[Ron111]

Bought a 9-20 for $300 at a HF sidewalk sale about 3 years ago. I have a shop task 3 in 1 that I use for milling and as lathe. I want to convert the 9-20 to a cnc. I want to maintain the original cabilities and haven't seen plans for the 9-20.

My thoughts are to couple to the original lead screw at the tail stock end of the machine and just engage the half nut for the long axis drive, and then do a direct couple (no gears or belts) to the cross drive axis.

Most of the work that I do with the lathe is to turn down m-16 stock buffer tube, which are about 7 inches long and roughly 1.25 dia. and I also make a small pull pin out of non hardened 1/4 inch dia 1 inch long steel dowel pins.

So, I currently, am not doing any threading, but I still want to maintain the capability if I ever need to.

I looking at of the plug and play controllers that cncresources offers. As far steppers, my uneducated gut feel is that a 200 oz/ in would be ok for the cross axis and maybe a 267 oz/in for the long axis. I already have mach II and am getting used to it.

I emailed Ron Steele to see whether his plans would be helpful and he didn't think that they would be too useful.

So, guys start scratching your heads and let's figure out is my thoughts about coupling into the existing lead screw makes sence at the tail stock end.

Give me your thoughts about the stepper motor sizes. (I would be happy with rapids say of 25ipm)

And some hents how to couple at both the Z axis and the long axis.

I've seen two or three guys that have indicated interest in converting these 9-20 so lets put our heads together and do this for the Gipper.

Thanks

Ron

[mxtras]

My intention was to eliminate the cheesy lead screw in favor of a smallish ballscrew for the saddle. I also have a 9X20 that's next in line for CNC but it's not a huge priority at the moment.

I will be following your threads - do this right so I can just copy you, ok? JK......



Oh - I was going to couple the motor at the tail stock end, too - either direct drive or a 1.5:1 reduction for torque and accuracy improvement. There are plenty of surfaces on that end just begging to be drilled into.

Scott

[wizard]

Here ae some negatives to consider:

1.
The 9x20 is less than ideal for CNC due to the spndles screwed on chuck. Of course this is not a problem is you have no intention of using the chuck. Also I've heard that the spindle itself on this lathe can be swapped out with a spindle from another to give you bolted on chucks.

2.
The leadscrew on this lathe and most lathes for that matter, is less than ideal for CNC. There are a number of problems from fast wear to lack of protection for the screw. An exposed screw on a CNC lathe with heavy usage will not last long.

3.
A CNC lathe cries out for a good variable speed spindle. This is an AC drive.

NOW this isn't much of a negatives list but some of the issues could be huge depending on your interests. Since the work you describe is a bit limited in scope you may do well with a CNC conversion. Here are some ideas for you that may or may not be usefull.

1.
If you expect to need manual operations such as threading in the future I personally would not touch the leadscrew that came with the machine. You have two approaches here. One would be a regular leadscrew mounted in a free position on the lathe, the other would be a rack drive system.

The alternative lead screw could be mounted either front or back side. For a rack drive I'd use the current postion but upgrade to better quality hardware. To do the rack drive you would need to implement a low backlash solution, probably with split gears. The nice thing here is that if you use a dual shaft stepper you could mount a handle on the other end and loose nothing with respect to manual operations. There might be slightly more work involved in the rack approach but I see it as the most flexible approach.

A rack Drive may be wanting for some CNC usage but in this case it may work well and has the potential for being less costly.

2.
On either axis I would prefer to see you gear the motor down. Not so much for torque, but to increase the resolution. I suppose one could go through the math but if you have 1.8 degree steppers you may find your resolution wanting. Further you really want more resolution than the process tolerance. I'd shoot for a postioning capability that is atleast a quarter of what you are hoping for relyiing on no more than half stepping the motor. Some may think that that is not good enough either.

As far as torque goes you will need somebody with more tooling experience than me to tell you how much torque you will need for a give cutter and material. In the end though you only have so much horse power to deal with here.

