10"x12" Scratch built lathe

[jakemestre]

Found another weird use for the old lathe.

[dbsharp]

Very cool build. Did you find any places selling or prices on hirth couplings when you did your research? If so why did you decide to make one vs buying?

Looking at your frame, it looks like it will have a relatively small moment of inertia in the area between where your rails are mounted and where it looks like your spindle is mounted. You might gain a lot of rigidity by adding more structure in there. I would think to make proper use of 7 hp you would need a lot of rigidity. When I designed my router, specific answers about machine rigidity is something I spent a lot of time looking for and I did not turn up much. This is funny considering that quite a few people build their own cnc's, and this is a pretty important factor in the performance of the machine, in my opinion. I'm just curious if you have estimated your machine rigidity yet?

Keep up the good work!

[bhurts]

I see you mention wanting to learn about grinding. I do od and id precision grinding on a regular basis. One of the best resources I have found is a ebook on advanced grinding practices. It is a very old book that is still very relevant today especially where speeds and feeds are concerned. Here is a link Advanced grinding practice: a treatise on precision grinding methods and the ... - Douglas Thomas Hamilton, Franklin Day Jones - Google Books to the book.

Ben

[jakemestre]

Very cool build. Did you find any places selling or prices on hirth couplings when you did your research? If so why did you decide to make one vs buying?

Looking at your frame, it looks like it will have a relatively small moment of inertia in the area between where your rails are mounted and where it looks like your spindle is mounted. You might gain a lot of rigidity by adding more structure in there. I would think to make proper use of 7 hp you would need a lot of rigidity. When I designed my router, specific answers about machine rigidity is something I spent a lot of time looking for and I did not turn up much. This is funny considering that quite a few people build their own cnc's, and this is a pretty important factor in the performance of the machine, in my opinion. I'm just curious if you have estimated your machine rigidity yet?

Keep up the good work!

Hello dbsharp,
There are few, if any, places to simply buy a hirth. I wasn't able to find one and my guess is it would have been crazy expensive. They're not hard to make assuming you have either a CNC or a rotary table and good practices on a manual mill.
As far as rigidity, this guy is extremely solid. Thanks for your thoughts on this one, I always am happy to hear concerns. The main plate is .75" thick steel, and the moment generated by cutting will try to bow the plate upwards if anything. I did a lot of analysis of the complete system including plate, headstock and turret and I'm not worried even in the slightest. I'm pretty sure the toolholders will snap before the rest of the machine will give. Also I spent a pretty penny on having the plate heat treated after all the welding, and I'm not about to do it again. lol Thanks again.

I see you mention wanting to learn about grinding. I do od and id precision grinding on a regular basis. One of the best resources I have found is a ebook on advanced grinding practices. It is a very old book that is still very relevant today especially where speeds and feeds are concerned. Here is a link Advanced grinding practice: a treatise on precision grinding methods and the ... - Douglas Thomas Hamilton, Franklin Day Jones - Google Books to the book.

Ben

Wow Ben, thank you so much. You rock. The spindle is still on the list of things to do, and that will be a valuable resource. I've been pretty occupied on several other projects for the last few months. Think I'll post a couple of photos this afternoon. Have a good one!
Jake

[jakemestre]

Sorry for the lack of updates.
I've been focusing on a prototype for a 3D printer that will I'll be selling in the 5-8k range. This is a very rough proof of concept to develop boards and control software along with extruder designs.
I figured a very appropriate first print after calibrations would be a champagne glass to celebrate...

The machine runs on Linux and is at minimum 10x faster than my makerbot Replicator 2. The champagne glass on full speed took a scant 3 hrs to complete. Because the machine is Linux based I'll be able to pull all the knowledge gained to finish the lathe here sooner than later. Gotta focus on the dollars first though.
Jake

[digits]

After a ton of experimentation on the short test section I decided that an annealed surface would be grind-able on my flimsy lathe. This was probably one of the stranger tools I've put in my toolpost.

I'll edit with a short video of the torch being moved based on the thickness of the material. Slower for thicker, and faster over thin for an even temperature. I took it to a gray, almost orange glow which should have been about 800-1000 degrees. Fire extinguisher was in hand the whole time. lol

That is fantastic! I guess that would be perfect for softening the ends of ballscrews before machining them too - even heating due to spinning might help keep them straight

[LeeWay]

That is a cool idea. It probably would help with annealing the ends of ball screws.
I did mine by hand and used a wet rag around to keep the heat from traveling too far. A sponge with a circular shape and a cutout could be used while it's mounted on the lathe. Just apply more water at times.
It would not need to turn very fast at all.

