The base and stand of the machine are primarily made up of T slot extrusion and plates that connect to the extrusion. Building the base consisted of cutting a ton of extrusion and bar down to length, countless holes getting tapped, a little bit of bandsaw/drillpress work, and installing a bunch of fasteners. I’ve already found a few things I would do differently if I had the chance to start over on this build, but so far no major roadblocks have appeared.
Plates
This project has a ton of 2D 3/8″ plates, most of which are used in the base and stand. I started the build by knocking out all of the plates for the whole project at once. I decided to laser cut some posterboard templates to help make the layout way easier. I was fortunate that the laser had enough power fidelity to merely engrave center-marks and labels before fully cutting the outline.
Once the layout was done, I was able to make quick work of the plates on the miter saw. A few of the small corners and rounded edges had to get finished on the bandsaw afterwards.
Most of the holes in the plates are M8 clearance, so I just used a punch to mark the hole location and drilled them out on the drill press. There were a few threaded holes as well (motor mounts, bearing mounts, sensor mounts) and while the drill press would have probably been fine for those too, I took the opportunity to clamp them in the mill and really dial in the hole positions. All of the threaded features were hand tapped afterwords.
I have access to a sandblaster through a makerspace I joined this year, so I thought it would be nice to blast the outside of all of the plates to give them a very consistent appearance. This is completely unnecessary, but I thought it was fun and looked nice.
In total, the plates only took a few hours to make. The commonality of drill sizes (mostly M8 clearance and M6 threaded), and assembly-line processing made them a quick build. They were made from 10 feet of 3/8″ x 4″ 6061-T6 bar costing just under $100, which is much cheaper than the $425 quote I got from send-cut-send.
The only thing I would change about this process for next time would be to include hole sizing information on the templates. Even though I do have drawings for all of these parts the templates were a much more useful reference in the shop.
Base
The first step in building the base is to process all of the T slot extrusion. Some of my large pieces I had cut to length by the supplier, but for the shorter pieces I had to mark the lengths and cut them down on the miter saw myself. After that, every piece was just some combination of tapping the end of the extrusion M8 and drilling clearance holes in the slots for M8 bolts.
For the longer segments of matching extrusion with clearance holes for connecting to the ends of a cross piece (IM, OM, RM parts) I made sure to clamp the pairs of extrusion together before drilling clearance holes to get a matched-fit. Some of the holes I just punch-marked and drilled out on the drill press, but for others it was easier to clamp them in the mill and dial in the exact positions without having to balance a large heavy beam on my tiny drill press table!
The first step in assembly was to join the four segments that make up the perimeter of the table. I assembled them with the table rightside-up to keep the bottoms of the extrusion flush, while also keeping the corners all flush. Luckily my two cross members were identical in length so I didn’t need to shim anything.
Next, I flipped the table upside-down and added the two large cross pieces. This kept the ‘top’ face of the four main cross pieces flush. Again, the lengths of the extrusion (which were cut by the supplier) were spot-on so I didn’t need to shim anything.
With the table still upside-down I added all of the blocking members to keep their ‘top’ surface flush with the rest. Because I cut these myself, I ended up shimming between 0.5mm and 1.5mm of additional length onto each blocking member to ensure that the distance between cross members remained constant.
Finally, to square the table I added corner brackets, turnbuckles, and steel cables. The corner brackets had an extra little alignment nub that I had to grind off in order for them to mount flush. My diagonals were off by about 5mm, but tensioning one cable slightly more than the other brought the whole thing into square! I’m extremely happy with my decision to use the diagonal cables and I honestly don’t know how I would be able to adjust the squareness without them.
One big takeaway I had from the base build was that it is way easier to cut extrusion to length on a miter saw than I anticipated. I was getting +/-0.5mm on all of my cuts, but I planned a cut dimension 1mm under nominal to account for a larger error. The result is that I ended up shimming all of my undersized blocking members to full length instead of just trying to nail the length on each cut and minimize shimming.
Another issue (that I really should have seen coming) was that I was unable to put nearly the amount of torque on all of my through-bolts as I would like. I intentionally threaded the ends of the extrusion 2-3 diameters deep so I could put a ton of tension on the through-bolts that hold the whole frame together. However, the tiny hex feature on button head bolts (which I selected to minimize their ability to snag on things) can’t handle nearly the amount of torque I would like to put on the fastener.
If I could do it over again, I would use flanged hex-head cap screws in order to increase the amount of torque I can apply. Luckily, I learned this lesson before ordering fasteners for other parts of the build so I can change my design for the gantry, carriage, and spindle to incorporate more substantial fastener heads.
Stand
The legs were very easy to add on to the base. Again, I had the extrusion cut to length by the supplier, so they matched each other very well. I had to thread 20 more M8 holes in order to install the foot plates and mount the legs. The foot plates also required the installation of a steel threaded insert so the swivel foot can adjust without destroying the threads.