3.
Right now I'm limited by time and money as to CNC projects, so I really can't just jump in and start a conversion on mine. Parts are slowly being collected though they are likely to end up going into a CNC mill of some sort. So any help I can offer would be based on other experiences. The one thing that did catch my eye is that you have very specific usages in mind which is a good thing, but it also leads me to a suggestion. That is to keep your manual lathe, because you will always want a manual lathe, and consider CNC a different lathe.

The advantage is that you won't need or care about the manual capabilities at all and can thus build the lathe to good engineering principals and design. The CNC won't be a compromise and the manual hardware can be completely eliminated.

There are other options also. One is rebuilding a used CNC lathe. Another is the build of a special purpose lathe.

All of these ideas revolve around keeping the current machin unencumbered with CNC stuff.

4.
You need to put on your list of stuff to consider is a spindle drive. Ideally you will want this under CNC control also. The problem is setting up a drive with a single reduction that will cover the spindle speed ranges you have need for. Since one part is steel and the other Aluminum, but vary in size, this might not be to much of a problem and solvable with a simpler drive. You probably already have enough experience here to know what the speeds are that you need.

5.
As to the specfic question about coupling into the lead screw at the tailstock end I'd have to say that it probably isn't the best of ideas but then again running through the quick change box might not be to smart either. Atleast with the current lubrication systems. In the end I think you will need a new lead screw to do a really good job here. Yes I've seen shaft extensions and the like on other lathes and I honestly don't think positively about them.

When going CNC a lead screw has to handle certain loads that are not seen on manual machines. So you really need a bearing system that handles axial thrust on the shaft well. You also need a mounting method for the pulley or coupling that handles the torque and the associated rocking back and forth. Even on good quality CNC lathes much care has to be put into these areas. Remember all torque for axis motion will now go though the pulley or coupling and is likely to be mor e intense than you can manage manually. This can all be done from the tail stock end of course but in doing so I'd remove the QC box completely.

6.
As for the y axis (crossslide) I'd consider milling that down and adding linear bearings. Either that or having tericite put on the ways. The reality is that I don't see the slide holding up well under CNC operation. Of course that could be argued for the whole lathe as it is manual, but I see the crossslide wear being confined to samll back and forth movements. This concentrates wear and makes it hard for hte lubrication system to do its job. On some industrial machines we had such a problem and programmed the machines to slew the axis back and forth every so often, over as much axis range as possible, to spread out the way oil.

7.
As to the controller I can see the advatages of buying a canned system. But you do have to make sure it fits your needs and thus need to know what the vendor specs are. My personal goal is to DIY the controller, but do realize going this route involes finding the parts. I'm not sure if you have settled on software yet but you might go to the LINUXCNC site to read up on info they have there. It is very good info and of a general nature. Also getting involved with the gecko site may be helpful as there is good info to be had at that site. You do have to make sure that he motors and dirves are matched up correctly.

Thanks
Dave

[Ron111]

Good input Dave,

Well, I have a prototype x & z stepper mount in place, and made a couple of oldham couplers (just have to drill final motor shaft size hole), now I just deciding on the control break out board and possible a cnchobby driver board which supplies the 3 amps which is more than the 282 oz in and the 453 oz in stepper motors require.

I currently examining the pulse generatoring circuitry in order to use Mach3 and it's threading capability. When I get proficent, I'll cut down a ball screw for the long axis.

Guys, need some input about pulse generating using either reflective or path breaking techniques.

Thanks,

Ron

[wizard]

Good input Dave,

Well, I have a prototype x & z stepper mount in place, and made a couple of oldham couplers (just have to drill final motor shaft size hole), now I just deciding on the control break out board and possible a cnchobby driver board which supplies the 3 amps which is more than the 282 oz in and the 453 oz in stepper motors require.

Sounds like things are coming along here. I'd go with Gecko drives as they seem to hit the seet spot as far as performance and cost. The idea of a little head room in the components is also a good idea.

I currently examining the pulse generatoring circuitry in order to use Mach3 and it's threading capability. When I get proficent, I'll cut down a ball screw for the long axis.

Guys, need some input about pulse generating using either reflective or path breaking techniques.

