[jasoneule]

Hello all-

Long time semi pro wood worker here. I have become increasingly interested in CNC tech after stumbling on it here 6 weeks ago.

I feel pretty informed now to be able to ask some questions after reading so much so here but I have some questions that are tailored to what I have in mind. Please feel free to suggest anything.

I want to make a router CNC that can cut 42" x 30". As of right now with what I do I do not foresee needing it bigger (famous last words....). I will be cutting wood (hardwoods, figured woods, occasional ply...never mdf) 95% of the time with some possible plastics thrown into the mix. I will be going with a big router, not a spindle. I have a full cabinet shop so I can/will continue to use that equipment for my main workflow. The main point of this machine to give me a competitive edge with the traditionally more labor intensive items....carvings, odd shapes, template making etc.

While I want to keep the price down..I learned over the years to buy the best you can afford. I have $1000 to spend.

With this $1000 I must be able to purchase the structure, liner motion, motors, drives etc. I already have plenty of routers and wood on hand. This $1000 does not include software or the computers used to run the thing...but does include the hardware package for the stepper motors (motors,drive, p/s etc)

What I would like to do is make a wooden framed CNC with some steel or aluminum plates in key motion spots. Working with wood is what I do so I feel confident I can get very good results. I would like to go with some 80/20 stuff but I feel my $1000 mark would be killed very quickly. I would rather put the money into the stuff that really makes a difference in workflow.

What I need help with it this:

1) I have decided Z axis will be lead screw. But for X and Y... I am torn on using RP or sticking with lead screw. What do you all think for basically a little smaller than a 4x3 machine? I have to keep it under a $1000.

2) I am still somewhat confused with what motor I need. Do really want to buy bigger motors than needed...but the are so close in price. I am thinking that this would be a good place to start

3 axis Kit with 380oz stepper with G540 drive from CNC ROUTER Parts

Do I need to go bigger...can I go less?

3) Lastly (at least for now)the motion. Skate bearings / Vbearings riding on angle iron seems very attractive considering price. Will I have long term success with this? Could this be possible with R/P... Would a linear motion carriage from "Fineline Automation" be an considerable up grade or more of the same in a different package?


Thanks

Jason
Wetland Wood Works

[jckstrthmghty]

For $1000 sure. That's easily obtainable. What the real question is accuracy. The bigger the cutting surface the more expensive it get exponentially while maintaining accuracy.

R&P would certainly add some price. Typically double the drive components for the larger x axis and the extra motor. I'm no expert but I don't think it would be necessary for 42" although it is a nice option for speed. For the price point you are after acme leadscrew is probably your best option.

G540 is a great option. I just purchased this myself recently. That kit should provide plenty of power for your configuration. You can certainly go less. The creator of joes2006 used a much cheaper kit from hobbycnc with 200 oz stepper and got good performance of 80 ipm using 1/2" 10 single start acme leadscrew.

I like Fineline Automation components myself. I plan to use much of their stuff in my next build but I've been trying to keep my current project in budget as well. I'd give it some thought to where you want to go with your project. Whether this machine will be the only machine or will you progress like many here have. You can then plan on what parts you will keep and minimize the expense on the next build.

So it comes down to what are your expectations. For $1000 there will certainly be some compromises but which ones are you willing to make? Speed, Accuracy, cutting area, and ease of assembly are some things for you to consider.

Btw I'm a total rookie and going through much of questions you are. I'm building a JGRO where the table itself cost me around ~$250 give or take and the rest will cost me ~$1000, not including the computer or software.

[ger21]

For a machine that size, you can go either r&p or 1/2-10 5 start acme. not sue which is cheaper, though, with the acme being about $100/6ft and the nuts about $25 each.

Depending on the design, you're probably going to want to drive the gantry from both sides, so you'll need 4 motors. The G540 4 axis kit at Keling is about $100 cheaper than cncrp, but doesn't include wires.

