Heavy Tabletop CNC Milling Mc 16" x 9" x 8"

[RET]

Hi,

I had been interested in CNC machining for a long time so back around 2005, I decided to purchase a "turnkey" CNC mill from Sherline. It came with everything, 3 axis mill, stepper motors, computer and software. The only thing I had to supply was a monitor for the computer.

This is what I started with; the little CNC mill from Sherline.

The software uses EMC code which can be downloaded from the internet for free like Linux. Now there is EMC2 which accommodates subroutines and one or two other enhancements from the original EMC. The Sherline computer incorporates the stepper motor driver board (4 axes, X Y Z and A) inside the computer case for a relatively compact, integrated system. The CNC expert at Sherline wrote a "front end" for the EMC mill software which makes things much more "user friendly" and is well worth having.

I used the system for a number of years to make parts mainly for "Dart," a 7 1/4" 0-4-2 English tank locomotive. These included two expansion links from oil hardening tool steel and two coupling rods from mild steel. Sherline also supplied a "textbook" on how to write "G" code which was very helpful. I learned a lot from the textbook and from actually using the software and making parts. Its actually amazing that you can make very complex parts from just a few commands for straight lines and curves (G0, G1, G2 and G3).

While versatile, I saw the limitations of the Sherline mill, so in 2011, I decided to make a much heavier, stronger and bigger unit but still using the same stepper motors and computer. This way, the computer wouldn't know the difference. I figured I could do this because at least half of the torque from the steppers was wasted in overcoming the friction of the dovetail slides used on the 3 axes. To eliminate play and backlash, the slides had to be tight and the 1/4-20 lead screws also had a double nut backlash elimination system which added even more friction. This was necessary for accuracy and to allow climb milling. By using Thomson ball bushings and ball screws, all of that friction would be eliminated.

This is what I wound up with (before I changed the spindle drive motor)

This will do for a start. I hope enough people will be interested in what I have done. Until I actually checked back, I didn't realize how long ago it was that I did this. I have been using the machine ever since and while it was expensive to build, it is an indispensable addition to my shop.

Richard Trounce.

[RET]

Hi,

Now I'll start to post the construction pictures, but before I do I'll give the key design considerations that are necessary for success.

First, you are designing and building a machine tool where the important thing is to keep deflection to a minimum. Second (for CNC especially), there has to be absolutely no play or slop of ANY kind anywhere because with CNC, you climb mill wherever possible because the finish is better and the cutter stays sharp longer.

This is the start of construction, the base of the "Y" axis.

The Thomson shafts are 1" dia. with support rails and the aluminum plate is 3/4" thick. The support rails are bolted to the Thomson shafts through 1/4-20 tapped holes on the underside of the shaft. Since the shafts are hard on the surface, you tell the supplier how long you want them and where you want the tapped holes (which have to match with the predrilled holes in the support rails).

I used the Sony readout on the Bridgeport mill to locate and drill all the holes in the aluminum plate. This way, they are all accurately located to less than .001" and on assembly, everything just bolts together with your fingers. As you can see, I used socket head cap screws everywhere because they are almost impossible to break.

Another view of the same stage of the assembly.

The original Sherline mill used 1/4-20 lead screws for the axes and the Thomson machine tool quality ball screws I chose were 5/8" dia. by 5 threads per inch, so I used a 4 to 1 reduction with timing belts on each axis. The timing belts and pulleys are relatively cheap, they are strong and they need no lubrication. They will last for a long time. This reduction means that the steppers don't have to be modified at all, they just need to be mounted. Because of the reverse mounting that you see in the pictures, the "Y" axis had to have the direction of rotation changed in the control parameters.

In this view, the "Y" axis table is mounted and the ball nut is fastened to it.

It is very important to keep dirt of any kind out of the system. For this reason, the ball bushings have lip seals and the ball screw nuts have dust wipers (they don't come with lip seals).

There's quite a bit more, but this should do it for now.

