Comments on: A new spindle for our Opti CNC mill http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/ Wed, 07 Jan 2009 11:41:04 +0000 http://wordpress.org/?v=2.6.3 By: Administrator http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-8777 Administrator Fri, 30 Mar 2007 10:28:56 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-8777 I think the measured run-out with our basic cheap beatings is around 0,01 mm to 0,015 mm I don't know if using higher grade bearings would help. It could also turn out that the cheap littlemachineshop spindle is not true to better than about 0,01 mm and using expensive beatings doesn't help. Do let me know about the results if you try the littlemachineshop spindle with abec-7 bearings or tapered roller bearings. The spindle (including the spindle box) is aligned to be perpendicular to the table using very thin spacers between the spindle box and the z-plate that is attached to the rails/ballnut. We haven't measured accurately yet how much the z-rails and the whole column is skewed relative to the true z-axis. In principle this skewness can be compensated for with tilted kinematics in EMC. Column tilt will only be an issue if you use tools of very different lengths, or if the part you are making is very tall in the z-direction. the motor is connected to the spindle with a 1:1 toothbelt, it's shown here: http://www.anderswallin.net/2006/12/new-spindle-installed/ the pro machines build the motor itself around the spindle, but I don't know how that could be done on a hobby machine... AW I think the measured run-out with our basic cheap beatings is around 0,01 mm to 0,015 mm

I don’t know if using higher grade bearings would help. It could also turn out that the cheap littlemachineshop spindle is not true to better than about 0,01 mm and using expensive beatings doesn’t help.

Do let me know about the results if you try the littlemachineshop spindle with abec-7 bearings or tapered roller bearings.

The spindle (including the spindle box) is aligned to be perpendicular to the table using very thin spacers between the spindle box and the z-plate that is attached to the rails/ballnut.

We haven’t measured accurately yet how much the z-rails and the whole column is skewed relative to the true z-axis. In principle this skewness can be compensated for with tilted kinematics in EMC. Column tilt will only be an issue if you use tools of very different lengths, or if the part you are making is very tall in the z-direction.

the motor is connected to the spindle with a 1:1 toothbelt, it’s shown here:
http://www.anderswallin.net/2006/12/new-spindle-installed/
the pro machines build the motor itself around the spindle, but I don’t know how that could be done on a hobby machine…

AW

]]> By: Cristian http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-8225 Cristian Wed, 28 Mar 2007 13:33:58 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-8225 Hi Anders, I was looking at the spindle on littlemachineshop and was wondering - how do you guarantee the Z-axis movement to be perpendicular to the table ( that is, is paralel to the gravity ? ) Any design on how to mount the spindle to a motor shaft? What about using very high precision ball bearings, such as ABEC-7 ( P4 european standard ) ( although very expensive... ) ... I noticed on littlemachineshop they have tapered roller bearings... aren't those bearings better than yours? Cristi Hi Anders,

I was looking at the spindle on littlemachineshop and was wondering - how do you guarantee the Z-axis movement to be perpendicular to the table ( that is, is paralel to the gravity ? )
Any design on how to mount the spindle to a motor shaft?
What about using very high precision ball bearings, such as ABEC-7 ( P4 european standard ) ( although very expensive… ) …
I noticed on littlemachineshop they have tapered roller bearings… aren’t those bearings better than yours?

Cristi

]]> By: Anders http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5367 Anders Mon, 12 Mar 2007 12:02:03 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5367 we have not measured the backlash of the ball-nut. It has an adjustment screw which adjust how tightly it fits over the screw. With a high tightness I imagine the backlash can be made very small, but with the disadvantage of high wear and more friction. I'll write something more about this when we get the machine assembled and working. For the rails, there is also no play or backlash that you can feel by hand. AW we have not measured the backlash of the ball-nut. It has an adjustment screw which adjust how tightly it fits over the screw. With a high tightness I imagine the backlash can be made very small, but with the disadvantage of high wear and more friction. I’ll write something more about this when we get the machine assembled and working.
For the rails, there is also no play or backlash that you can feel by hand.

AW

]]> By: Cristian http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5365 Cristian Mon, 12 Mar 2007 08:32:18 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5365 Hi Anders, just wanted to ask if you have any idea about the backlash in a 16*2.5 ball screw? As you said about the linear guide ways, I thought you're probably right and decided to use the MGW15, especially as I am concerned about precision. Is there any play in the railways, on transversal axis? If yes, how much? Hi Anders,

just wanted to ask if you have any idea about the backlash in a 16*2.5 ball screw?
As you said about the linear guide ways, I thought you’re probably right and decided to use the MGW15, especially as I am concerned about precision. Is there any play in the railways, on transversal axis? If yes, how much?

