CNC XY Table for clock making no Z axis

[Stephen BensonParticipant @stephenbenson75261]

[Stephen BensonParticipant @stephenbenson75261]

I have been wanting to make a robust CNC XY table to fit on top of my mill table with a 150x120mm or so area for making parts for clocks in up to 1/4 brass sheet. The Z axis would be manual using the mills own spindle.

I would want this for short production runs of several hours and for it to be accurate to about 0.10mm so I could be in commercial machine territory but I have no room and maintenance costs scare me. I do not want to reinvent the wheel but I have not seen any plans that fit the bill anybody got any ideas?

Edited By Stephen Benson on 18/05/2011 20:15:55

Edited By Stephen Benson on 18/05/2011 20:16:37

[Tony JeffreeParticipant @tonyjeffree56510]

There are plenty of X-Y tables around, available from the usual suspects. For example, Arc Eurotrade (catalogue distributed with this month’s MEW) has 3 sizes of table on page 43 of the catalogue; he smallest has a table 200mmX90mm and a table travel of 130mmX70mm. These are manual tables, but would be simple enough to convert to stepper drive if you so desired, using the existing leadscrews.

 

Regards,

Tony

[John Stevenson 1Participant @johnstevenson1]

Arc has an example of the small table converted to stepper drive with two of the tiny type 16 motors fitted.

 

I did this about two years ago as an exercise and it was just a bolt on, no modifications were made to the table at all.

 

Image dated July 2nd 2010

 

John S.

Edited By John Stevenson on 18/05/2011 23:16:32

[Stephen BensonParticipant @stephenbenson75261]

Thanks for the replys but is using a manual machine table the way to go, I would of thought that it was a huge compromise due to the backlash, weight and drag issues.

 

What I had in mind was a ball bearing lead screw with linear bearings to hold may be 8-10mm thick flat aluminium table sorry for the misunderstanding.

But if converting a manual machine is considered the correct way to go how do you over come the above problems.

[Tony JeffreeParticipant @tonyjeffree56510]

Posted by Stephen Benson on 19/05/2011 08:42:03:

Thanks for the replys but is using a manual machine table the way to go, I would of thought that it was a huge compromise due to the backlash, weight and drag issues.

 

What I had in mind was a ball bearing lead screw with linear bearings to hold may be 8-10mm thick flat aluminium table sorry for the misunderstanding.

But if converting a manual machine is considered the correct way to go how do you over come the above problems.

Whatever way you go, there are compromises – your proposed direction of linear bearings and aluminium table will be a compromise on stiffness compared with a starting point of a cast iron X-Y table, for example. Also, while ballscrews have advantages (low rate of wear and low friction, for example) eliminating backlash isn’t necessarily one of them, unless you pay more for anti-backlash nuts. There are ways of reducing backlash in conventional screws, and often a significant part of the backlash can be eliminated by fitting decent bearings on the leadscrews, especially if the original ones were plain bearings.

 

Regards,

Tony

[Stephen BensonParticipant @stephenbenson75261]

I have been using industrial CNC machines most my working life and the two runner with centre ball screw model seems to be common to most the industrial machines I have experienced so I suppose I expected to see it in the hobby machines as well. Certainly I do not see stiffness to be an issue as stiffness is more about design rather than mass especially in the small table size (150x120mm) I propose.

[Tony JeffreeParticipant @tonyjeffree56510]

Posted by Stephen Benson on 19/05/2011 17:15:34:
I have been using industrial CNC machines most my working life and the two runner with centre ball screw model seems to be common to most the industrial machines I have experienced so I suppose I expected to see it in the hobby machines as well. Certainly I do not see stiffness to be an issue as stiffness is more about design rather than mass especially in the small table size (150x120mm) I propose.

OK…but why go to the bother when you can start from a ready-made X-Y table?

 

Regards,

Tony

[PatParticipant @pat]

Hi Stephen

 

Can we assume that 0.10mm means that you are looking to maintain positional accuracy across the table envelope of + to – 0.005 mm or better? The problem of part interchangeability and co-ordinate tolerance needs to be factored into any production run unless you are prepared for selective assembly of parts and some scrap parts.

 

You will need to think very carefully about the above tolerance requirements as you could quite easily get into the area of servos that use linear scales as feed back elements. This would vastly increase the cost of the table. For clock making the co-ordinate tolerances are relatively relaxed as the gear centers are run with a very loose mesh as the need is for minimal friction. Backlash in the gearing is not an issue also the speeds are very low so noise produced by a loose mesh is not relevant. Clock makers also use gear tooth forms that are designed for minimal friction unlike engineering tooth forms that are designed to provide smooth rotational transmission of torque, low back-lash and quiet running.  Make a clock with this sort of gearing and it will require too much power to be viable.

 

The simple table pointed out by John Stevenson would probably achieve all that you need for both clock plates and gear wheels. The actual weight of that table would help with getting the required fine finish on the gear tooth profile with out too much polishing.

 

The backlash is simply controlled by approaching any critical dimension from the same direction each time. Easy to do the CNC as it will repeat the direction each and every time in stead of getting fed up with the extra handle twirling! There would remain the problem of lead screw wear and this is one reason to convert to ball-screw drive. The dove-tail wear should be relatively low provided you use a lubricant that is designed for slide-ways. This oil is a specialty oil designed for minimal stick slip and the constant reversal in direction of travel. If the production run is long then the linear bearing skates and matching rails would be required. These are expensive and result in a loss of usable ‘Z’ axis travel. The round rod type of linear slides unless very carefully designed could give you rigidity issues due to the construction of the insert bearing.

 

Suggest you fix the co-ordinate tolerance you need then look at the accuracy obtainable from the various table axis drives then look at how much each design costs. Then revise to get a good working solution. I would go for the simple table with dove-tail slides and stepper motors as I think this will give you the repeatability required for clock making. Convert to ball-screw if the lead screws wear out and modify the CNC tool paths so that each critical dimension is approached from the same direction each time.

 

Hope this helps to get you cutting brass – Regards – Pat

 

 

 

 

 

Edited By Pat on 20/05/2011 12:24:03

[PatParticipant @pat]

Hi Stephen

 

To give you and idea on pricing you may care to look at this site:-

http://www.slidesandballscrews.com/hgr25-profile-rail-p-404.html?cPath=40_120_122

Zapp Automation is a source of precision linear slides and trucks by Hinwin. Cost is around £600 for 400mm ways and eight trucks. They also have the servos but think in terms of £300 for the DRO and another £500 plus for two servos and £300 for ball screws and B12 bearings.

They also do the cheaper linear ball bearing type but these are more appropriate for a light weight mechanism. If you reduce the feed rates tool pressure you might be able to use the light weight liner round rail but the accuracy will be difficult to determine as information of the rigidity is difficult to come by as these bearings are not intended for the precision table application.

 

Regards – Pat

 

[Stephen BensonParticipant @stephenbenson75261]

Thanks everybody some really great information and I am very grateful I am almost convinced but I have a serous paradigm to break but I suppose I have not lost too much if I have to upgrade the table at some point.

So John if you could give a little more info on your table conversion I would be very grateful.

Cheers Steve

 

[Author]

Posts