I flipped the base rightside-up, keeping it elevated on a workbench so I could install the legs below. It took a ton of fasteners, but I was able to get all four legs and seven gusset plates (the location in the pattern where the electrical box goes does not get a gusset plate) installed and flushed. I’m not too worried about the legs being square to the table as long as they are securely mounted.
Rail and Screw
Hiwin rails are actually quite flexible and rely on external features to maintain their straightness. Because I didn’t machine a lip into the extrusion that I am using as a mounting surface for the rails I had to dial in each rail to the straightest reference feature I could find.
I decided to use a long carpentry level laid across all four cross members of the base as a reference, then mounted a dial indicator to a bearing on the rail. By running the dial indicator along the rail and measuring the face of the level I could determine the relative position of the rail to the table.
After about 30 minutes of adjusting, tightening, loosening, and re-tightening I was able to get the first rail dialed in to within 0.001″ of the level. I repeated this on the second side, but with the additional constraint of making the height of the rail relative to the bed (measured with calibers) match. There are some obvious flaws with this approach, but without a large reference flat surface, CMM, or inspection arm I don’t know how else to make the rails flat and co-planar.
The motors and ball screws were easy to mount on their respective plates. I will be adjusting the position of the ball screw plates later after I add the gantry plates because I can use the ballnut on the moving assembly to help me dial in the screw position.
I don’t have anything specific that I would do differently during this step on the build other than maybe search for better ways to align the rails. Once I add the gantry we will see how much shimming and adjustment I will need to do to the ball screw mounts.
Moved And Leveled
The entire assembly weighs well over 100 pounds at this point, so with some help I was able to use the work bench as a dolly to wheel the router over to its final spot in the shop and place it on the floor. I used my giant carpentry level again to adjust the swivel feet and get the table mounting surface as level as I could. I double checked squareness and it was still ‘perfect’ (within my limited ability to measure it).
There are a few pieces still missing that will be added later:
- E-Chain and mount that I will add after the electronics box is mounted
- Y Axis forward hard stop plates that I will add after the gantry is installed
- Y Axis rail guards that I will add after the gantry is installed
- Y Axis homing switches and limit switch flags that I will add with other electronics once they arrive
- Steel ballast/shear plate on the bottom that I want to add after the machine is ‘running’ so I can do a before/after comparison
- Leg stringers that I will add after the machine is ‘running’ so I can do a before/after comparison
- Hard stop bumpers that I will add all at once so I can prep all of the surfaces for adhesive at the same time
Time and Cost
I’ve been tracking my time based on the feedback from someone during the reddit review. So far, I’ve spent 14 hours on the base and 4 hours on the stand. The base was about equal parts plate prepping, extrusion prepping, and assembly/calibration. The stand was mostly plate prepping and extrusion prepping.
The biggest time sink that I did not anticipate was how long it took to thread all of the M8 holes in the extrusion. There are 20 threads in the stand and 58 in the base! Because I want to maximize the preload on my M8 steel fasteners most of the holes were threaded 20mm deep, which often required two separate passes with the tap to clear out the chips. Some extrusion suppliers, like misumi, offer options to pre-thread the ends of the extrusion. I would definitely consider this option in the future because it is probably the most cost-effective way to reduce build time in the stand and base assemblies.
The extrusion cost really added up quickly. The base extrusion cost $614 while the stand extrusion cost $228. All of the plates were under $100 combined, and shared stock with plates from a few other assemblies. About half of the material was used just for the stand.
The fasteners also added up quickly. The long M8 bolts (52 in total!) to connect together the frame cost just over $1 each. Each pair of 20-25mm M8 and T slot nut cost almost $1, and there are 20 installed on the base now and another 14 coming with the rail covers. The stand added another 65 pairs of M8 t slot nuts and screws. These common sizes were all bought on amazon from no-name suppliers to save on cost, although maybe 1 in 50 fasteners had a bad thread that I had to fix with a tap or die.
The tension cable and related base hardware cost $130 from mcmaster. The swivel feet and threaded inserts for the stand cost $40. Overall this isn’t too steep of a price to pay for a leveled and squared machine. In the future I will need to buy some kind of bracket to help me bolt the router to the floor.
The rails and bearings cost $113 for the pair and the screws cost $128 each. The screws included the bearings and motor couplings I needed, which is nice. I’m going to account for the cost of the motors in a future controls post.
The grand total as shown for the base is about $1200, while the stand is about $375. These numbers don’t make for a great comparison yet because the stand still has some stringers to add ($125) and the base still needs its steel shear panel and hardware ($200). It looks like my stand is going to be only about $100 (17%) cheaper than the avid leg kit, although it is probably slightly stiffer. I’ll need to wait and see how the gantry, carriage, and spindle shake out to compare my design to the main cost of an Avid.