[jakemestre]

Hey forum,
Long time no see! Sorry I've been gone so long, needed to spend effort on the making money front. I've been working nonstop on my 3D printer to bring it to market. I completely rebuilt the proof of concept and am getting some really awesome results. I'll post pictures as soon as the final product is ready to launch.

I've got a bit stuck on the software for the printer for the last couple of days so I figured I'd try my hand at finally grinding this spindle. I've been sort of scared to do it since it requires I hit a very tight mark on a lathe with .02" backlash. And it was extremely hard to find the right spindle to begin with.
I finally succeeded and yesterday evening I hit +-.0001 diametrically after two days of grinding. I had to remove about .1" off the diameter. I ended up taking passes of .0001" running 50rpm with stepover of .125". Thank god that machine is a CNC or I would have died turning hand cranks.
One of the things that was holding me up was that I didn't have a way to dress the wheel once the part was in the chuck. I had been using a boring head held in the spindle to get things true. So I printed a little holder which had recesses for magnets so I could dress the wheel by way of magnetic attachment. Pretty sweet little tool actually.

Here's a photo of the hodgepodge setup.

This little tree grew from the shredding of metal!

On the first few passes the wear pattern of the plain bearings was visible.

And the finished grind.

Now I'm onto grinding for the rear bearing, machining key-way, and threads for lock-nut.
Here's an overall of the lathe.

[DICKEYBIRD]

Very interesting stuff Jake. Is that the T/P grinder from Little Machine Shop? If so, looks like it does a fine job.

Did you have any trouble with the lathe cutting a taper initially?

[jakemestre]

Very interesting stuff Jake. Is that the T/P grinder from Little Machine Shop? If so, looks like it does a fine job.

Did you have any trouble with the lathe cutting a taper initially?

Thanks. Yep and Yep. It's a Little Machine Shop toolpost grinder. I had to make a mount to adapt to the 9x20 but no big deal. If you get one order some spare grinding wheels, and a couple belts. I toasted the one belt(middle of the second to last pass) but managed to hobble through. It runs super hot.
Lathe was cutting a taper initially but I knew that would be an issue. Just had to adjust the tailstock grind straight.
Jake

[jakemestre]

Hey guys,
Progess!
Added a rail for the tailstock(slight design oversight) I think I'm going to find a cheep ballscrew and motorize the tailstock. I think that's actually easier than having to design a movable head, and a locking mechanizm, and, and, and...


Once that was done I loaded everything into the enclosure I've been designing. Pretty straight forward.

I'll be making more but probably not much until the new year. I've got a printer to launch by the 1st, so my energy will be devoted to that.
Happy holidays to everybody!
Jake

[jakemestre]

Happy new year!
Update and absurd setup of the week.
The CNC lathe couldn't deal with a 1-7/8 x 20 thread on the back of the spindle and started screwing up the passes in the threading cycle. So I had to order a die that would get the job done to proceed with assembly of the spindle.

As you can see I didn't have a die holder for such a large diameter so I ended up putting the die into the spindle. But I couldn't clamp it too tight or it would change the shape and undersize the threads so I used some shimstock to keep the die from rotating while cutting threads. The spindle was kept from spinning by the chuck key against the ways. lol And the tailstock I had to move as I was cutting the thread to help support the spindle from tweaking the threads or die. I love trying to figure out how to do this kind of stuff!

I machined the main bearing retaining ring and lipseal. It was actually kind of a pain in the butt to have such a thin piece of 4140. That and I was doing it over the weekend so I couldn't just go buy the right tools either. Half done on the mill and a tiny bit done on the lathe.

I packed the bearings with good ol Kluber and stuck the spindle and bearings in the freezer for a day.
Sorry I didn't get any photos of the process but I heated the bearing housing on the stove plopped the frozen bearings in and when they had come up to temp I inserted the spindle. Everything went in like a dream! measure 100x, cut once in the case of this dumb spindle. Now I know why they cost an arm and a leg!
And lastly the assembled product:

Now onto pulleys, disc break spindle stop, way covers, doors, chip conveyor, and tuning.

[DICKEYBIRD]

Nice work, I enjoy wierd setups too...especially when they're successful!

I forget, are you running Mach on the small lathe? Parallel port w/single pulse/rev indexing or external motion control with encoder? I have varied amounts of success threading with my Mach & P/P equipped lathe. The bigger stuff usually screws up & I end up switching to a manual lathe to straighten things out.

Thanks for the update & keep posting as it progresses!

[vtheanh91]

you can tell me if you use what the compressed air system and solenoid valve type, thank you, my english is not good.