Optical techniques will work both reflective and beam breaking. The trick is to find a optical sensor fast enough to handle the speed of the interrupted or reflective patch. So it is up to you to determine the method you want to use. If forced to I'd go the beam breaking route with an attached disk. Do be aware that either of these approaches is sensitve to dust, oil and chips.

You may want to look into hall effect sensors. Some of these are made secifically for picking up the passing of the teeth of a gear. Others respond to magnetic sources mounted on the shafts. Honeywell makes these as do other sensor manufactures. If you go with a sensor designed to pick up gear teeth then the only other thing you need is a bracket.

The question becomes how fast of a pulse train can something like MACH handle from the spindle? It is concievable that the wrong combination of gear teeth and spindle speed could result in an alias being presented to Mach.

Thanks,

Ron

[Ron111]

While I'm waiting for some components to come in. I starting to investigate good attachments for the spendle speed sensor. Above the gear transmission on the front of the lathe is a approx. 4x6 inch data plate with the gearing selections and speed charts. Behind that cover looks like an ideal location for either type of sensor. It has good access to the shaft and I'm thinking of painting the spindle shaft flat black and then painting a white line if I use the reflective pulse generating system. I'll have to look at mach3 turn because they specify a minimum pulse width that is expected which will have to be calculated at the highest spendle speed of my 9x20, which I think is a little below 2000.

So smart guys, scratch your heads, and see if we can come up with the same width of the white for the minimum pulse width. Mach3 specifies that one pulse can be used on the spendle. If you use more that one, the main pulse has to be wider (I think roughly 50%) that the other pulse(s).

By the way behind the forementioned cover, the area appears to be pretty clean and oit/grease free, as least on my lathe. It you know something that would make this not a wise location, speak now or forever hold your piece.

Well Gents, let me know what you think,

Ron

[Ron111]

Wizard,

I appreciate the input. I use an hall effect sensor on my senor project when I designed and built an electronic speedometer (back in the early 80's when I was finishing my BS). I was strongly considering the approach and then I became concerned about mounting the maganets and balancing issues. Maybe it wouldn't be a problem. What do you think? Infact has anyone on board utilized the pulse processing capabilities using Mach3 turn and has been threading?

Well we'll procede with caution!

Ron

[wizard]

Wizard,

I appreciate the input. I use an hall effect sensor on my senor project when I designed and built an electronic speedometer (back in the early 80's when I was finishing my BS). I was strongly considering the approach and then I became concerned about mounting the maganets and balancing issues.

Hall effect switches come in many forms many of which require a magnet to operate. Fortunately the magnets don't have to be huge and can be embeeded in a disk of some sort. Blancing could be an issue especially at high speed.

There is a class of sensor, which I believe is hall effect, that are designed to capture or count gear teeth. No magnet is needed for these. I know of at least one machine I worked on in the distant past that used such sensors for dynamically triggering events as an axis was moving.

Maybe it wouldn't be a problem. What do you think? Infact has anyone on board utilized the pulse processing capabilities using Mach3 turn and has been threading?

Unfortunately I have no experience with Mach at all. I'd suggest taking their suggestions seriously.

Though I know it has been done I have a hard time accepting that good results can be had from syncing up to a spindle that generates one pulse per revolution. Seems like rock steady spindle speed control would be requred.

Well we'll procede with caution!

Ron

As to your other message about location of the sensor. There was a web posting someplace where a guy did exactly what you suggest. He went behind the cover and mounted the sensor there. Forgot what kind of sensor it was but I believe retro reflective.

It is not a bad location given that you can go a bit futher and design for access. That is put a little door there for maintenance on the sensor and the retro reflective material. Some sensor are more sensitive than others to dust and other accumulations.

As to retro reflective materials be careful here. We use a sensor that did not work well at all with the reflective tape that we initially used. The problem was that the reflectve tape has a crosshatch pattern on it, and the sensor would actually count the cross hatch pattern so instead of getting one count per rev we would get many. This is one of he many reasons why I prefer not to use optical sensors for this sort of thing. They are maintenance intensive.

