Technically this is my 2nd attempt at converting a Grizzly G0704 Milling Machine into a CNC Milling Machine – the first attempt fell through due to lack of funds. Sidenote: That is actually a big reason why I stress planning and budgeting so much when someone wants to try to make their own machine, it always costs more than you expect. Anyways it has been a long time coming for some sort of documentation on this project, and now that I am no longer living in an apartment I can finally start to finish this project.
So I have a few goals that I am trying to achieve with this conversion. First and most importantly, I want to be able to simultaneously machine with all 3 axes, and be able to make complex 3D shapes and organic curves like a sphere or something. Secondly I want the machine to be able to hold at least 0.003″ tolerance (assuming the part is setup and machined correctly). Finally since I am going to be running this machine inside a house, I need to limit the mess it creates. This means a full enclosure to contain metal chips and coolant.
A few of the lesser priority goals I have are the ability to do High Speed Machining toolpaths, a flood coolant system, an optional 4th and 5th axis, linear encoders and a dedicated computer/monitor for a controller. There are a few others but I already have a full plate with these goals so I am taking things one step at a time. I had originally decided that I was going scrape all of the dovetail ways on the machine, I even bought a surface plate and made some scrapers, but I realized I would rather just replace the dovetail ways with linear guides from Misumi. If done correctly I will still have enough rigidity and improved accuracy, as well as being able to square the X and Y axis up to each other easier. More on this in the future.
So there is a lot of stuff to cover for this conversion, and I am going to be putting plenty of pictures into the post so to help I will try to break things up into an outline format (if you couldn’t tell from the CNC Courses page, I love outlines). As I add content to each section I will link to that part of the page with the outline below so if a line is clickable, that means I have added content.
A. CAD Model – updated Aug.3 2015
1. Y Axis – updated Aug.3 2015
2. X Axis – updated Aug.3 2015
3. Z Axis – updated Aug.3 2015
B. Parts Needed to Purchase
II. Machining – Modifying Stock Parts and Fabricating Custom Parts
A. Changing the Stock Pieces
B. Custom Pieces
III. Assembly – Putting it All Together
IV. Wiring – All of the Electronics
V. Programming – Software and Stuff
As I said above, I have several goals for this project. Here I’m going to go a little more into detail on what I want to do. Completely stock, the dovetail ways of the G0704 mill are not very accurate, causing quite a few problems. The ways have a tendency to be too loose at some spots, and so tight that they bind in other spots, they aren’t very square to each axis, and have some geometrical errors as well (curvature, warp, twist, etc.). Ryan’s G0704 post on cnczone.com does a really good job of measuring and showing these errors. Most people usually lap the ways which at least gives a consistent fit which eliminates binding/looseness, it doesn’t correct for geometry errors. In Ryan’s G0704 post he scraped the ways, resulting in a much more accurate surface as well as eliminating the looseness issues. However, I am going to eliminate the dovetail ways entirely and instead replace them with linear guide rails from Misumi. I know that these rails will provide the rigidity and accuracy required, and should make some other modifications slightly easier. I will also replace all of the stock leadscrews with ballscrews, each having two spring loaded ballnuts to completely 100% eliminate backlash. My reasoning for this is if I eventually want to make a 4th and 5th axis and do simultaneous 5 axis machining, I need all backlash eliminated from the main 3 axes to help reduce the total backlash in the 4th and 5th axis. While backlash isn’t necessarily an issue as far as accuracy is concerned (backlash can be compensated for with software), it is an accuracy with rigidity and vibration – this is one of the main reasons people don’t try to machine with a 4th or 5th axis on a DIY CNC machine. I have a few ideas on how to deal with backlash in the 4th and 5th axis, but that will be covered in the future.
A. CAD Model
I had to model the G0704 in SolidWorks so I could plan out the mounts for the motors, ballscrews, and linear guides and stuff like that so I grabbed a pair of calipers and spent a few days finishing up the model. I used the head from cnczone.com member Onocyclone’s model – he had an AMA25LV which is another mill almost identical to the G0704. I checked the measurements of his model to my parts and they were slightly off, but I plan on replacing the head anyways so I just used his to save time. Otherwise the parts in the below picture are from my model (I’ll admit, I spent much more time than I’d like to admit getting the color of green correct…). After that is a picture of the modified model so far, with the next couple of sections showing exploded views and stuff.
Since I am going to be replacing the endcaps on the table and the handwheel mount on the Y axis and the cap on the Z axis I didn’t bother modelling those. I will be posting my CAD files eventually, if anyone wants them before let me know with the Contact page.
