HF round column mill drill build

[109jb]

Great build I am planning on doing the same to my mill some time soon (still a newbie). What do you mean by reversing the quill spring? Do you simple turn it around and what does that do? Do you still get back last in the z using the fine feed adjustment and not a ball screw?
Can you show more details of your machined ball screw ends?

Thanks!

Reversing the quill spring - The quill spring is set from the factory to return the quill to the up or retracted position. This is fine except that most all cutting tools are up-cutting and want to pull farther into the stock. If you just put a motor on the fine feed and don't reverse the spring, the tool can pull itself down far enough to take ot the backlash in the worm drive of the fine feed mechanism. On my mill, that backlash was 0.028". Unacceptable!! By reversing the spring, the quill is now spring loaded into the extended position, so with the fine feed engaged, the tool can't pull down because the quill is already against that end of the backlash of the worm drive. Simple but this simple solution probably wouldn't have occurred to me.

On the turned ends of my ballscrews, one end is just turned down for a short way to 12mm diameter. This is the floating end. The other end, I turned a bolt with a 12mm straight portion, a 12mm threaded portion, and the very end is 3/8" diameter. The lengths were to suit my particular installation and the bearings I used. I then welded the turned bolt to the end of my ballscrew. I can get you some pictures, but not right now because the machine is together I will be making some new zero-backlash parts soon for my ball nuts and will take the machine apart to install them. When i do I can get you some pictures.

[109jb]

Well, since I'll have my round column mill working in 3 axes soon, I need something to make with it. I sold some stuff and was able to get enough hobby money to buy a G0704. I still want an RF-45 size machine, but the G0704 will fill the bill for at least a few years. All I had time to do today was uncrate it and get it on a Shop Fox mobile base I had. I did plug it in to make sure everything works but that's all. Shipping was quick. I ordered monday and git it today (Friday). Nothing damaged, and no chipped paint.

[109jb]

Well I finally got time to complete the re-work of my anti-backlash ball nut setup for my x and y-axis screws.

The original anti-backlash setup didn't have strong enough springs and allowed some backlash unless the gibs were too loose. The new setup uses 2 wave washers that each have a compressed force of 101 pounds. With the setup I used I am able to get infinite adjustment up to the max which in this case is 202 pounds. I adjusted both to about 175 pounds of preload.

I also received my 3rd TB6560 single axis driver board and got it hooked up to my Z-axis drive. I did a little more tuning of the whole system in EMC2 and did some backlash testing. I am proud to report that I have zero backlash and no backlash compensation entered into my EMC2 setup. The only thing I would still like to do is to map the screws to account for any manufacturing deviations in the rolled screws. This will hopefully get me 0.001" accuracy.

As a test, I milled a bearing pocket for a 22mm (0.8661") diameter bearing I had. I decided to shoot for a 0.867" diameter pocket. I used a new 4 flute center cutting end mill and measured it at exactly 0.375" diameter using my micrometer. The program I wrote uses cutter diameter compensation. I entered the cutter values into the tool table as tool #1. The program makes 2 roughing cuts each 0.100"" deep and 0.010" undersize on the diameter. The final full depth pass is a climb milling cut to clean up the diameter. I did a dry run and managed to drag the tool across the top of the part. That is where the scratches in the video came from. I guess I should have set my z-zero at 0.250 above the part instead of 0.200 since the final depth is -.200. Brain fart!! The linked video is the first real run of the program. As you can see the bearing fit nice and snug. After I shut off the video, I measured the pocket with my inside micrometer and can report that the pocket is within 0.0004" of perfectly round measured at -45deg, 0deg, 45deg, and 90 degrees. First pass and the size was within 0.0005". The diameter range of the 4 measurements came in at 0.8669, 0.8665, 0.8667, 0.8669 respectively. The target was 0.867. Pretty darn good if you ask me.

[109jb]

Well it has been too hot here in Illinois to work in the shop without air conditioning and I found out that our electric service was at very low voltage when demand was high. During these 100 degree days we have had the voltage would get as low as 89V. At that voltage, the air conditioner in my workshop would just trip the breaker after about 5 minutes. The utility finally got a new regualtor installed up the road from me and we now have good solid 110V. So, with the heat, no air, vacations and work I haven't been able to do a whole lot on either machine until now. I should have all my parts made for my G0704 build hopefully this weekend.