Dave

[Ron111]

Well guys,

I got the 9X20 going with the cnc. It works well. I had .006 backlash on the z axis and .0015 on the x so I plugged in the backlash offset in Mach3 and all is going well. At this point I have ordered a proximity sensor from automationdirect.com for $20.50 and will begin installation and testing hopefully this weekend. With this in hand and providing spendle speed reference, I will be able to take advantage of the threading wizard in Mach3. (I'm starting to sound like a Mach3 fan)

I need the threading capability in order to cuttdown and thread the ball screws.

Well guys let's start scratching our heads and come up with the best source for ball screws and ball nuts and which technique that I will use to eliminate backlash whether to use a second ballnut or use the ballnuts with the backlash compensation. I not versed in either method so feel free to jump in and bend my ear.

We are in this world together.

Thanks for your inputs,

Ron

[ZipSnipe]

Wheres the PICS?

[HayTay]

Wheres the PICS?

Ditto!!!

I was wandering through Harbor Freight yesterday and checked out the lathe/mill combos. I haven't gotten up enough courage to give one a go, just yet. Plus, I don't have anywhere to put it if I did buy one. For someone without any manual mill/lathe experience, are they worth it or should I look elsewhere?

I managed to make it out of the store with only a couple of packs of cheapo spiral bits and a 4" cross slide vise.

We seriously need some PICS of your HF lathe/mill CNC conversions.

If a lathe/mill CNC conversion has taken place, and there are no pictures to prove it, has the lathe/mill CNC conversion actually been done?


HayTay

[sanddrag]

I was wandering through Harbor Freight yesterday and checked out the lathe/mill combos.

Don't do it. Those things are pretty junky. Save up for independent machines. I would recommend the Habor Freight "8x12". Same exact machine as the LatheMaster 8x14. Good quality. Not usually found in stores though.

[Ron111]

Well Guys,

I going to try to attach at couple of photos. I have very slow dial up service so sometims it works and sometimes it doesn't.

I did simulate the pulse sensor using a momentary switch and pushing it on and off where the proximity sensor would be connected in the circuit and it worked as Mach3 was able to pick up the pulses, so if the sensor works as I think that it does I'm a step closer to cutting threads (and machining the end of a ballscrew).

Bach to ball screws, I saw a couple of ball nuts on ebay which were square and had 4 mounting holes drilled into them. I don't know if the holes went all the way through it or if it was a tapped hole. It would make it easy to mount onto a block behind the apron verticle plate.

Give me some input about different configurations of mounting a ball nut behind the apron. I want to have a space to cover the length of the ball screw. How often are ball screws and ball nuts lubricated?

Take care

Ron

[Ron111]

Just drilled and tapped a 10-32 hole midway in the spindle drive shaft between the Spindle drive pulley and the spindle for the insertion of a 10-32 socket head screw to be used as a pickup for the proximity sensor. I think I figured out how to correctly attach a couple of photos.

Ron

[Alex_Cole]

I just wanted to say that I am very interested in this as I just bought my 9x20 lathe on sale from HF in toledo ohio no more than a month ago. Thanks for the info so far...


Alex

[BobWarfield]

Woohoo, it went together pretty quickly, didn't it? Nice job! :cheers:

Replacing the leadscrew with a ballscrew will be an interesting task. If you can find one on eBay that has the bearing mounting block you're miles ahead IMHO. Decent angular contact bearings are not so easy to come by, while raw ballscrews are very available.

RE lube and mounting, the following page may contain some useful hints:

http://www.rockfordballscrew.com/manual.htm

I'm guessing you are going to want spindle speed control under CNC too at some point. For an inexpensive lathe like this a VFD setup seems like overkill. Find one of the ubiquitous DC treadmill motor setups. I've got one on my 9x30 and it has inputs so I can control it via a knob or a CNC controller.

I'm curious why your first priority on the ballscrew is the saddle, rather than the cross slide? While you did measure more backlash there, my thought would be to focus on the axis that reverses directions (WHILE CUTTING) most often. Seems like when you are profiling a curve the cross slide can go in and out in a way that could be a backlash nightmare. The cutter is typically retracted before the lead changes direction. Backlash compensation can't really help if you reverse directions while cutting, but it can do a decent job otherwise.