[jakemestre]

Wow, it's been a really long time! Sorry I disappeared, I've been working on building a company based on a 3D printer design. I need my lathe though so time to resurrect the project. I've been building out the controller this week and getting a whole bunch of stuff running. I have controlled motion with Linuxcnc at the moment, and I'll be posting some updates as soon as I get some pictures. I think I'll probably also do a video on how to get Linuxcnc to work on a setup. Dickeybird I hope that answers your question. Vtheanh91 for the Turret I think is what your question is asking, I'm using a double acting air valve and corresponding solenoid like this one. It runs on 120PSI. The turret is going to be an interesting beast to get working properly, as there is little out there on turrets.
https://www.automationdirect.com/adc...)/AVS-3111-24D

[mountaindew]

I noticed your spindle head is cropped out in your pictures So I cant see what your using there. An idea that come to me a while back was a Rapid turn spindle head would make a good starting point for a small cnc lathe build. Just passing on ideas

[jakemestre]

So much update! I've almost got the machine running. A couple bumps along the way but everything seems to be working out. I'm really hating and loving LInuxCNC all at once. There is so much you can do with it, but programming everything and learning how to troubleshoot is a seriously steep learning curve. Here are a few photos:
Pulley Making time: I can't wait to have coolant on a machine again and the ability to break a chip.




The backside had to be recessed on the mill to allow for the jaws to clamp and jbelt grooves machined.



This is the spindle encoder I've printed a mount to install and tension. It's a very simple AMT103 encoder from digikey. It reads beautifully in linuxcnc, and I have to say it's really cool to see the spindle number changing when I'm barely rotating the chuck at all.



I added proximity switches to the turret to both index, and know if the turret has reengaged correctly before cutting starts again. This is the final reassembly photos. I will at some point print some TPU gasket to protect the hirth teeth from chips and coolant.


I'm using a nema17 stepper with a Mechaduino board to make sure the turret arrives in position.

Cable chain is installed and a friend was helping me and designed a small junction box that affixes the chains and keeps the cables free from chips.




The rats nest: I had just switched to a brand new computer to run the machine. The old computer kept having cpu overheat shutdown events and I didn't want that to happen while running a program.
Also I wanted to use DIN terminals to get signals in and out of the 7i76. It is definitely not a break out board, but rather a signal in/out board. It is really nice once you know what you're doing but power, signal, ground, shield are not broken out together in labeled headers.
So you have to use these beautiful beautiful diagram that Norbert(newbynobi on the linuxcnc forum) created for the 7i76e. That's the Ethernet version instead of paired with the 5i25 pci card.


Next up cleanup.

which is slightly better, nothing like the Grizz but it'll work.

A little better, and all the 120 outlets are powered and controlled via software.

I'm waiting on the VFD as the old one had some issues. TECO has been absolutely amazing to work with. Seriously, they are amazing, one of the best companies I've ever done business with.

I've got the machine moving, homing and each axis controlled with the switches and MPG on the control panel. It was about a 3 day process to figure out why linuxcnc would not work with the MPG. Finally after leaning how to understand what's going wrong with the hal file, the custom.hal I created for jogging was perfect and I could see everything in the halmeter. Turns out the old MPG wasn't outputting signal and the code worked great after all the troubleshooting. It uses a 3position switch to toggle between the x, z or not selected and a rotary 2 position switch to send the bit logic to hal which selects the unit. .0001, 001 .01 or .05 I find that using it at the .01 is great for direct jogging. At anything greater than .01"/pulse the the speed you're spinning the wheel exceeds the max speed of the closed loop steppers and the machine keeps going after I stop spinning the MPG.



If anything it makes a good nightlight for the shop...



As is:


Up next is get the turret working and changing tools based on what's selected. I need to make some tool holders as well and finish off the enclosure. I think the chip conveyor is going to go on the back burner for a bit and I think at some point I'll build in a touch-off probe.

[jakemestre]

Turret is almost working. Now to figure out how make ClassicLadder execute this gcode simulated toolchange.

[pippin88]

The other option for the turret is a linuxcnc 'component'. They are just short C programs with connections to hal. That's what I used for my turret. I'm more familiar with C and find the classicladder logic odd / difficult.

[jakemestre]

The other option for the turret is a linuxcnc 'component'. They are just short C programs with connections to hal. That's what I used for my turret. I'm more familiar with C and find the classicladder logic odd / difficult.

Hey pippin,
Thanks for the suggestion. I did a little more digging with classic ladder and there's not really a good way to control an axis really with the outputs. Unfortunately I'm not so good with C, so a remap is what I need. Working on that part now. For the moment I'm using the motion.output to control the turret lock, and an A axis coordinate to swing the turret around. I'll post an update in a little while.