My CNC build-up with video and photos + everything a beginner should know

[dmoore]

First let me give a little back ground. I've looked at hobby CNC from the sidelines for some time - I first looked at the PlasmaCam machine back in the late 90's since I do mostly metal fabrication - of course the prices usually start (as they still do) at $10,000 and go up from there. I also had issues of space in my prior 600 sq/ft shop. I've since moved into a larger 1,300 sq/ft shop and now have the time and money to jump into CNC so I decided to wet my feet with a small MDF machine with the intention of moving up to a larger Plasma/Router unit such as the MechMate in the future.

For those still on the side lines that haven't built your machine, I'd like to share some of my experiences that might save you some time and frustration. First, start with a small CNC machine - it's not really the CNC machine itself which is the most complex and expensive part. If you purchase high quality parts (steppers, controllers, breakout boards and software), they can all be moved to another larger machine in the future - plus, you'll learn so much building and making mistakes (even with CNC machines build from plans) that the next CNC machine you build will be different anyway, so why build a larger CNC machine with the flaws the first time around?

To start with, think of CNC as a long chain of parts and systems - like a car (motor, transmission, drive train, suspension, electrical, etc). They include, but are not limited to: CNC machine construction material, power transmission type (belts, screw, rack, etc), limit and home switches, motors (stepper or servo), controller (Gecko, HobbyCNC, home built, etc), breakout board, interface type (USB, parallel port), computer, CNC controller software (Mach3, etc), CAM software (lots!) and finally at the top is the design software (AutoCAD, V-Carve, SketchUp, Solidworks, etc). So, as you can see, there is a long list of systems and technologies to learn and the vast majority are completely applicable to a future (and larger or more complex) CNC machine you built. So, start small, think about where you want to go in the future and build up from there.

While I'm sure that a lot of people build CNC machines just to build them, they are intended as a means to an end, a tool - what exactly are you trying to produce with the CNC machine? Do you build model steam engines? Make signage? Cabinet or custom wood working? 3D prototype (for molds, etc)? Car parts? Of course you want a versatile machine but for the same reason they make adjustable wrenches and non-adjustable wrenches - there is usually a tool that best fits the job for the level of quality you want to produce (for the same reason you don't rebuild an engine with an adjustable wrench.) So, decide what you want to make and that will define the materials you want to cut in it. It's very unlikely that an CNC machine made from MDF is going to mill ferrous metals - so instead, look at a mill conversion package. If you are going to work with aluminum, look at the machineability of the grade of metal and speeds you need to accomplish this in - maybe you can use a wood/aluminum based machine and "get away" with making it work but if you do any serious work, you'll need a mill conversion. If you plan to do simple signs, or basic parts, you can get away with an MDF/Wood/Aluminum based machine - just look at the sizes of materials you want to work with and go from there. Size is another issue - as I mentioned above, I'd start small and while I know most people starting out would like as large a machine as they can get, with each increase in size comes a whole host of issues - cost in the form of larger and/or longer lead screws, rails, support structures. Complexity in the need to use ever more accurate parts to overcome the cumulative errors in tolerance in parts as they become longer. Issues of space and storage - a 4'x8' machine won't generally fit into a 4'x8' spot in your garage. Larger and more complex materials as the machine grows in size - it's just not likely you'll be able to build a 4'x8' machine out of MDF, but MDF is great for a 2'x3' machine - so you'll need to move to metal and the additional complexity that entails. So, don't start with picking a machine, pick the project and let that dictate the machine type you will need.

There is a now a huge quantity of suppliers for the now fairly generic CNC parts market. You have plenty of choices with every level of software, power transmission, stepper controllers, breakout boards and PC's. You can get a fair amount of it used or assemble it yourself (controller and breakout boards). Of course there in lies the problem - there are so many choices, how do you make sure they all work together? We'll one option is to purchase a set of plans that details ALL the parts in the chain - which stepper controller, stepper motors, drive screws/belts, rails, etc. Of course that takes all the confusion out of it and is great when all these systems are new to begin with. Of course another option is to pick and choose your options but be prepared to work for it, not only on the front end of selection but also on the back end of making it all work together. Picking the best of breed that meets your needs will better suite you on future machines and on your current build - just spends LOTS of time reading the boards and asking questions and seeing what others have made work together. Either way, there are plenty of well designed plans out there and there are plenty of completely one-of-a-kind designs. Above all, do your homework - expect to spend weeks if not months doing research before you purchase a single item.

