Alan's Grizzly 9x19 CNC Conversion

[acondit]

I previously posted this info in another thread. However, since I am now working on my conversion in earnest (I got my router running). I thought I would go ahead an start my on lathe thread.

Here are a few pictures of my variable speed conversion.

I used a Pacific Scientific DC motor that had a shaft the same size as the OEM Emerson motor (5/8"). This allows the use of the original pulleys and belts.

I made the motor adapter plate from a piece of 3/8" aluminum and a piece of 3/4" aluminum. I used a round-over bit to mill the recess for the motor.

I made the controller bracket from 1"x1"x1/8" HR steel angle. It bolts to the same holes as the original switch bracket.

I bought the Dan Foss C150 controller on ebay. It is currently using a pot to control the speed but it supports a 0-10v DC signal to control the speed. This can be obtained with the use of a board like the Digispeed boards from Homann in Australia.

Alan

Similar Threads:

• 9x19 G4000 CNC Conversion
• Alan's G0704 Conversion Planning
• New Machine Build- Just got my Grizzly X3. CNC conversion to come...
• Grizzly X3 CNC conversion..help

[acondit]

Here are a couple pictures of the power supply that I am building for my 9x20 CNC conversion.

I designed a power supply board to make building power supplies easier.
The first photo is a picture of the PS board. The second is the components mounted on an aluminum chassis that I formed at a friends shop.

My ps chassis has my PS board, an Antek toroid transformer 23V 400VA to provide power, a 35,000uf 80v filter capacitor, (2) Superior Electric SS2000MD4 drives, and a CandCNC mini-io breakout board. It outputs about 32V dc at around 16amps. Should be enough??

The power supply board has a full-wave bridge to rectify the AC, a surge limiter to keep input surge from doing bad things, a bleeder resistor, connections to the filter capacitor, and four fused outputs to the drives.

For the lathe I only installed two fuse holders and two drives. I plan to use 2 PacSci 253oz. steppers, one to drive the leadscrew, and one to drive the crossfeed.

Alan

[acondit]

This part of the project originally began as a bearing support modification to the crossfeed leadscrew. I was never happy with that part of the design. It had a bronze bushing in the back end and a single ball bearing in the front end. For a manual screw it would have probably worked fine, but it always nagged at me. I had thought about several different schemes for mounting double bearings at the front end to take out any backlash there. One day I decided to just do it.

I turned off most of the pocket for the front bearing. Then I cut another piece of aluminum and cut the remainder of the back bearing pocket in it. Then I flipped it around and cut the front bearing pocket in the opposite side. So the back bearing aligns the front part of the bracket with the rest of the body and together they support the screw radially and act as angular contact bearings.

Next step was to design a method of driving the crossfeed with a stepper.

Alan

[Ron111]

Looks good. Are you planning to use a ball screw and ball nut on the Z axis?
Like your X-axis hardware, is that the stock screw? I cnc'ed a 9X20 last year and it had a 16tpi acme that performed ok with about 200in/oz.
Ron

[acondit]

Looks good. Are you planning to use a ball screw and ball nut on the Z axis?
Like your X-axis hardware, is that the stock screw? I cnc'ed a 9X20 last year and it had a 16tpi acme that performed ok with about 200in/oz.
Ron

Yes, I am planning on using a ballscrew on the Z axis. I have a 6' length of 5/8"x5tpi ballscrew and a double ball nut.

My X-axis has an extra 3/4" of travel. I replaced my original crossfeed lead screw with a 7/16"x20 leadscrew. The new crossfeed nut has an adjustment to reduce backlash. Eventually I hope to replace it with a ballscrew as well, but I haven't reached a final design for that yet. I am leaning towards a rotating ballnut and fixed ballscrew. That would allow me to have the crossfeed ballscrew in the original tunnel.

Alan

[acondit]

Once I figured out the manual bearing support setup for the crossfeed, the CNC drive was easy. I took a piece of 3/4"x4 aluminum bar and figured out the center for a 60T .2 pitch belt gear. Next I figured out the center location for a M23 Stepper motor.

