Lathe conversion from stock to CNC

More
16 Jul 2012 15:17 #22003 by Soundreflections
Hi,

I am looking to upgrade my lathe (BV20B-L) from stock to CNC. It has always been in the back of my mind to do, though recently my motor decided it is no longer going forward, it is still happy with reverse though. So I upgraded to a 3 phase motor, with VFD, which is one of the first steps to CNC, as it can now easily be controlled by a computer, I can also run it from 20Hz, to 100Hz, instead of just 50Hz, so I am looking forward to some flexibility there. I must just get a pulley for the new motor, ten I should again have a working lathe!

I figured DRO should be my next stage, as I need some form of DRO for manual work and one needs DRO for CNC. I was looking at www.shumatech.com/web/products which is open source based. Starting to investigate what I had considered the final stage, the CNC processor, I find LinuxCNC, which seems to have it all! I would like to ask a few questions, which are probably silly, newbie questions, but I have not yet found the info I am looking for on the site, or forum. Please refer me to a source if my questions have all been answered. I am an electronic engineer, so the workings I can understand, I am new to the mechanics & drive train sections, so may yet kick myself for some ideas / questions in time to come.

I am planning to have manual / auto control, so that I can make something by hand if I want (often for simple pieces I can see doing a CAM file for it could be tedious), I also enjoy hand turning at times. Can LinuxCNC use steppers as encoders, if they are not driven? I would need a DRO in manual still. Otherwise can one interface scales to LinuxCNC?

Stepper / servo? What are the implications of the route one chooses? What is the most common in use, and why?

My lathe has a power feed for thread cutting, how accurate is servo based cutting? Would I be best off to motorize the lead screw and not have drive gears, or use the gears when thread cutting and have a servo drive for normal cutting?

These are the main questions I can think of now, ironically, as I have an old computer lying around and I have the live CD , I in theory have all the processing require, just no motors. I can not at this stage see an advantage to having a CNC PC just controlling spindle speed!

Thanks in advance for your inputs
Regards
Peet

Please Log in or Create an account to join the conversation.

More
16 Jul 2012 16:14 #22007 by andypugh
Soundreflections wrote:

I am looking to upgrade my lathe (BV20B-L) from stock to CNC

I have converted a similar style of lathe. Perhaps you can learn from my mistakes.
The first mistake is, possibly, to start with such a lathe. They are not that great, and you may find the limitations of the underlying machine.
Some pictures: Firstly, a modification I made to the saddle keeper strips, and a few shots of the machine itself:
picasaweb.google.com/1081645046564043805...er=0&feat=directlink
Rather "dilute" pictures of it amongst other stuff, including limit-switches and spindle encoder (look for the dark blue colour)
picasaweb.google.com/1081645046564043805...er=0&feat=directlink
A CNC-zone thread on converting the X axis to ballscrew:
www.cnczone.com/forums/mini_lathe/63621-...ew-2.html#post509784

I must just get a pulley for the new motor, ten I should again have a working lathe!

You probably don't need any of the original spindle drivetrain. I converted to a single run of dual-V belts. Probably overkill considering that the original belts last several hours at a time.

I am planning to have manual / auto control, so that I can make something by hand if I want (often for simple pieces I can see doing a CAM file for it could be tedious), I also enjoy hand turning at times. Can LinuxCNC use steppers as encoders, if they are not driven?

In theory steppers can be used as encoders, I don't know how well it works. Scales would be easier, or you could fit actual encoders to the the stepper motors.
However, I find I prefer to use my lathe in a semi-auto mode, where I select start and finish Z positions, feed rate and surface speed, and press a button marked "Face" or "Bore" etc.
JT works the same way, and uses ngcgui, I wrote my own version a while back. Ngcgui is probably the better solution. I have half a plan to change my version to Glade, with pictures in tabs.
www.bodgesoc.org/lathe/lathe.html
My version runs G-code routines with parameter inputs. I think ngcgui creates fresh G-code each time. (but I could be wrong). Both work much the same way in practice.

Stepper / servo? What are the implications of the route one chooses? What is the most common in use, and why?

Steppers are much cheaper and more straightforward than servos. Steppers are more common, for that reason. In your situation I would definitely go for steppers unless you find some very cheap servos. I think a full-price servo setup on your base machine would be a waste. Servos would make true-manual operation easier, having encoders and no unpowered cogging, but I doubt you will actually find yourself doing any manual work. I left myself the option, and never have.