A.1 Y Axis
So the first thing I needed to do was plan out the linear guides. I plan on using the SE2BD16 linear rails from Misumi. For the Y axis I plan on adding a large steel spacer between the base and the Z axis column to get more Y travel, which makes the 390mm rail the perfect length. For one rail and two carts it is ~$130, and I need two of these sets for the Y axis. Below is a picture showing the planned rail setup. As it turns out, the rails are the perfect height for the dovetails. The dovetails are roughly 0.6″(15.24mm) and the linear rail + cart is exactly 16mm. As you’ll see later on, this isn’t really important anyways. This also shows the spacer on the Y axis. At this point I should also note that the spacer is not my idea, it belongs to HossMachine at www.g0704.com. His DVD/Download is well worth the money, he has a lot of great ideas and information. I will not post any information regarding his modifications – for example I won’t post the CAD file for the spacer. But seriously his DVD is well worth the money, and now if you purchase it you have the option of him mailing a DVD or getting a download link.
Now I know for a fact I’m going to get some people who criticize my decision to replace the dovetails with linear guides, especially for the intimidating cost of $900. A few of the things I’m going to hear are: “You’re not going to enough rigidity”, “Linear guides are only accurate if you set them up right, the ways would’ve been better”, “For the money you could’ve done (insert something else here)”, and just stuff like that. Here’s the thing though, while I’m not a stubborn person or opposed to anyone else’s ideas, the biggest reason I want to replace the dovetails with linear guides is just because I want to do it. There isn’t any harm in experimenting a little, so just hush. Now, proper dovetail ways should be more rigid than linear guides. But due to how poorly machined the ways are on a stock G0704, linear guides are actually just as rigid if not more rigid – or more specifically they are more consistent. Plus, each cart has a dynamic load rating of around 600lbs, and an allowable static moment of around 20 ft-lb (which when the static loads are considered, there will hardly be any torque applied to the carts). This is plenty of rigidity for machining. And as far as positioning the linear rails accurately, as long as the flat surfaces that the rails are bolted to are machined in one setup to ensure parallel surfaces, and the holes are drilled in the same setup, then it should be just fine. Anyways enough ranting, but this just seemed like something not many people have done and documented (besides a great post on cnczone where a guy did the Z axis), so I figured I’d give it a shot.
Moving on, I got started designing the bearing mounts, motor mounts, etc. I am using a double ballnut assembly, which I will spring load so eliminate backlash. Using some washer springs I can apply a range of force between 10lbs-50lbs. I’m not sure how much force I should apply yet, so I will be using some shims to adjust the force. Below is two pictures showing this assembly. The small parts between the ballnut and the mounting bracket in the first image are the springs.
The next few pictures are going to show the rest of the Y Axis, then I will explain some of the decision I made.
So I designed a simple bracket that will hold the Angular Contact bearings, and will be bolted to the base using the original two screw holes. There will also be three additional holes drilled & tapped in the base for extra holding strength. The pulleys have a 3:1 reduction which still leaves me with plenty of speed – in fact I’m still concerned about the torque I would get from that specific motor so I might try for a larger reduction or a different motor. I will be able to adjust the preload on the Angular Contact bearings using some shims between the outer races.
Here I had to make a decision to make my own saddle. Once I realized how much work I would have to do to the original saddle, it was pretty obvious that it would be well worth my money to make one from scratch. This gives the added benefit of having the choice to make it thicker, giving more rigidity, and now I don’t have to worry about the holes that were already drilled or the slots, etc. So in the above pictures, that is a piece of stock roughly 7″x7″x1″ 1018 steel. And again, I probably spent a bit too much time playing around with Solidworks and actually putting a screw in every hole…
So that’s it for the Y Axis, once I finish the rest of the design I will come back and clean the model up, make it look a little better and plan the rest of my bolt locations and fasteners. Then I will be posting my CAD drawings eventually.
A.2 X Axis
Alright, now for the X Axis. Hopefully I won’t have to type much here since everything is pretty much the same reasoning as the Y Axis. The rails are longer, in fact they are the same length as the Z axis rails – they cost about $160 for one rail and two blocks, so the estimated total for rails is around $900. Hopefully it is worth it! Also the ballnuts were arranged slightly different due to bolt locations, but the same spring loading concept remains the same. I still have to figure that out and model it, but that shouldn’t be too much work.
Free End Bearing of X Axis on G0704
So far I have racked up an impressive sum. Here is a list of parts I need to buy, and once I finish up my model I will add the price of stock. This won’t include any controller electronics or software, just motors and stuff.
SE2BD16-390 – $130 x 2 = $260
SE2BD16-630 – $160 x 4 = $640
CPM-SDSK-2310S Teknic Servo Motor – $257 x 3 = $771
Bearings – ~$100
Stock – ~$200