don't know what an ACME thread is, but they carry the taps!
> Not the rod, screw, nuts, etc..but yes to the taps! And, besides, if
> I was to use a plastic material, I could just use a short offcut of the
>threaded rod itself, as a tap. I've done it before!

yes you could. I understand that delrin makes a good plastic follower
nut, with low backlash, because it tends to shrink around the threads
a bit. It will wear moderately quickly, though.

>
>>>I appreciate any help or insight you can provide...
>>>
>>>
>>
>>Some of this is possibly OT for this list, so please contact me
>>directly if you need more information. All these places do ship by
>>UPS, and even in a major metropolitan area, you'd be surprised about
>>what you can't find, and how expensive it is otherwise.
>>
>
>I can't see how it'd be OT...it pertains, directly, to the design and
>construction of a CNC machine. Admittedly low end...but a CNC machine,
>which is right in line with this list's charter.

I'm somewhat new to the list, and am more used to a list that has a
little less stringent OT policy. I am still feeling my way around
what is and what is not permissible.

>
>If you want to contact me off list, though, just hit reply to sender.
>
> -- Chuck Knight

Probably not needful, though.

Harvey

>
>
>