The G540 and 380 ox motors are definitely the way to go.

My machine uses skate bearings, and while adequate, I think the cncrp carriages would be a big improvement.

However, I see you budget as being much too low. You're going to spend your entire budget on electronics and drive components. And don't be surprised to spend $200 ore more on miscellaneous nuts and bolts and other small stuff.

[brtech]

I'd look towards the cncrouterparts carriages. Usually, they are mounted on 8020, but at least for your 4' axis, you can build a wooden frame as long as it's flat and stable. Perhaps a torsion box out of baltic birch ply. Mount CRS rails to that, and use the extended carriages.

You can try doing the same thing with the gantry, but you may be better off with a piece of 8020 for its better tradeoff of weight, strength and straightness. You can build the Z out of MDF.

At 4', I think double drive R&P really is the best option. The shorter axis can be screw.

Look at the finelineautomation 4 x 3 for ideas.

I don't think $1000 is enough, for the same reasons others do: the motors and mechanics eat too much of the cost right off the top, and you don't want to skimp there.

I guess if I had your skills and tools, I'd try torsion box X and Y, and MDF Z, with CRS rails and cncrouterparts carriages. Lots of MDF Z's around to copy from.

Look at how Nate (fineline) drops the bed of the router so that the X/Y interface is the Y gantry extrusion supported by the carriages. Contrast that with the cncrouterparts way of doing it with a riser. There are tradeoffs to both designs, but if you are doing it with wood, you might like Nate's approach better.

But can you build a torsion box gantry beam that doesn't sag more than .001 or so over 3' with the weight of the CRS and the Z zinging across it?

A hunk of 8020 does that pretty easily

[louieatienza]

It's a lot harder to do nowadays since everyone is like a hawk on eBay, but there are still deals to be had. Some junkyards/scrap metal places may have linear motion components you can use.

An inexpensive drive setup could consist of timing belts and pulleys. You could even run both sides of the gantry with one motor and a jack shaft. You also have the option direct driving the belts off the stepper (if powerful enough) or gearing them down, or having the stepper fixed or moving. I believe the MechMate plans have the option for belt drive. Just one thing to look at. And it wouldn't be to hard to retrofit r&p afterwards.

I think adding a few hundred to your budget would allow you to compromise less and build a better machine in less time, in the long haul. The problem is making a linear bearing system from scratch could end up costing more if you try to engineer it yourself, and spend a lot of time and money on something that won't work to your satisfaction. You can get supported linear shafts and all the bearing blocks from Glacern Machine Tools for $600 and it would be money well spent.

[ssutton]

My 2 cents...........take your time and build what you need now AND what you need later. I took 3 years to build because of not having the money now. But, I now have a machine that will meet my needs for years to come. Its cheaper to build it right the first time than build it two or three times wrong.

Just my 2 cents, no offense intended.

Scott

[Balcomwoodworks]

I too am a professional Cabinet maker. I believe that you should really consider waiting until you can afford to build a larger, stouter, 4x8 machine. you will find more than enough uses for it. Instead of cutting out templates you can be cutting out parts for instance.

Have you considered a chain drive until you can afford and retrofit a R and P ? There is a guy on the mechmate forum who did this successfully for a while (quite a few commercial jobs.) it seemed to be his opinion that it would have worked well on a machine with less use than his was seeing. I plan on going this route for my design/build.

[jasoneule]

Thanks all for your input. Its helping.

Let me clarify a few things though.

The primary intention of this machine is to cut wood, but not full sheet goods. If I am building cabinets or anything that requires a box for that matter without irregular shapes my cab saw will breakdown boards very quickly and accurately with little hassle, so a 4x8 machine really is to much (space and money)

I am interested in cutting odds and ends. Mostly curved parts....mostly hardwoods. Also carvings...ornate to simple. This is what kills me labor wise. I am in the transition process of backing off of cabinet work all together unless I really want to take the job. I have found that furniture and commissioned pieces to be much more lucrative and a whole lot more fun.