My lathe has a power feed for thread cutting, how accurate is servo based cutting? Would I be best off to motorize the lead screw and not have drive gears, or use the gears when thread cutting and have a servo drive for normal cutting?

Servo/stepper thread cutting is great. You can choose any pitch you want, for a start. And the machine can pull the tool out and reset every time. Example:

Please Log in or Create an account to join the conversation.

More
16 Jul 2012 16:46 #22013 by BigJohnT
Soundreflections wrote:

Hi,

I am looking to upgrade my lathe (BV20B-L) from stock to CNC. It has always been in the back of my mind to do, though recently my motor decided it is no longer going forward, it is still happy with reverse though. So I upgraded to a 3 phase motor, with VFD, which is one of the first steps to CNC, as it can now easily be controlled by a computer, I can also run it from 20Hz, to 100Hz, instead of just 50Hz, so I am looking forward to some flexibility there. I must just get a pulley for the new motor, then I should again have a working lathe!

I figured DRO should be my next stage, as I need some form of DRO for manual work and one needs DRO for CNC. I was looking at www.shumatech.com/web/products which is open source based. Starting to investigate what I had considered the final stage, the CNC processor, I find LinuxCNC, which seems to have it all! I would like to ask a few questions, which are probably silly, newbie questions, but I have not yet found the info I am looking for on the site, or forum. Please refer me to a source if my questions have all been answered. I am an electronic engineer, so the workings I can understand, I am new to the mechanics & drive train sections, so may yet kick myself for some ideas / questions in time to come.


I have two shumatech DRO's one for my manual lathe and one for my manual mill. I never did get around to installing the one on the lathe. The one on the manual mill works fine but doesn't have any application in the CNC world.

I am planning to have manual / auto control, so that I can make something by hand if I want (often for simple pieces I can see doing a CAM file for it could be tedious), I also enjoy hand turning at times. Can LinuxCNC use steppers as encoders, if they are not driven? I would need a DRO in manual still. Otherwise can one interface scales to LinuxCNC?


Far better than CNC/manual is to install a MPG which you can configure to cut angles as well as in the X and Z planes. Also the MPG is much better way to touch off your tooling and material.

Stepper / servo? What are the implications of the route one chooses? What is the most common in use, and why?

My lathe has a power feed for thread cutting, how accurate is servo based cutting? Would I be best off to motorize the lead screw and not have drive gears, or use the gears when thread cutting and have a servo drive for normal cutting?


The accuracy depends more on your mechanics than the controls and drives and such. Almost 0 backlash ball screws will be the main component to get accuracy.

These are the main questions I can think of now, ironically, as I have an old computer lying around and I have the live CD , I in theory have all the processing require, just no motors. I can not at this stage see an advantage to having a CNC PC just controlling spindle speed!

Thanks in advance for your inputs
Regards
Peet


My BP mill is in that state with my D510 computer and LinuxCNC just controlling my spindle and the Anilam controls still running the servos. If you were lucky enough to purchase an AutomationDirect VS2 drive LinuxCNC has a built in driver for it via modbus. In any case soon to be converted...

And like Andy I share the same feelings about converting a manual lathe to CNC... I'm temped to build a strap on CNC X and Z axis for my Samson manual lathe but have resisted so far.

John

Please Log in or Create an account to join the conversation.

More
16 Jul 2012 17:28 #22016 by Soundreflections
Thanks for the replies,
I often tackle these projects out of a technical interest more than having a production type requirement. (Women can often not understand these things, even when one points out the potential one will have, once everything is done). I have this lathe, no more space for a bigger lathe (I technically don't have space for this one, but space is relative). I certainly don't have a budget for a better lathe, with better tolerances. I agree that it is a bit of a shame to fit CNC to such a machine, but if I can do it on low budget, try to tighten tolerances, etc... I can take that experience (and possibly equipment) forward, when I do get a new, improved lathe, I can then also more easily make the decision to look for CNC, or the best non CNC and convert.

Somewhere I need to get a mill, but I will need more space and budget for that, standing between the metal and the wood lathes is one thing, I cannot see my dear lady accepting a mill in the living room, no matter how nicely I decorate it!

Please Log in or Create an account to join the conversation.