When entering CNC, the focus to start with is often on the mechanicals end of things. But, just like someone who falls in love with a 7 MPG truck and purchases it, the reality is that you have to put fuel into that truck each week even if that wasn't at the front of the decision making process to start with. So, starting with your focus on what projects you will be producing, plan out the entire CNC machine from hardware all the way through the software. You may be surprised to find that the $1,500 you spent building a CNC machine is cheaper than all the software (CAD/CAM/CNC Controller) you'll need to produce the results you want. Also look at the electronics that are between the software and the hardware also - a fair chunk of money can be spent on motors, stepper controllers, breakout boards and PC's with enough horse power to run your applications. Of course there are trade offs here also, you can spend less on build it yourself, full step controllers or go all out and get micro-stepping controllers - the question is, what level of quality do you need and what speed can you live with?

Another area to look at before jumping in is power transmission - lead screws (ACME? Threaded Rod?), lead screw nuts (basic nut, anti-backlash nut, ball screws?), gearing or direct drive?, stepper or servo?, pulley or chain? One or two motors for X axis?, rack and pinion? There are a lot of options and each one has cost, quality and speed advantages and disadvantages. I'd first ask what level of precision do you need? If you are making signs and speed isn't an issue, you could easily use basic threaded rod from Lowes/Home Depot for all of $4 and it will work fine. If you are making PC boards, you might just need something with more precision and of course costs. The downside (generally) is that with precision comes loss in speed - for the same reason you can better navigate tight areas on a bike as apposed to a car - but trying traveling 50 miles on a bike. Somewhere there is a sweet spot for your needs. I would keep in mind that you only need as must precision as you need to accomplish your goals and no more - it may be great that you can position your CNC machine with .00004 but if it takes 4 minutes to travel the 18 inches to get there, you may become a bit frustrated.

Another area to plan for is all the "extras". Just like when you buy a car, you don't "have" to have air conditioning but there is a point in the future that you'll think you made a bad decision for not getting it. Once you have yoour carefully planned out machine built and working, you'll discover that you are missing some things that were not completely obvious at first. Some of these "extras" are - a good clamping system. This could run from the complex such as a vacuum hold down system to the simple, with clamps to the machine base. Of course again we have many alternatives with advantages and disadvantages. Vacuum would be great, unless you have small parts, coolant requirements, no space to keep a source of vacuum generation, etc. Of course clamps can’t always reach where you want them, they can mar a surface, limit the area of machining and can damage your unit when the gantry goes crazy and hits the clamp. Another area is debris removal and this of course is dictated by the material you are removing - a simple vacuum for wood, a much more complex coolant system and container for metal chip removal. You can rest assured that a system for debris removal will move to the top of your list the first time you make a project and have to clean a layer of dust or chips from everything within the vicinity of the machine. Yet another "extra" is safety - for your machine (I'll hope you are operating it safely) in the form of limit and home switches. Yes, you can often set "soft limits" within software but one mistake or lack of attention and you can end up with a machine that will destroy itself in a second flat and require new lead screw nuts, pulley belts, motors or the machine itself. Without these safety items, I suspect it's not if but when an accident will occur. Yet another "extra" is the tooling that will be required - router bits, mills, etc, and these often are not cheap. Often a simple router bit will be 15-20$ each and you might need a dozen of them to do different type of milling or cutting. Plan to spend a few hundred on tooling alone - trying to use your left over drill bits will only get you so far. Of course you'll need the actual materials you will be working with, be that MDF, wood, plastic, steel, aluminum, etc. Of course every metal for the last 7 years has been expensive but even the basics such as foam and plastics have increased in price also. Even MDF can run $25-30$ per 4'x8' sheet. So, factor this in. Also factor in where you will be storing this material once you get started.

Another thing to consider is the placement of your system. A CNC machine with a 2hp router, doing a 2 hour job in your basement, right below the TV room where your wife is watching Desperate Housewives might not fly. You need enough space for the machine, space around the machine to place the materials, space to store your materials when not in use and space to house the computer to run the machine. Take into account if the machine will be near an area where sound and debris could be an issue, such as a basement or garage. You'll be surprised how a CNC router working on wood and without a vacuum system will manage to put dust in areas you didn't think it could go. Also keep in mind that a lot of materials and systems make a lot of noise in the process. My shop is 75 feet from the house - with the garage door open (that isn't even facing the house), my wife can hear the router going from inside the house. So, think of these issues before the neighbors show up with pitch-forks.

Another area is the tools required to build the machine. In my case I started to build mine with an old, less than accurate table saw. I later had to abandon the saw and purchase a $600, accurate table saw to get the level of precision I wanted in my build. Of course you'll need lots of other tools, some that can be a time saver and others that are a necessity. You don't have to have a drill press but making accurate holes with one is much easier than without one. Keep this in mind when building your machine. Of course there is starting to be a larger market for "bolt together" machines starting in the $500 range (such as www.buildyourcnc.com and with Joe 2006's ~$700 machine parts) that require only the most basic of tools.