First up was to thread my ball nut brackets for the 15/16"-16 thread. I decided to thread mill it using an old 3/8"-16 tap as my thread mill. It was a 3 flute so I ground two flutes off and left all of the teeth on the third flute since the thread pitch is the same. Here is a video of the whole process. It came out good and the threads are tight just how I wanted them.

[ngr1]

Hello, thank for sharing your build with us.

With respect to the thread milling using a 3/8 tap, what value did you use for the size or OD of the tap within the wizard or program that you created?

I am trying to use a thread mill wizard in MAC3

Thx.

[109jb]

Hello, thank for sharing your build with us.

With respect to the thread milling using a 3/8 tap, what value did you use for the size or OD of the tap within the wizard or program that you created?

I am trying to use a thread mill wizard in MAC3

Thx.

I can't comment about Mach3 because I am using LinuxCNC, but I used 0.281 radius in the program I used for thread milling that 15/16"-16 hole with the 3/8" tap ((0.9375-0.375)/2=0.28125). I wanted the threads tight, so I started with 0.281 for the radius. The program I used is as follows. I added comments in red to explain:

G90 Set absolute mode
G53 G0 Z0 Send z home
G53 G0 X0 Y0 Send x and y home
The first 3 lines are how I start all my programs
G0 X0 Y0 X0 Y0 for this part was the center of the hole
G0 Z0.1 Send z to 0.1" above part
G1 X-0.281 Move out to the radius
G91 Incremental mode
o103 repeat [10] Start a loop that repeats 10 times
G2 Z-0.0625 I0.281 circular milling with a -z equal the thread pitch
o103 endrepeat End the loop
G90 (Absolute mode)
G1 X0 Send X back to hole center
G53 G0 Z0 Rapid Z up to home
G53 G0 X0 Y0 Rapid X and Y to home
M02 End program
The last 3 lines are how I end all my programs

I used the loop in this program because I am using LinuxCNC 2.4.6. Versions after 2.5 don't need a loop to do this. The loop runs 10 times and each run through the loop make one turn of a helix with a 0.0625" z component, so in total it goes 0.625" deep. The G2 command only uses a Z and an I word because in LinuxCNC it is assumed that if X and Y are omitted that it makes a complete circle ending at the start point, and the J is assumed to be 0 if left out. Not sure how Mach3 handles this stuff. Also not sure if Mach3 does loops in the same fashion as linuxCNC.

Hope this helps.

[amyers]

Hello, thank for sharing your build with us.

With respect to the thread milling using a 3/8 tap, what value did you use for the size or OD of the tap within the wizard or program that you created?

I am trying to use a thread mill wizard in MAC3

Thx.

hoss used the mach 3 thread milling wizard in his video from last year. he was doing the same ballnut thread.
http://www.youtube.com/watch?v=Suwl_HzouVU&feature=plcp]G0704 Economically Threadmilling a Ballnut Mount - YouTube
am

[109jb]

Posted this in the wrong thread of mine This is where I intended to put it

So I cleaned up my wiring and when doing so i started at the motors. I made theconnections to the Y and Z axes with piece of 4 conductor cable before i cut it in half to terminate the controller end. It was kind of funny to be able to turn the Y axis motor and watch the Z axis extend.

I also finally figured out why I wasn't able to get the Z axis feed rate I expected. If you look at my pictures for my Z axis setup you will see that I had A 1:2 pulley ratio with the smaller pulley on the motor. Set up that way I was only able to feed about 1/4 as fast as the X or Y axis. I switched it around with the big pulley on the motor and now I can feed as fast as the other axes. I can get about 100 ipm since I am running at 24v and 3a instead of 50v and 5 a. I have it slowed down to 72 ipm for a bit of a buffer to insure I don't lose steps. Working great now. If only I could find time from work and necessary household things to actually make some parts.

[109jb]

Still haven't had much free time to work on the new machine, but I'm to the point where I am ready to sell my round column Harbor Freight mill. This is the one in this thread. There is an ad in the classifieds section of the forum with a link t a demo video for it. Mill is in Morris, Illinois about 1 hour southwest of Chicago.