So sticking with something close to 4x3 seems about right and I certainly want to do it right.

The 2nd motor for the X axis is noted....and I will definitely go that route.

In my order of wants it goes like this:

Accuracy
Quality of Cut
Speed

So more than anything I just don't want to screw up a $200 piece of wood because a motor skips a step or it burns it.

With that being said I would like to dive into more discussion on R/P vs leadscrew.

From what I have read recently it seems that mostly everyone used leadscrews...and half of them complain about backlash or whipping or something else. Secondly, as Ger21 mentioned...lead screw is not cheap once you add up everything.

I am trying to wrap my head around the R/P because it seems to be what everyone "wants" to do but eventually defaults to lead screw. It also seems that using some prefabbed stuff from Fineline that it is going to be around the same price as lead screw...am I wrong on this? Besides its supperior jog speed....does R/P exibit less backlash problems known to lead screws?

I looked into the Glacern stuff. Its nice...very nice. I just do not know if spending that coin would get me any more precision than a fine line carriage on cold rolled steel. I mean I know it probably does...but could I even notice in my work?

As for the Y beam and it sagging...I was going to have some CRS spaning it for the carriages to ride on. this would more than likely make it completely sag free. However I will certainly consider the 8020 for the Y beam. Is there any one in particular size W x H that can be recommended.


Thanks all. I will be documenting the build once I start and recontribute.

Also as a side note. Since most of you run dust collection (I hope)...you may find what I did here interesting:

Video
[ame=http://www.youtube.com/watch?v=dhnBvdeaWS8&feature=related]Dust Separator - Thien Concept 5 Final Test - YouTube[/ame]

Writeup
6" Inlet/Outlet Separator How to with Video, Pics and Text

Jason

[louieatienza]

I wouldn't go too crazy with accuracy, especially in a wood router. Solid woods especially can grow and shrink larger than the accuracy of your machine! Having a ridgid machine with quality linear bearings will go a long way in terms of cut quality, as well as having as much a backlash free drive, regardless of choice. Basklash-free also means repeatability, which is way more important than accuracy to me - the ability to repeat a position or path as precisely as possible. You can use microstepping to "gain" back some resolution.

As far as the Glacern rails, consider that the cost of the CNCRP trucks will run you over $300, and that doesn't include all the cold rolled steel and drilling. There are advantages and disadvantages to both, and it helps to look at all available options. One can be a better fit over the other for a particular design.

[RossMosh]

I'd look at the CNCRouterParts linear guide system and Joes CNC rail system. You're going to want to use one or the other for your cnc machine. After that, you need to add up the cost to build a proper frame. You're also going to have to decide if you want to use 8020 or go with a more raw material.

In the end, $1000 seems light. I think you should look at doubling the budget at minimum. Really, it's more realistic to think you're going to spend between $3-4000. Remember, software alone you're probably looking at $600 (V Carve Pro) and $175 (Mach 3). Add in router bits, dust management, the router itself, cable management, and the other bits and pieces and you'll realize pretty quickly $3000 comes pretty quickly.

Also the big benefit of R&P vs leadscrew is simple: R&P is a no worry system. With leadscrews you need to make sure you pick the right screw otherwise your machine will be too slow. It might cause whip. It might not run well with your motors. R&P just works.

[KriegKuts]

The FLA-100 from fineline cost me about 2k, with a cheap 2hp harbor freight dust collector. Note, that I chose to cut, drill and tap myself, which saves around 400+ off the bat. This is with a g540 and hitachi spindle and a bit or 2.

[racedirector]

You may also consider something like a Joes 4x4 which can be built in pretty much any configuration you can think of. I plan to one day build a 2x4 version of it, others have grown it to 8x4. The plans themselves are $100 but well worth it to see what makes up the machine and the support forum for it has a wealth of information on building one in various ways. If you are infact interested, take a look at the Joes2006 forum or even visit Joes site (joescnc.com) and email him about your plans and his.....