Now, that all said, once you get everything working and you finally use your machine to cut out a project, you'll be amazed you built this thing by yourself with parts from eBay, the hardware store and your garage. Even the most basic machines can, with the proper software, produce astounding results. Once you have a working CNC machine, you all the sudden realize that there is a world of possibilities. Things that you got out the saber saw for, used the band saw or drilled by hand can now be done with a few clicks of the mouse to near perfection over processes performed by hand. Again, the key is the software and any education needed to learn these products.

All in all I am very happy with the outcome of my machine. As mentioned before, I plan to start with this machine and move up to units that better meet my needs in the future.

Here are the details behind my machine and what drove my decisions.

I built the basic structure of my machine from the "plans" and videos from Patrick's website, www.buildyourcnc.com. While he didn't provide drawings, he does provide good examples and he's done everything inside his house. So, clearly in a well stocked shop I felt I could build the same thing with ease. I decided to build a slightly larger machine than Patrick built - my machine ended up with a usable area of 39.7" for X, 28.8" for Y and 3.60" for Z. This is more than enough for the projects I was planning - signs, models and engraving. I decided to go the MDF route and it required two sheets of 4'x8', 3/4" MDF to build. Of course had I made less mistakes I could have fit it into a little more than a single sheet. On my first build I did the 1/4" "all thread" for lead screws mainly due to price. I later upgraded to 1/2"-10 TPI, two-start (a total of 5 turns per inch of travel) for my X and Y and 1/4"-16, single start for Z. The reason I upgraded was to gain speed. I also upgraded to "dumpster CNC" (www.dumpstercnc.com) anti-backlash nuts, couplers and lock nuts. I purchased my steppers on eBay used - just generic LIN Engineering units that have worked out well. I didn't purchase new, nor larger steppers since I knew in the future I would be buying new ones for another machine. For the supports and rails I used Patrick's method of using aluminum angle with skateboard bearings. These have worked out pretty well, especially after Patrick showed me a method to better secure the rails to my gantry (see photos).

I purchased four Gecko Drive stepper controllers (www.geckodrive.com) even tough I only needed three since they were on sale for $127 each and I knew I would use two steppers for the next CNC machine's X axis. I purchased the Gecko 203V's since they seem to generally be considered "top of the line" for hobby stepper controllers, though I could have gone as low as $100 for all three axis's with other less complex kits. While not technically required, I purchased the PMDX-122 breakout board from PMDX (www.pmdx.com) to provide easy hookup, buffered inputs and isolation. I am running the entire thing from a $150 HP PC I purchased from a local classified ad, 2.somthing GHz and it works fine. For CNC controller software I run Mach3 - an outstandingly complex program written by a hard core techie for hard core techies. The price is right at about $150 for everything you get. Next up is the big dog - the design software. I intended to do signs, models and parts and since I don't have any experience in CAD other than with SketchUp (which isn't "really" CAD), I needed easy to use software for design. The single leader in this space is Vetric (www.vectric.com), I purchased their Cut3D, Vcarve Pro and Photo Vcarve bundle. Of course I downloaded their trial applications which are fully functional and only limit you to the samples they provide in the trial, and I was amazed how easy it was to product great quality work. Of course the bundle was $950 USD… again, this is an example where the software alone is twice the entire cost of the hardware.

The "extras" included a cheap dust collection system from Harbor Freight (www.harborfreight.com), a self-built rolling metal table with storage and my computer is mounted in a rolling computer cabinet I picked up from a local hospital auction for $225. The rolling computer cabinet (nick named the "bat" cart) is great since it came pre-made with everything you need for a mobile CNC computer station - touch screen monitor, filtered power supply, wheels, cabinet doors to keep the dust down, keyboard stand, mounting locations for the PC, power supply and controllers. I also added wireless remote jog controller from Logitech called the RumblePad - it works great to jog to the work piece without any cords.

Here is rough breakdown of my costs (excluding tools):

Item Quanity Price Total
MDF - 4x8 sheets 2 $25.00 $50.00
Screws, bolts and nuts 1 $150.00 $150.00
1/2" Lead Screw - 6ft 2 $56.00 $112.00
1/4" Lead Screw - 3ft 1 $25.00 $25.00
Dumpster CNC Nuts, Couplers, etc 1 $100.00 $100.00
Bearings 6 $7.00 $42.00
eBay Stepper Motors 3 $40.00 $120.00
Breakout Board 1 $95.00 $95.00
GeckoDrive Stepper Controllers 3 $135.00 $405.00
Monitor Computer Cabinet (local auction) 1 $225.00 $225.00
Wiring and wiring management 1 $100.00 $100.00
Mach 3 Software 1 $150.00 $150.00
Rails, misc parts 1 $75.00 $75.00
Limit switches 6 $10.00 $60.00
Vetric "V-Carve" suite 1 $950.00 $950.00
Dust Collector 1 $100.00 $100.00
Dust Collection Hose and fittings 1 $45.00 $45.00
Steel for stand 1 $30.00 $30.00
Logitech Rumble Pad Pendant 1 $35.00 $35.00
Power Supply 1 $115.00 $115.00
Router 1 $25.00 $25.00
Personal Computer 1 $150.00 $150.00
Computer Cart 1 $250.00 $250.00
Router Bits 10 $15.00 $150.00
Misc (cables, dust shield, etc) 1 $75.00 $75.00
Totals $3,634.00

Here is a link to a video of the unit in operation on YouTube:

You Tube

You Tube

Thanks,
David Moore

[Darroll]

WOW, this should be a sticky!

