PDA

View Full Version : BUILD LOG: Hybrid Mill Design



Chaz
29-11-2015, 05:04 PM
Hi,

So I would like to get a larger mill that is designed mainly for cutting Aluminium. This means high speed spindle and fairly quick feed rates.

Heavily inspired by this design (http://uk.dmgmori.com/blob/172034/d7bd0ec76e8ecc8c4b672d1fd67df02a/pm0uk13-nmv5080-pdf-data.pdf), see pages 4 to 7, I came up with the following design. This is also very similar to the Ultimaker 3D printer design where you have a 'box' and you keep all the movement on top and place everything centrally over the table.

The design is drawn using 1605 Leadscrews (C5 grade) and 400W Servo Motors (4, 5 or 6) - Cant decide if Z and Y need 2nd screws or not. Rails are Hiwin 20 (for Z) and 25 for X and Y.

The spindle will be a 4KW Chinese Watercooled spindle.

Cutting size based on the design is around X and Y - 600 mm x 600 mm x 250 mm.

For the frame, I want to look at using something 'interesting' like Epoxy Granite or something similar. For any parts that are needed, these will be done out of either steel or aluminium.

Thoughts / critique welcome. Thanks.

https://c2.staticflickr.com/6/5821/23398593945_6194aec56e_b.jpghttps://c1.staticflickr.com/1/675/23290135082_b0b6f48fe6_b.jpghttps://c1.staticflickr.com/1/664/22770221434_e173337f8f_b.jpghttps://c1.staticflickr.com/1/653/22771424983_f5a11a6afe_b.jpghttps://c1.staticflickr.com/1/782/23102818360_5fdb24e993_b.jpghttps://c1.staticflickr.com/1/622/23316024731_b0b0076430_b.jpg

https://c2.staticflickr.com/6/5766/23290136872_c94ee7e6e1_b.jpg

Neale
29-11-2015, 05:24 PM
Chaz - can't help with any of your questions but I've seen that you are using Fusion 360 for this design. I'm using the same thing (and gradually getting to grips with it) although I haven't looked at any fancy renderings yet! However, can you give a pointer to a source of the Hiwin carriages and rails? Are these from Hiwin, and what drawing format works for you?

Thanks,

Chaz
29-11-2015, 05:43 PM
Chaz - can't help with any of your questions but I've seen that you are using Fusion 360 for this design. I'm using the same thing (and gradually getting to grips with it) although I haven't looked at any fancy renderings yet! However, can you give a pointer to a source of the Hiwin carriages and rails? Are these from Hiwin, and what drawing format works for you?

Thanks,

Hi, how did you know it was Fusion? Hehe.

There is a place that I found the CAD documents on .... I import them as STEP file IIRC.

http://hiwin.partcommunity.com/ - you need to register, quick and easy.

Neale
29-11-2015, 05:51 PM
Hi, how did you know it was Fusion? Hehe.


Because you mentioned it in another post! I thought the question fitted better here, though. Many thanks for the answer - I'll go take a look.

I have to say that things like F360 make 3D CAD much more accessible even to a novice amateur like me, and your pictures above are a great demonstration of what can be done.

Clive S
29-11-2015, 06:25 PM
I have just been watching a video here https://www.youtube.com/watch?v=o-GBpUZ3piY If you have used SW it seems very easy to use. I am impressed and will give it a try.

Chaz
29-11-2015, 06:45 PM
Any comments on my mill design ;p ?

routercnc
29-11-2015, 10:36 PM
Hi Chaz,

Some thoughts -
1.The plates which hang off the Y axis (and hold the Y bearings) look weak as they cantilevered off the side.
2. How will you clamp the spindle? Are you planning on using grub screws? Not sure that is a good idea.
3. Are you only using 1 bearing per rail on the Z axis? I would use 2 per rail.
4. I would use 2 screws on Y to get the most out of a machine like this.
5. 2 screws in Z is worth looking at but it takes up Y travel and I think one screw is still pretty good.