Richard Trounce.

[RET]

Hi,

Here's the next installment. The first view continues from the previous installment. If you look closely, you can see that the ball nut isn't fastened to the intermediate table yet so that part can still roll easily if you aren't careful.

Another view showing the intermediate "y" table mounted.

Again, all the holes are located and even tapped on the Bridgeport so on assembly, the alignment is perfect. I don't even do any layout scribing on the parts, its all done by using the co-ordinates on the readout. While the readout can go down to one micron, the mill can't, so I have the readout set to two tenths of a thou. on all 3 axes.

In this view, the upside down pieces are the Thomson shafts for the mill table.

As you may have noticed, for the initial building up I have everything sitting on a wooden stool. In spite of the lip seals, because of the low friction of the Thomson shafts, if anything tilts at all, suddenly things move. That can be awkward at times if you aren't expecting it, but it also showed me how much less power it would take to move the different axes.

Where I could, I used material that I had, but the middle 3/4" plate I had to buy and also the 3/4" steel plate for the table. The most expensive parts by far were the ball screws and the Thomson shafts. By the time I got them here, each axis cost me $2,500.00 Canadian. For those of you who live in the U.S., it would be quite a bit cheaper, no exchange, brokerage fees, etc.

This shows drilling the holes in the end of the table for the "X" axis stepping motor mount.

The picture above shows the table mounted in the Bridgeport with the right angle attachment mounted on the quill. Milling the table was the hardest part of the whole project, especially when it came to cutting the table "T" slots. That's something I hadn't done before and it took a little doing. The table was clamped down to the milling table and the two sides cut down to the proper height, then I used end mills to cut the slots and finally Woodruff keyseat cutters to cut the wide part of the "T." I broke one cutter and it took quite a while to get all the little pieces out so I could continue. Finally, I used the Bridgeport to locate all the mounting holes in the table.

This shows the machine with the table in place on the assembly.

In the above view, if you look closely, you will see a longer capscrew in between the two Thomson shaft bearings. The ball screw and nut aren't in place yet, so the capscrew is the only thing that keeps the table from rolling off onto the floor. In one of the previous views above you will see one of the lip seals out of place between the two shaft bearings. I had to reverse the lip seals so they would keep the dirt out, not the grease in and that one wasn't put in right.

This is as good a place as any to stop for the moment. More tomorrow.

Richard Trounce.

[FKreider]

Very nice, would love to see some parts that have been made on it!

[datman]

Great thanks Richard.

Ian

[RET]

Hi,

Frank and Ian, thanks for the interest. If you want to see parts that I've made with the machine, look at the Big Boy thread "Union Pacific Big Boy in 3/4" scale." If you follow through it after 2011, you will see a number of things, two number plates, power reverse parts, etc. There is another thread called "3/4" Scale Coupler Update" which also shows quite a bit of the work that I've done with this machine.
Now we'll continue with the build pictures.

This shows a second view of the table in place but without the ball screw.

Because all the holes in the 3/4" plates are perfectly located, lining up the Thomson shafts and pillow blocks on assembly is really easy. You just assemble everything finger tight and push one of the shaft assemblies to the side and lock it down by tightening all the bolts. Next, you snug up the other shaft-pillow block assembly a bit and run that axis through its complete travel, which will push the shaft and pillow blocks into alignment with the first one if any adjustment is required. Tighten up the second set and that's all you need to do. Works perfectly every time!

The ball screws come cut to length (obviously you have to tell them what you need for both length and annealing) with the ends induction annealed so you can machine them to suit the pillow block and timing pulley mounting. They were a little bit on the expensive side, but I used the appropriate pillow blocks from Thomson. If you look closely, you will see that the pillow block at the driven end is always bigger because that one has two bearings in it and locates the ball screw axially with no play. The bearing on the other end just floats in the smaller pillow block. This is the proper way to mount the ball screws. When I was machining the ends of the screws, sometimes the annealed section wasn't quite long enough and I found out that I could machine the unanealed part of the screw if I used carbide tooling. Make sure that you finish machine the ends to a close slide fit and that the floating bearing is a press fit on the screw shaft.