More
16 Jul 2012 18:47 - 16 Jul 2012 18:48 #22021 by BigJohnT
Consider making an XYZ attachment that clamps to your ways for your lathe and use the spindle for your tools, might even make it convertible to XZ and a tool holder for turning ops. That way you would end up with three machines in the footprint of one.

John
Last edit: 16 Jul 2012 18:48 by BigJohnT.

Please Log in or Create an account to join the conversation.

More
16 Jul 2012 19:13 #22028 by Soundreflections
BigJohn, thanks for that idea! I can get quite exited thinking about that, though I probably am being a bit limited in my understanding of machining. I see this as basically putting a QCTP on, then clamping the workpiece in the QCTP, and the tool in the chuck? That gives me basic mill & lathe. I have often wondered about a way to have a indexable table mounted vertically, or something like that.
Can you perhaps elaborate? I would really appreciate input / ideas. Presumably I can have different configs in LinuxCNC to select my tool setup?
Thanks
Peet

Please Log in or Create an account to join the conversation.

More
16 Jul 2012 22:30 #22042 by BigJohnT
Lots of machines run dual configs, it's just a different file you load. Andy has one machine like that.

The lathe conversion would be the simple one to make with some linear rails and plates to clamp onto the existing ways of the lathe. Of course careful design would be needed to limit the amount of height lost to the XZ table.

The mill conversion would involve adding a Y axis (that would move up and down in your perspective) with some kind of work holding device on it.

Things to ponder...

John

Please Log in or Create an account to join the conversation.

More
17 Jul 2012 13:45 #22077 by andypugh
BigJohnT wrote:

The mill conversion would involve adding a Y axis (that would move up and down in your perspective) with some kind of work holding device on t


Mount it on a counterbalance?

www.lathes.co.uk/meyerburger/

(Not an entirely serious suggestion)

Please Log in or Create an account to join the conversation.

More
17 Jul 2012 15:12 #22089 by Soundreflections
Thanks for the ideas and inputs. Some of those machines look quite interesting!

Firstly a CNC question: Is it more practical / common to motorise the "apron" drive (can't remember my terms here), or the lead screw. I take it my X2 need not be motorised, just locked. Looking at my lathe it would be a bit interesting, but quite do-able to modify it for motors, would especially help if I can rig a mill mode for it!

I hope to have added a picture to this post, of my cross slide. To add a "mill" table I have had an interesting thought, to which I would appreciate some comment.

If I have a form of channel bolted on where the follow rest should be (never got one), I then have a vertical "table", with a male form of the channel, with a hinging mid point that can attach to the tool holder (I am planning to go for a QCTP somewhere soon). I can then use my X2 axis to set the angle of the "table" and still have Z axis travel with the tool post being a fixed point and the "table" sliding in the channel. This is quick thinking and I can immediately see a few snags however I would appreciate some comments and maybe this is something that can become a workable option.

Thanks
Peet
Attachments:

Please Log in or Create an account to join the conversation.

More
17 Jul 2012 15:36 #22093 by andypugh
Soundreflections wrote:

Is it more practical / common to motorise the "apron" drive (can't remember my terms here), or the lead screw.


Typically you add two motors, one to control the cross-slide and one to drive the leadscrew, generally changing the leadscrew to a ballscrew.
The compound slide can be eliminated and replaced by a solid block, it has no function and adds flexibility.

I removed the entire apron and replaced it with a mount for a ballnut and a bracket/drive for the cross-slide motor.

If I was doing it again, I would do it a little differently; I would try to keep the ballscrew much nearer the bed than mine ended up. In fact I might even design and make a complete new saddle casting, with more room for a cross-slide ballscrew. And a new cross-slide, incorporating a toolchanger. Hmm, that would be serious feature-creep.

Currently my Z-drive looks like this: picasaweb.google.com/lh/photo/20UsgDxjZ7...pFm0?feat=directlink
If I was doing it again, the mounting bracket would bolt across the end of the bed, the motor would point the other way, and the drive pulley would be "hooked" round the back of the bed to get the ballscrew tighter in to the bed.
One gain from this would be adequate room for spiral-spring or telescopic leadscrew protection.

At the saddle: picasaweb.google.com/lh/photo/SewmOYCpUT...pFm0?feat=directlink I made things a bit difficult with my adjustable keeper plates, but a Mk2 design would step the ball-nut down and in.

Please Log in or Create an account to join the conversation.

Moderators: piasdom
Time to create page: 0.406 seconds
Powered by Kunena Forum