I machined front and rear bearing pockets to match the bearing pockets on the manual bearing support.

Then because the stepper shaft was too short to go clear through the 3/4" material and support a belt gear on the other side, I milled a pocket on the reverse side to set the M23 stepper part way through the plate.

Now I have to figure out the belt tensioner.

Alan

[acondit]

I took the spindle apart because I had read that part of the finish problems that I have had over time could be because of the spindle bearing adjustment. I had to drive the spindle out with a hammer and a block of hardwood. When I mic'd the rear bearing journal, it was 1.3780. I mic'd the inside of the rear bearing and it was 1.3776. I also checked the inside micrometer against the outside micrometer and they were in agreement. I chucked of the spindle in my other lathe and used a fine flat file and some 400 wet or dry sandpaper to achieve a snug sliding fit.

Then I discovered that the gear and pully wouldn't slide on either. There, I discovered that the key didn't want to come out of the spindle slot. I finally clamped vise grips on the key and used a brass hammer to tap the vise grips. It took about twenty tries before I finally got the key loose. Without the key in both the gear and the pulley were an easy sliding fit. So, I dressed the sides of the key until I could hand press it in and pull it back out with pliers. Then everything would go together properly.

I cleaned up the bearings, greased them with "Super Lube" a teflon bearing grease. I put it all back together and adjusted the fit.

Now I need to cut something to see if it improves the finish.

Alan

[acondit]

I know that I ultimately don't need the DRO after I have it CNC'd, but I need it right now. So I am trying to mount my Shumatech DRO-350 on my G4000 as a temporary measure.

I already had the DRO, QC-100 quadrature to chinese signal converters, USDigital encoders and encoder strips(7" and 22"). So it is a question of mounting.

I took a small piece of delrin and machined a holder for one of the encoders. I machined a slot into the holder so that it would slide on the edge of a piece of 1" square aluminum that I cut a 0.625 slot in the bottom.

I still need to mount the linear encoder strip in the aluminum tube and mount the whole thing to the cross-slide. Then I have to do the same thing for the bed to carriage encoder.

Alan

[twocik]

Nice job!

[acondit]

twocik,

Thanks for looking, I am reworking the encoder mount/slide. I am trying to resolve interference between it and the gib adjusters on the crossfeed.

I thought about mounting the encoder on the chuck side of the crossfeed but then it limits how close I can get to the chuck (while at the back it limits how close you can get the tailstock to the crossfeed). It makes some of the ideas about using a cable driven encoder sound enticing. Oh, problems, problems. Well a new tailstock design would solve that, but I promised myself no new projects till I finish those already on my plate.

Alan

[acondit]

I got the encoder carrier redesigned. Now it has slots for slides on both sides and it will clear the gib adjusting screws on the cross-slide.

I got some 3/8 x 3/8 ground stock to mount the encoder strip. I am going to machine a couple of delrin plugs so that chips can't enter the ends of the square tube. Then I have to fasten the encoder strip to the 3/8 x 3/8 spacer. Then I will have the first tube done. I want to hook it up to the DRO and test it before I mount it on the lathe.

Then I still have to cut the slot in the Z-axis tube, put the Z-axis all together and mount everything on the lathe.

Alan

[acondit]

I have been working on a controller that will allow the lathe to work in manual mode (no gcode) using the cnc electronics. I designed a small circuit board that has two modes. The first mode acts as a variable speed automatic feed and the second mode uses an encoder to generate steps as the handle is turned. You can select either the X or Z axis in either mode.

Alan

[acondit]

I used one of my optical switch boards to feed the tach on my DRO-350. With one of my little PacSci steppers and my "manual" controller I can vary the speed from 252 rpm up to 3072 rpm before the motor shuts down. Some of my other steppers shutdown at about 600 rpm.

I think I am going to adjust the speed range down. I am looking at something like 15 rpm to 1500 rpm. With my current gearing this would give me feeds from about 0.187 ipm to 18.75 ipm. Does that sound OK? Or is it still too high?