I machined and fitted the "Y" axis screw just fine, but when I did the "X" axis, I forgot to remove the burrs on the shaft ends. When I put the ball nut on, I got bits of steel in the ball nut and it didn't turn freely. Oops!

Now you aren't supposed to be able to take these nuts apart; they come with a cardboard tube in the middle and if you take the tube out before the nut goes on the screw, all the balls run out onto the floor and that's lots of fun. I had to take the balls out, wash everything out in Varsol, clean things as best i could and reassemble the nut. I wound up having to do this 5 times before I got all of the metal chips out. After the first time, I used an ultrasonic tank I had to help with the cleaning operation.

In all this, I found out that the nut and screw are preloaded so there is zero play. This is done by having a very slightly different pitch on the nut than the screw has, so when it is assembled on the screw, the preload is built in. When reassembling the nut, you have to put the balls in and fill the recirculating tubes first (there are two of them because of the preload), then you carefully add more balls and keep the ones you've already added in place by using the cardboard tube to keep them from getting away on you. Lots of fun, especially when you have already done this four times before! Because the ball screws are machine tool quality, they are first machined slightly oversize, then hardened and finally ground to a final finished size. I know for sure I wouldn't want to have to make one!

This shows some of the machining of the column assembly and base.

The picture above shows some of the machining of the base and column, again using the digital readout on the Bridgeport.

In this picture, I'm checking the overarm to make sure it is OK for size.

The above picture shows the overarm being checked for size. You can see the original Sherline mill in the background and that the overarm is not bolted in place but has weights on the back part to keep it from falling off. Both the column and overarm are made from 2" x 5" steel bar because I got that for free from work (yes, I had permission to take it).

Again, this is a good place to stop for the moment.

Richard Trounce.

[RET]

Hi,

Here's the next installment. By the way, the numbers in the heading above are the axes travels, X=16", Y=9" and Z=8." The table is 25" long.

This is the base and column painted and in place.

The picture above shows the start of the permanent assembly on the table with the base and column in its permanent position. You can also see that the base channels are machined to accept the bottom aluminum plate.

A further stage in the permanent assembly.

This shows the bottom plate in position. Its starting to look like something.

Looking good.

More progress. While its nice to see things happen in the pictures, there's a lot of planning and machining that has to take place in between these pictures and it took a while to do it all. It was also pretty obvious that the machine wouldn't be stable enough without some further lateral support, so I added the four 1" dia. round bar aluminum feet at the corners of the lowest plate. This picture shows them in position.

This one shows the Sherline "Z" axis in place.

This view shows the Sherline "Z" axis in place. In this picture, the machine isn't usable because the X axis stepper isn't mounted.

This view shows the X axis stepper motor and lead screw in place

Finally in this picture, the machine is usable. I used the machine this way for a while with the Sherline Z axis, at least partly because for some reason I had difficulty getting the ball screw and nut for the Z axis.

I had made parts for "Dart" with the little Sherline mill, among them were the two coupling rods. The rods were too long for the little mill's capacity, so I had to write code that did one half and then turn the rod end for end in the fixture to do the other end. When the new machine was usable, I decided to make a further aluminum dummy rod just to try the new machine out. After making the necessary code changes, I tried it and it worked just fine. Its always interesting to sit and watch the machine go through its paces. The side rods were of the "fishbelly" type with a 6" radius. With conventional machining, doing that takes a complicated set up. With CNC, the machine doesn't care, it just does it!

Another good place to stop, but we're getting there.

Richard Trounce.

[RET]

Hi,

The first picture shows the actual machining of the aluminum coupling rod test piece I talked about in the last posting.

Machining the dummy Dart coupling rod.