Cheers

[racedirector]

Nobody seems to have mentioned HobbyCNC equipment. There are heaps of people who use this equipment to power wood routers, even 4 axis versions. They have driver boards and motor kits although the boards are kitset. There are, however, numerous build threads on them around the web. I went another way and bought SLAmstepper boards, again kitset, will be buying motors from automationtechnologies, power supply from somewhere else etc etc as I too am on a tight budget.

Just thought I would throw some other options out there....

Cheers

[vtx1029]

I'd say lead screw is going to be cheaper, but probably not by a ton. If you use CRP's carriages it would be easy to go R&P its just bolt on. If you go with something different it may be a wash if you have to buy or build a bunch of pieces to go lead screw. I just bought my R&P from More gear and it was about $50 shipped per 6' piece add in $90+ shipping per R&P Drive from CRP so your looking at around $150 for each axis.

I agree with others. Your going to be hard pressed to build something for under a grand. 2500-4000 is a more realistic area. The little pieces nickel and dime you to death!

I was going to build a frame and base out of LVL but after looking at it going all 8020 was not that much more $ wise for me being I was able to find some deals on used extrusion. But LVL or other wood bases can be appealing and they do have more mass then a 8020 built machine.

[brtech]

The general rule of thumb around here is 3' and smaller use screw, 4' and larger use R&P. This is not hard and fast, but it's common. So your X probably should be R&P and your Y could be screw.

Then you need to decide on the linear motion. The cncrouterparts carriages are a popular option, they ride on CRS rails. Other options include V-Groove bearings running on the edge of angle iron (see for example the buildyourcnc machines), and skate board bearings mounted on angle aluminum running on pipe (see for example, the solsylva machines). Someone mentioned chain drive, the buildyourcnc Blacktoe is an example. The other choice is a supported linear bearing like the Glacern stuff, but there are many other vendors for that kind of thing. You want supported, not unsupported.

Then you decide on construction. You can choose Steel, aluminum extrusion or MDF (or other wooden sheet goods). While 8020 extrusions are a popular choice, lots of successful MDF machines have been built. You could do a hybrid (wood supported CRS on X, 8020 supported CRS on Y, MDF Z). The Blacktoe is a 4 x 2 MDO machine.

Lots and lots of design choices. Look, for example, at how Nate (fineline) mounts his Y gantry beam right on top of the carriages, and lowers the baseboard to give the Z travel (look at the FLA-300, a 4 x 4' machine). Compare that to how Ahren (cncrouterparts) raises the gantry beam with steel risers (CRP4848). Note that when you buy from fineline, you are getting cncrouterparts drives. Nate cuts, drills and taps the 8020, drills the CRS and the rails, and puts all the fasteners and other parts together. You can download the plans for free, buy the carriages, R&P drives, electronics and other parts from cncrouterparts and DIY the same thing. Ahren is roughly doing the same deal with the CRP4848.

Accuracy is measured in thousandths. An 8020 machine with screw or R&P will get within a few thousandths of where it should be over 4'. An MDF machine with chain drive probably can't.

Gantry beam sag calculation is not trivial. You need to know a lot about the materials. 8020 Inc has a deflection calculator that you can use to estimate what their extrusions do when you hang the CRS and Z carriage on one of their beams, but it probably takes a Mech E with some tools to figure out what an MDF or BB ply torsion box beam might do.

[ger21]

Accuracy is measured in thousandths. An 8020 machine with screw or R&P will get within a few thousandths of where it should be over 4'. An MDF machine with chain drive probably can't.

if you throw out the chain drive, an MDF machine can and should be every bit as precise as an 8020 machine. Possibly more, as aluminum can grow quite a bit with temperature changes.