Thanks David this has helped me tremendously.

[seanreit]

Not so fast dust breath! May I ask how you intend to hold things you are cutting on that table?

[dmoore]

Originally Posted by seanreit Not so fast dust breath! May I ask how you intend to hold things you are cutting on that table?

How observant you are! I have purchased a bunch of toggle clamps from Harbor Freight (http://www.harborfreight.com/cpi/cta...emnumber=96237 http://www.harborfreight.com/cpi/cta...emnumber=96234) and I will be installing T-Nuts into the base and then screwing in clamps where I need them, even in the middle of the bed (after I modify the base of the toggle clamps).

[dmoore]

Originally Posted by Darroll WOW, this should be a sticky! Thanks David this has helped me tremendously.

You are more than welcome. Sometimes I think we get so caught up in the hardware specs and details that we loose sight of the bigger picture.

[Gir]

Excellent writeup David, this was very useful. Your setup looks very professional, good job!

[jdell42]

Originally Posted by dmoore.com How observant you are! I have purchased a bunch of toggle clamps from Harbor Freight (http://www.harborfreight.com/cpi/cta...emnumber=96237 http://www.harborfreight.com/cpi/cta...emnumber=96234) and I will be installing T-Nuts into the base and then screwing in clamps where I need them, even in the middle of the bed (after I modify the base of the toggle clamps).

Let me know how those toggle clamps work out... and if you have problems with your dust collection shoe and the clamps not getting along so to speak.

[vcooney]

Hey dmoore,
Was there anything special you had to do to make the wirless controller work?

[dmoore]

Originally Posted by jdell42 Let me know how those toggle clamps work out... and if you have problems with your dust collection shoe and the clamps not getting along so to speak.

JDELL42 -

I've completed the clamping system and it turned out pretty good. I installed about 75 T-Nuts in the bed of my CNC machine. I took the toggle clamps and welded on some 3" long 1/4-20 bolts and tried it out. Overall it works pretty well. My suggestions would be:

- Use larger nuts... 3/8" or 5/16 since they should be a bit more stable
- Use a closer spacing. I used ~2" spacing and depending on your size clamp, it could be too short. You should use the distance between your bolt on the clamp to the end of your clamp, less a little bit for a clamping area on your material.
- The toggle clamps (such as those from Harbor Freight) come in vertical and horizontal. Get the horizontal since there will be less interference with the gantry - in my case the smaller clamps will fit under the gantry.
- When installing the T-Nuts, drill the hole just slightly smaller than the "neck" of the N-Nut - this should make it less likely for the T-Nut to come out.
- Use a adjustable drill or an impact wrench (1/4") with a bolt and a two or three washers on the top of the deck to "pull" the nut into the base of the deck.
- Use pliers to hold the T-Nut when pulling into the base of the unit, otherwise it may twist the "spikes" before it is fully seated.

Here are photos of the completed clamping system:

[dmoore]

After mounting my new Hitachi router, my old dust shoe wouldn't work. So I set out to make a new dust shoe. I wanted the basic shoe but I wanted it removal able to ease bit changes. So, I pulled some left over hard drives from servers in the data center and disassembled them to pull out the super-magnets. I screwed some washers to the bottom of the Z carriage. Then I used the CNC machine to make the holes for the router spindle nut and the dust removal hose. I screwed the magnets onto the base of the shoe and then attached some clear plastic around the outside. All in all it works well, though I think I'll change out the dust skirt to a different material. It's great to be able to remove the shoe to change the bits. Photos below:

[dmoore]

Originally Posted by vcooney Hey dmoore, Was there anything special you had to do to make the wirless controller work?

VCOONEY - I did have problems when I installed it on the orginal machine that was only 800mhz, the software that comes with the Logitech controller that captures the output and "presses the keys" - a problem other people have reported. I upgraded to a 2GHz machine and that solved the problem with the Mach3 slow down. The other minor problem is that the controller "goes to sleep" and you have to press a button before it will "wake up" since the joysticks will not wake it up. I'd say it's a toss up between wireless and wired. Still a pretty good deal for $29.

[dmoore]

Originally Posted by Gir Excellent writeup David, this was very useful. Your setup looks very professional, good job!

Thanks!