Next we finally start on the permanent installation of the Z axis. This is after using the machine for several months with the Sherline Z axis while I prepared and collected all the bits & pieces I needed.

The start of installing the Z axis.

In the above picture you can see the Thomson shaft support rails, the pillow block mountings and the angle bracket for the ball nut. You can also see the aluminum chip shield on the table and the plastic cover for the Y axis. As I said before, keeping the chips and dirt out is essential.

For each of the three axes, it has always been a problem to get the ball screw nut in place and bolted in. I need smaller fingers, or perhaps a little helper about two feet high.

Finally the installation is complete.

In the picture above, the installation is finally complete including the counterweight bar that takes the load of the Z axis. This way it is just as easy for the stepper to raise the Z axis as it is to lower it. It took a little doing to mount the Sherline spindle on the Z axis plate so it behaved as though it was an integral part of the plate. I finally had to drill and tap the spindle housing in the four corners and bolt it on with four socket head capscrews from the back side. Once everything is assembled and the capscrews all tightened its great, but sometimes getting there isn't easy. That applies to a lot of the steps in building this thing.

Another view of the same setup.

This view is of the same setup, but it shows what is to the left and right instead of what is up on top. Finally I had the machine assembled the way I thought it should be and I used it that way for about a year.

I installed graduated hand wheels on all three axes and one of the first things I noticed was that after every pass, without fail, the graduations would always return to the SAME position they started at. This told me that the machine wasn't missing any steps and that it was accurate.

Just having a larger machine was a big help, it increased the size of the parts I could handle, but I gradually realized that I needed more rigidity to handle heavier cuts with larger end mills like 30 thou. cuts with 3/8" dia. end mills. I could use the larger end mills, but only with lighter cuts of about ten thousandths, so I started looking at how to increase the stiffness between the base aluminum plate and the column.

Time to quit for today. It took about 6 months to get to this point from the time I started on the project. It was very rewarding to see everything take shape and perform just the way I had envisioned it.

Richard Trounce.

[RET]

Hi,

As I said before, after using the machine for a while I could see that the top of the column needed to be laterally stiffer to allow heavier cuts so i started to install side braces. Although I started with the right side, the first picture is of the left side brace because I can take a picture of it easily and it gives a better idea of what the finished braces look like. The right hand one is the same but it is covered up now with the DC control for the spindle motor.

This is the brace for the left side.

I began with the right side brace and built from both ends. I started with a piece of 2" square mechanical tubing with a 1/4" wall because I knew I was going to tap holes in it to fasten it to the underside of the bottom aluminum plate. Tubing is better because closed sections are stiffer. I also chose steel for the braces because steel is easier to weld and has a higher Young's modulus (30,000,000 vs. 10,000,000) so it would deflect less. I used 1/8" wall for the rest of the tube part of the brace because it was just going to be welded, tapping wasn't required.

Next I made and mounted the 3 1/2" square by 3/8" thick plate on the side of the overarm. From that point I built from both ends, figuring out what shape each tube piece needed to be, cutting the tube to shape, machining the faces then unbolting the mounted end pieces, taking them out to the garage and arc welding the bits together. After welding, I would bring the parts back and reinstall them on the mill. Eventually, the last piece of tube was welded on at one end and then both ends were fitted together. I was very pleased to find that when I made the last weld, the completed brace actually went in place without having to be modified at all.

This is part of the column brace for the right side.

I went through the same procedure for the left side brace and that worked out the same way. In the picture, you can see a strap that I added to stiffen up the pillow block mounting plate. That also helped.

With the addition of the braces the machine is a lot stiffer. Now I can take up to 30 thou cuts with a 3/8" end mill, although most of the time I don't try to push the machine nearly that hard. For many parts like the Big Boy shields, I use very small cutters, like .025" dia. for the second shield. Cutting that one was a very slow process, but watching it work was fascinating. The program causes the cutter to go and do a little bit here, then go somewhere else to do some more and eventually come back to where it started to add a little more. After breaking one cutter, I found out that I had to take only one or two thou depth at a time and spray the part with WD40 and brush the swarf away with a toothbrush. With that setup, everything worked OK but it just took time, a lot of it. I used Synergy for that part, there is no way in the world that you can write the code for that piece manually, it would literally take forever.