[cknowlto]

After conversion, if you want to manual mill, are the steppers interfering with your moves or do you have a way to disconnect them so you can move smoothly?

[109jb]

After conversion, if you want to manual mill, are the steppers interfering with your moves or do you have a way to disconnect them so you can move smoothly?

My machine doesn't have any handles on the table, although you could easily put some on there. The quill fine feed is has the pulley mounted on it, but once again, you could put a handle on there. Also, the quill is spring loaded down now and that would make using the quill handles, the ones like a drill press has, more difficult because the quill wants to go down rather than up. The problem on the table is that the ballscrews turn so easily and can be back driven. I would be afraid of the table moving while taking a cut. Again, you could devise a lock for the screw as well.

In a nutshell, you could preserve all of the manual features of the mill pretty easily; However, once you use a CNC'd mill you will never want to crank handles anyway. It is very simple to use Jog mode or MDI mode to accomplish the things that you would do manually.

[109jb]

A little while ago someone PM'd me asking if I had any drawings of the conversion because he had the exact same model machine. My inbox was full and in cleaning it out I inadvertently deleted your message. If you see this, send me another PM because I was able to find the drawings. At least the most important ones.

[CNC-Joe]

Nice job on the build, 109jb. What did you layout your control panel in ? It looks sharp!

[109jb]

I used Autocad to lay out the control panel. Unfortunately I never built it. The reasons are many, including not enough time, the keyboard shortcuts in LinuxCNC are adequate, and I was spending a lot of time getting our house ready to sell. I have been steadily getting rid of stuff in the shop in case we had to move and it is a good thing since we have accepted an offer on the house.

[CNC-Joe]

Very nice build log, 109jb! I also have a round column mill that I've been considering converting (( although I have a nice CNC'd SuperMax YCM-30 knee mill, you can never have enough CNC machines, can you ? )). Could you put together a list of your material used, where you bought them, and approximate cost ? Example your Angle Contact bearings, driver cards, etc. etc..

[109jb]

Very nice build log, 109jb! I also have a round column mill that I've been considering converting (( although I have a nice CNC'd SuperMax YCM-30 knee mill, you can never have enough CNC machines, can you ? )). Could you put together a list of your material used, where you bought them, and approximate cost ? Example your Angle Contact bearings, driver cards, etc. etc..

Ball screws (3ft 5/8" rolled) and ball nuts (4) were from McMaster Carr about $200
Angular contact bearings were from E-bay about $8 each x 4 = $32
Aluminum for bearing mounts - I already had this but probably got it from Speedy metals. Total of about $10 of aluminum in this project
15/16 used tap from e-bay was I think $15
(3) single axis TB6560 stepper drivers - $15 ea x 3 = $45
Chinese 5-axis BOB from e-bay $15
(3) 570 oz-in NEMA 23 steppers from automation technologies - about $150
timing belt and pulleys from econobelt. about $30
e-bay 24V power supply I think it was about $25
Computer and monitor - free from a company that was upgrading. Just had to add a hard drive which I already had.
Control software - LinuxCNC (free)
Misc stuff was probably another $20 to $50

Total about $500 to $600. Seems about right

Note though that I had always intended to swap out the power supply for a 48V and the drivers for something else but I never did. I sold this machine with the tb6560 drivers and got a G0704 before I changed anything on this one. The motors were severely under driven with 24V when they want 50V. It worked OK but didn't have the speed that it could have with other drivers. I am using Keling 6050 drivers on my G0704 and they work great. About $40 each. I am also using a 48V server power supply that cost me $35. This setup works great and has plenty of power.

[CNC-Joe]

Thanks 109jb ! I didn't realize that you had stopped work on the round column mill and sold it. Your information on a 48V Server Power Supply for $35 - is really interesting (( I've been spending much more on power supplies, although I have been putting in 80 Volt power supplies on my machines )). I did build an Unregulated Power Supply (aka UPS) for my CNC router. It generates about 57 volts - and cost me a WHOLE lot more (Toroidal transformer, bridge rectifier, large capacitor, bleed resistor, etc. etc.) I still want to convert this round column beast to CNC and add a tool changer... I know its an addiction.