In any event, you may wind up replacing both axes and you need to start somewhere. Great project, thanks for the pix.

Best,

BW

[Ron111]

Bob,

Good comments, and an excellent point of I should consider ball screws also on the x axis also. In my world I mainly turn down small pins and approximately 7 in. long M-16 buffer tubes. They have a .5" by .5inch square shape on the bottom to keep the colaspable stock from rotating. So, I have to turn that portion down to make the tube round for use in my system.

The small pins that I turn down, I use .25" dowel pins and I leave .3" at the .25dia and turn the remaining .75" down to .157 to .167", As long as I stay in that range, I'm ok. The pins are used as pull pins for quick releasing retractable pins. I have pain about $16 each for retractable pins, so that why I choose to make them myself.

It does get exciting holding the .3" end in the chuck and machining the .75 off. I think the jig I use gives about .025 clearence between the tool and the chuck. The first couple of passes, I keep one hand on the kill switch, but so far I haven't hit the chuck using my shopmaster.

So in short, I'm not machining any contours, but I should climb out of my focused world and listen to requirement that others would expect in a machine, because I may need a contour and would be disappointed it if didn't turn out too well. So, thanks for the input, Bob.

I had visited your home page, and you do first rate work!

Well guys, when I last left you, I had just connected the spendle Proximity sensor and got good spendle speed measurements. Today I made the cables for I/O out of the control box and for connection on the machine, and cut mt first threads. It worked very well using the Mach3 wizard for simple threading. So I attached a picture.

Let's take a moment and savor this event, This is the first thread that I have ever in my life (with a lathe manual or CNC). It's also as momentous as a kid riding a bicycle for the first time. (maybe I'm taking this a little far) but anyway, I attached a picture.

I did buy a $20 Targus laptop keypad and velcro it the the front of the lathe for local jog control and programmed a couple of keys (using keygrabber (a utility in Mach3))to give me a shift function and a pause function along with using the arrow keys for jogging. So I stop the program, quickly if the cutting tool is getting to friendly with the chuck.

So, we made good progress today. Some give me a good source for a preloaded 5/8 ball nut, and a couple of options of mounting behind the saddle. And where is a good place to get a 15/16 X 16tpi tap (I believe that is the standard thread) for mounting a ball screw. I understand that you can use two back to back ball screws for anti backlash, if anyone has a couple of pic's and a brief description, it would be helpful.

Well, that's all for now folks!!!

Take care,

Ron

[Trapper14]

congratulations on your first thread ever!! looks decent enough!!

[Ron111]

Well guys,

I just received the 5/8" ballscrew in from McMaster Car ($1.25 an inch ($45 for a 36 inch screw) and $23.45 for each ball nut). I'm going to venture out in a little while and begin to turn down the ball screw end. These screws are hardened to around a rockwell hardness of 60, so we will see how it goes. I'm still thinking of an arrangement for arranging a spring between the ballnuts for the preload. Any one with some good suggestions please feel free to speak up.

I have run a couple of trial runs using 36" inch 5/8" unhardened steel rod fed through the rear of the spindle and didn't have any problems with whipping as I was cutting at 150 rpm. I must say the the lathe so far has preformed very well on the acme screw, but the whole purpose up to this point was to be at a point where I can turn down and and thread the end of the ball screw. Then I will machine the bearing block mounts and install the angular bearings, and I will be ready to start designing the ball nut mount and the preloading arrangement.

Later,

Ron

[Runner4404spd]

the easiest way i found to spring load these that takes up the least amount or room is

1. create a bracket that will bolt to the carriage to mount your ball nut. then make a ring with 8 holes evenly spaced at 45° apart. mount one ballnut to the mount.

2. mount the ring onto the other ball nut. place a spacer and a set of spring washers between the ballnuts. now install two screws through the ring with holes in it and tighten it down to the bracket.

3. this will allow you to adjust your preload and it works pretty well. i don't know if i have any pics of the setup but i will try and post a drawing later.