Recently I burned out the Sherline spindle motor, so I replaced it with a much heavier one. I also built a rotary headstock for the mill and I've found that to be a very useful addition.

That's it for the moment.

Richard Trounce.

[RET]

Hi,

If you would like to see the rotary headstock in use, look at the latter part of the "Union Pacific Big Boy in 3 1/2 gauge" thread. I used the rotary headstock to make some of the parts for the power reverse unit. I wanted to see if I could make the parts look if they were cast and I think I was successful in that.

These are a balanced Pelton wheel design.

The above turbine wheels were made using the rotary headstock. The buckets were cut using a modified 1/8" dia. two flute end mill. The end mill was modified on a Quorn cutter grinder with the cutter relief ground by hand with a Dremel using a binocular microscope so I could see what I was doing. Its a pretty simple program to write; when you look towards the buckets, cut the left one first, then when you cut the right bucket, that removes the burr that is thrown up by the cutting of the left bucket. Easy, right?

I think that pretty well covers the building of the CNC milling machine. It easily works to tenths and is very versatile. As I said before, the machine has turned out to be an essential part of the shop and I'm very pleased to have it.

Richard Trounce.

[RET]

Hi,

Just a few more words and pictures, then we're done.
While I haven't said so before, I used Anvil1000 to make CAD drawings for the hole layout for each aluminum plate and the mill table. I even made a 3D drawing of the table in Synergy to show what it looked like. The CAD drawings help me with the design and with them, I know that everything I draw will fit perfectly before I build any of it.

As I may have said before, if I make a drawing that will be used to write "G" code, the line intersection co-ordinates must be accurate to at least 5 decimal places, otherwise the program will give error messages and won't work. The Sherline software can run all four axes at the same time if your "G" code program is complex enough to need it. By the way, if you are looking for the Big Boy thread, it is on page 3 now.

The new spindle motor after i burned out the Sherline one.

This is the spindle drive motor that I installed on the mill after I pushed the Sherline one too hard and burned it out. This motor is rated at 2 hp @ 130 VDC., but the control I was able to purchase puts out 90 VDC. so that means that I'm getting 1 HP. at full RPM., still a lot more power than before. With the step up pulley arrangement that's there now (in stead of the previous step down setup), I'm getting a top speed of 6,000 rpm. a lot higher than I could get with the other motor.

7 1/4" gauge truck side frame.

This is something else I decided to try. It's a scaled up side frame for a 7 1/4" gauge arch bar truck made from aluminum. Still a work in progress, but its interesting to see how versatile a CNC mill can be.

Synergy 3D CAD drawing of side frame.

This is the Synergy 3D CAD drawing for the side frame.

Finally, for some of the smaller parts I've found that carbide cutters can be quite useful and its worth having a number of different sizes in your cutter inventory. They can break fairly easily, but if you work within their limitations, they hold their cutting edge a LOT longer than high speed steel.

I've also found that if you want to machine thin parts, soft soldering them to a piece of steel bar works very well as a device to hold them. For materials like aluminum, you can also epoxy the part on, but take lighter cuts because the bond isn't as strong. An advantage to using epoxy is that when you reach the join between the part and the holder, you can tell because the epoxy chips obviously look different to the metal ones. This makes it easy to see when you have cut through the part.

I hope all this is helpful to at least some of you. Don't be afraid to go and make your own version. Its an interesting road and the end result is certainly worth it.

Richard Trounce.

[Benjamin Maggi]

Not terribly related, except that you mentioned "Dart" in the first post. You and I conversed via Chaski many years ago as I was thinking of building one. I still might, once my current engine is done.