[louieatienza]

if you throw out the chain drive, an MDF machine can and should be every bit as precise as an 8020 machine. Possibly more, as aluminum can grow quite a bit with temperature changes.

Unless the linear bearing system and drive system are also made of aluminum, I don't see how having the aluminum "grow" affect the resolution of a r&p, or leadscrew, given that these are usually bolted onto the t-tracks?

[adprinter]

I have to agree with others on this post, $1000 is going to get burned up just in the electronics (Motors, Power Supply, Drive, Energy Chains, Limit Switches, Wiring, Plug adaptors, limiting resistors, etc.). You mention your familiarity with wood working, but not with machining. Do you have access to a machine shop, or at the very least, metal working tools such as a lathe, band saw, drill press, taps and dies, etc.? I ask this question, because if you have to hire for the metal work, your budget is gone already! I cannot say with absolute certainty, but my estimate to date in the construction/redesign/reconstruction and improvements made to date on my machine is probably in the neighborhood of $2,000 invested. I designed my own linear rail system, which has proven to be a very acurate system, driving a gantry which probably weighs over 300 lbs. All bearings used on the X and Y axis are skate wheel bearings. Cheap, and easily replaced. On the Z axis, DO NOT SKIMP! For this is the heart of any CNC system, in terms of acuracy. I used traditional linear rails with recirculating ball bearings on the slide, and a ball screw on the drive mechanism. Unfortunately, the ballscrew purchase was a used ebay item, which had to be replaced. (I went with acme lead screw, since unemployed, I just could not afford a replacement ball screw assembly). So, now all three axis are using 1/2" 10 TPI acme lead screws. The X axis makes use of my custom designed linear rails, which is basically .25" thick 1.5" square tubes acting as the rails, with 1/8" thick 4" square tubes as the trucks with skate wheel bearings mounted onto swing bolts. This provides an adjustable truck assembly, which grips all four sides of the 1.5" rails. The lead screw is mounted down the center of the table, and used a single motor to drive the X axis of the gantry. Since I am using a Gecko G540, and 900 oz in motors, this allowed the 4th axis to be added later (1200 oz in motor on that one). The table is constructed of steel angles, square tubes, and 3/4" thick aluminum slats for the table surface. This provides table-length slots, which allow the use of 1/2" bolts for hold down clamps. I use it for days at a time, cutting Oak with a 1.5 HP Router. There are no welds on this machine, which allows for adjustability to true everything up. An important feature of any machine. Fine tune adjustments would not be possible on a frame which has been welded together! (And you will find, that the ability to make adjustments are absolutely essential to achieve precision). Once you build one, you will find that it will be a life-long task of constant improvement, and refinement. The one thing I would like to suggest for you, going in, is to design your machine to be adjustable! If the frame assembly is designed in this manner, with mechanisms assembled for this purpose in mind, it is a simple matter of "dialing it in" once the machine has been assembled. It is also useful, for tightening things up, as parts begin to wear. You did mention "Acuracy" as the top of your priority list. With "Speed" being last on the list. The machine I built I generally run at 50 IPM on roughing cuts in solid Oak, and 20 IPM on finishing cuts using carbide bits. Hope this helps to convince you not to skimp on the electronics- your budget will go almost totally for just the electronics (IF you expect Acuracy). As for the mechanical structure, it depends on the materials you already have on hand, your access to the tools needed for metal working, and your abilities to work with metal which will largely determine the budget needed for the construction.

[ssutton]

I have an MDF sacrificial surface on my machine and it "grows" and "shrinks" a lot with humidity. IMHO aluminum would be infinitely more stable than MDF as a construction material.

[LeeWay]

I live in South Alabama abut 20 minutes from the Gulf. While it's not as apparent when I run my AC and keep my doors closed, humidity really gets into MDF swelling every edge and screw hole. Anything built with MDF has to be made fast and sealed and finished ASAP. I have a personal ban on anything MDF in my shop because of this.