]]> By: Administrator http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5287 Administrator Fri, 09 Mar 2007 06:40:24 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5287 if you have a 420mm plate fixed to a 140mm block then that does not sound very sturdy. if you will be cutting wood, plastics, or light aluminium then maybe it will work. If your cut's are heavier then I would suggest four blocks placed in the corners of the plate. It does require longer and thus more expensive rails. Have you estimated your cutting forces ? As stated above, for a bench-top hobby machine I believe the 15mm rails will do just fine. A professional machine like a Haas Minimill or Toolroom Mill would use something like 35x35mm rails. I found the knuth.de picture I was thinking about. It's on the second page of this pdf: http://www.knuth.de/pdf_2007_russisch/Seite_052.PDF That design seems to be pretty generic for VMCs, so maybe worth studying. Anders if you have a 420mm plate fixed to a 140mm block then that does not sound very sturdy. if you will be cutting wood, plastics, or light aluminium then maybe it will work. If your cut’s are heavier then I would suggest four blocks placed in the corners of the plate. It does require longer and thus more expensive rails.

Have you estimated your cutting forces ? As stated above, for a bench-top hobby machine I believe the 15mm rails will do just fine. A professional machine like a Haas Minimill or Toolroom Mill would use something like 35×35mm rails.

I found the knuth.de picture I was thinking about. It’s on the second page of this pdf: http://www.knuth.de/pdf_2007_russisch/Seite_052.PDF
That design seems to be pretty generic for VMCs, so maybe worth studying.

Anders

]]> By: Cristian http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5270 Cristian Thu, 08 Mar 2007 08:58:23 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5270 Hello, one more question about linear bearings. I was thinking about using HGH35HA, one per rail instead of using 2 HGH15HC per rail. The board on the rail is 42 * 42 cm, and HGH35HA is 14 cm long. It would be placed bellow the board, 2 rails, in the middle of the right and left rail. Would this be better instead of using 4 HGH15HC, 2 on each rail ? Hello,

one more question about linear bearings. I was thinking about using HGH35HA, one per rail instead of using 2 HGH15HC per rail. The board on the rail is 42 * 42 cm, and HGH35HA is 14 cm long. It would be placed bellow the board, 2 rails, in the middle of the right and left rail. Would this be better instead of using 4 HGH15HC, 2 on each rail ?

]]> By: Administrator http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5233 Administrator Mon, 05 Mar 2007 22:32:32 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5233 15mm rails: the rail size you choose will depend on the expected loads. I don't remember the exact specs on the 15mm rails, but if you take a look at them I would think the reasonable forces are well beyond what you can expect for a small mill (i.e. a 10mm end-mill will brake long before the rails will give up) From a purely accuracy point of view I would also think that smaller dimension rails are actually better ? We use two bearing blocks per rail, I don't think any more would help. For a very precise machine I think the key with the rails and ballscrews is to get a machine with good repeatability (zero backlash and play in all directions). Then if there are systematic errors they can be measured and corrected for. Some precise machines use linear glass-scales as a secondary encoder for this purpose. Others just do an interferometric scan when the machine is built and program the control to correct for any nonlinearities or skew in the axis orientations. design: most commercial vertical milling machines use a conventional design where the spindle moves with the z axis, and the table moves in y and x directions. If you want a really stiff machine you use a 'carriage' design for the table (see for example the knuth.de site for a good picture) Yes, we are fitting the 15mm linear rails to the BF20. I would not advertise this for anyone else as being a simple exercise. They do fit if you think about how you want to fit them and the work on the implementation a bit, but it's not a machine that is designed for linear rails. We are not using the stock BF20 spindle or table, we have built our own. Anders 15mm rails:
the rail size you choose will depend on the expected loads. I don’t remember the exact specs on the 15mm rails, but if you take a look at them I would think the reasonable forces are well beyond what you can expect for a small mill (i.e. a 10mm end-mill will brake long before the rails will give up)
From a purely accuracy point of view I would also think that smaller dimension rails are actually better ?

We use two bearing blocks per rail, I don’t think any more would help.

For a very precise machine I think the key with the rails and ballscrews is to get a machine with good repeatability (zero backlash and play in all directions). Then if there are systematic errors they can be measured and corrected for. Some precise machines use linear glass-scales as a secondary encoder for this purpose. Others just do an interferometric scan when the machine is built and program the control to correct for any nonlinearities or skew in the axis orientations.

design:
most commercial vertical milling machines use a conventional design where the spindle moves with the z axis, and the table moves in y and x directions. If you want a really stiff machine you use a ‘carriage’ design for the table (see for example the knuth.de site for a good picture)

Yes, we are fitting the 15mm linear rails to the BF20. I would not advertise this for anyone else as being a simple exercise. They do fit if you think about how you want to fit them and the work on the implementation a bit, but it’s not a machine that is designed for linear rails. We are not using the stock BF20 spindle or table, we have built our own.