Alan

[acondit]

I've been busy with other jobs but I finally got a few free minutes. When I put the stepper plate on the crossfeed, the belt was loose on the pulleys. I had not designed in any adjustment, but I had thought about the solution. I had a little adjuster that I got off of another stepper I bought on ebay. So, I machined a pocket for it to sit in and voila "instant adjustment".

Next step is to figure out the Z-axis screw. I thought about leaving the stock lead screw there but I have a number of double ballnuts and a 6' length of ballscrew, so I think I will go ahead and replace the standard lead screw with a ballscrew now. Eventually I plan to replace the crossfeed screw with a ballscrew as well.

Alan

[acondit]

I decided that the best way to deal with the Z-Axis was a ballscrew and that the best way to do that was to machine an end to attach to the ballscrew that would fit through the bearing block and connect to the motor. It has a 0.500 bearing surface with a 1/2x20 adjusting nut and a 0.375 connecting end. I also need to shorten the connecting end shaft by about 0.25".

I am using an existing Nema23 motor mounting plate, but I had to make new hexagonal standoffs because the ones I had were too short for the adjusting nut and the lovejoy connector. The new standoffs are 2.5" long and both ends are drilled and tapped 10x32.

The bearing support will fit in the same location as the stock bushing for the stock leadscrew. Now I have to cut off a section of ballscrew, anneal the end, and turn it to fit in the end of the fitting I made. I am thinking of attaching it with silver solder. Is that a good idea or not? I could braze it or pin it but silver solder is not quite as hot as brazing, less chance of warp, I think.

Next I need to decide how to mount the other end. I am thinking about annealing it and turning it to fit into the hole where the original leadscrew connects to the standard gear box. However, I could just build another bearing support to mount to the bed. Any suggestions.

Finally, I am planning to remove the existing carriage that attaches to the saddle and attach a plate to which to mount the ballnut.

Alan

[jhowelb]

I left my gear box alone and just turned that end of the screw to a couple of thou under the size of that mount fixture for the original screw and just let it turn inside there. No problem to date.

On the adapter, you could turn the end of the screw to about .002 larger than the pocket you've built and press/drive it home. I've had good luck doing that cold. For a firmer joint , make the shaft a bit larger then cool the shaft with dry ice and heat the pocket then press them. The wheels on rail cars are attached to the axles with this exact process....no weld or solder!
jb

[acondit]

I left my gear box alone and just turned that end of the screw to a couple of thou under the size of that mount fixture for the original screw and just let it turn inside there. No problem to date.

That is where I got the idea.

On the adapter, you could turn the end of the screw to about .002 larger than the pocket you've built and press/drive it home. I've had good luck doing that cold. For a firmer joint , make the shaft a bit larger then cool the shaft with dry ice and heat the pocket then press them. The wheels on rail cars are attached to the axles with this exact process....no weld or solder!
jb

John,

Thanks, I read that on your thread before but had forgotten about it. It sounds like it might be worth trying. Although I live in a small town and I am not sure about the availability of dry ice.

Alan

[jhowelb]

Frankly, Alan, the cross feed screw on my lathe took no ice or heat and has never slipped.
The purpose of heat is to expand the size of the hole and reverse with the cold. It will shrink some in your fridge or freezer and expand on the BBQ.
Heck, give it a try! The worst that can happen is that you learn from the experience. If you botch it completely, you can always braze it or even spot weld.

That reminds me check this out! Post#5
http://www.cnczone.com/forums/showthread.php?t=29339

[acondit]

Well I took John Howell's advice and turned the end of the ballscrew about .002 larger than the pocket in the end piece. Now I just have to find some dry ice to chill the end of the screw.

Alan

[jhowelb]

Looks like a nice job on that screw, you might just ride the leading edge with a file for a taper to help get a start. I think you'll find putting the two together will be anti climatic. I've come to rely on this method a lot!!