Anders

]]> By: Cristian http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5201 Cristian Mon, 05 Mar 2007 09:14:29 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5201 And one question more - do you plan on using the linear bearings on the BF 20 vario? Is it possible to convert the dovetail table and attach linear bearings to it? And one question more - do you plan on using the linear bearings on the BF 20 vario?
Is it possible to convert the dovetail table and attach linear bearings to it?

]]> By: Cristian http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5199 Cristian Mon, 05 Mar 2007 08:54:22 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-5199 Hello Anders, guys, I only did some investigation about milling machines and decided to build my own since I admit I am obsessed with precision and I will certainly go for ball screws and linear bearings. I took a look on metallstore and found those ballscrews, 16*2.5 to 25*20 , they are all available. I took a look also at your CNC machine and noticed you' re using 15 mm guideways, and wanted to ask - isn't this too little? I was planning to use 25mm or even 30 mm since I intended to take the same movement approach as the BF 20 vario: X axis table ( down ) carries the Y axis table ( on it ) and the Z axis is executed vertically, carying the milling head. What sizes do you think are fit for normal milling ( also in steel ) and how many guideways per rail are fit? Moreover, I intend to use only steel plates since I will need some times to machine steel pieces. A different approach is to have the Y axis down and the X and Z mounts to be vertically and move the milling head. Is this better? Hello Anders, guys,

I only did some investigation about milling machines and decided to build my own since I admit I am obsessed with precision and I will certainly go for ball screws and linear bearings. I took a look on metallstore and found those ballscrews, 16*2.5 to 25*20 , they are all available.
I took a look also at your CNC machine and noticed you’ re using 15 mm guideways, and wanted to ask - isn’t this too little? I was planning to use 25mm or even 30 mm since I intended to take the same movement approach as the BF 20 vario: X axis table ( down ) carries the Y axis table ( on it ) and the Z axis is executed vertically, carying the milling head. What sizes do you think are fit for normal milling ( also in steel ) and how many guideways per rail are fit?
Moreover, I intend to use only steel plates since I will need some times to machine steel pieces.
A different approach is to have the Y axis down and the X and Z mounts to be vertically and move the milling head. Is this better?

]]> By: Administrator http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-4935 Administrator Wed, 28 Feb 2007 19:04:58 +0000 http://www.anderswallin.net/2006/10/a-new-spindle-for-our-opti-cnc-mill/#comment-4935 Hi Christian, For cnc-use ballscrews will be much better than the normal threaded screws that come with the Opti. Our ballscrews are 16x2.5mm screws by HiWin and they are bough from http://metallstore.de/ For the ways, some machines use dovetail ways successfully, but we decided to replace them with linear ways. Our guideways are made by IKO, 15mm wide, and bought on ebay. HiWin makes similar ways, available also from metallstore.de When our machine was running with steppers and the stock screws, if you are careful about the setup and know what you are doing you can probably reach 0,05 mm accuracy for the parts (or a little better). With ballscrews we are hoping that number will be closer to 0,01 mm I don't think it's possible to buy the column separately. Opti does sell the cross-tables separately, and if you inspect the models you will see that some of the tables sold separately are exactly the ones used on the Opti mills. If you are serious about building stuff yourself, and can find someplace with a big mill where the base of your machine could be face-milled, then custom building the base is also an option. Check out the forum om cnc-zone for inspiration. good luck, and do send me some pictures if you build your own machine! Anders Hi Christian,

For cnc-use ballscrews will be much better than the normal threaded screws that come with the Opti. Our ballscrews are 16×2.5mm screws by HiWin and they are bough from http://metallstore.de/
For the ways, some machines use dovetail ways successfully, but we decided to replace them with linear ways. Our guideways are made by IKO, 15mm wide, and bought on ebay. HiWin makes similar ways, available also from metallstore.de

When our machine was running with steppers and the stock screws, if you are careful about the setup and know what you are doing you can probably reach 0,05 mm accuracy for the parts (or a little better).
With ballscrews we are hoping that number will be closer to 0,01 mm

I don’t think it’s possible to buy the column separately. Opti does sell the cross-tables separately, and if you inspect the models you will see that some of the tables sold separately are exactly the ones used on the Opti mills.

If you are serious about building stuff yourself, and can find someplace with a big mill where the base of your machine could be face-milled, then custom building the base is also an option. Check out the forum om cnc-zone for inspiration.

good luck, and do send me some pictures if you build your own machine!

Anders

]]>