I've also been working on a new design over the past few weeks based on the DMG Mori - here is my attempt so far:

Iso metric:
16658

Side:
16659

Top:
16660

Y axis:
16661

Box Z axis:
16662

Underside of Z axis showing spindle clamped by front plate:
16663

uli12us
30-11-2015, 01:49 AM
Yes, forget it. How big should the machine be?
That dubious concept isn't as easy as it looks. I think they have a gantry Axe on Y (if the normal X-Y is correct) only, because Hermle have a patent for a centralised Spindle. Gantry have no advantage against 1 spindle in the middle, except in a really large machine.
And your control must handle 3 of it. Can it do this.
The same for X and Z. This construction is ok, if you have a spindle with a ton or more. If not its much too complicated.
And the Z axe don't work, you have 4 guiding (http://dict.leo.org/ende/index_en.html#/search=guiding&searchLoc=0&resultOrder=basic&multiwordShowSingle=on) carriage (http://dict.leo.org/ende/index_en.html#/search=carriage&searchLoc=0&resultOrder=basic&multiwordShowSingle=on) (I hope the word is correct, the slider on the linear guide) 4 or eventually 3 is correct, but not in these layout.
You need 2 sliders per linear guide. And if you want 4 side by side each must
parallel within a rather tight tolerance.
If its for yourself and you don't want to sell it, then you can make it with a central spindle.

A gantry spindle brings no advantage, but much more effort, each Motor and driver must bought double.

The concept with the u-shaped bed is good, but all other too complicated. Unfortunately I don't know, how I can add a pdf-file.

Chaz
30-11-2015, 12:50 PM
Yes, forget it. How big should the machine be?
That dubious concept isn't as easy as it looks. I think they have a gantry Axe on Y (if the normal X-Y is correct) only, because Hermle have a patent for a centralised Spindle. Gantry have no advantage against 1 spindle in the middle, except in a really large machine.
And your control must handle 3 of it. Can it do this.
The same for X and Z. This construction is ok, if you have a spindle with a ton or more. If not its much too complicated.
And the Z axe don't work, you have 4 guiding (http://dict.leo.org/ende/index_en.html#/search=guiding&searchLoc=0&resultOrder=basic&multiwordShowSingle=on) carriage (http://dict.leo.org/ende/index_en.html#/search=carriage&searchLoc=0&resultOrder=basic&multiwordShowSingle=on) (I hope the word is correct, the slider on the linear guide) 4 or eventually 3 is correct, but not in these layout.
You need 2 sliders per linear guide. And if you want 4 side by side each must
parallel within a rather tight tolerance.
If its for yourself and you don't want to sell it, then you can make it with a central spindle.

A gantry spindle brings no advantage, but much more effort, each Motor and driver must bought double.

The concept with the u-shaped bed is good, but all other too complicated. Unfortunately I don't know, how I can add a pdf-file.

Thanks. I understand most of what is being said. Ill respond later when I have a bit more time regarding some questions.

Please email me the pdf - ctintinger at gmail dot com.

Thanks

komatias
30-11-2015, 01:01 PM
Chaz,

You seem to have money to spend on this which is why I am keen to see what you come up with. My suggestion on the spindle arrangement is to make the box section adjustable so that you use two of the rails as data to bring the other two in to take up the tolerances.

Getting two carriages per rail should go without saying, but you may get away with two per rail on the back and one per rail on the front.

Do you have the powersupply for the 4KW spindle?

routercnc
30-11-2015, 01:17 PM
Chaz,

. . . My suggestion on the spindle arrangement is to make the box section adjustable so that you use two of the rails as data to bring the other two in to take up the tolerances.




Agree with that. My plan was for the green sections on the Y axis to have the vertical end plates arranged so that I can fix one pair of carraiges and then 'slightly pre-load' the other plate up to the opposite side so everything is aligned. This assumes the machining on the Z axis plates is good where the rails sit.

Chaz
30-11-2015, 01:37 PM
The reason for the grub screws to hold the spindle was that I did not want to distort the 'box' by slotting it and then clamping the spindle but wasnt sure if the grub screw method would work.

The same for the 4 way / 4 carriage system. Im not sure how the Octagon system works on the benchmark mill .... need more info / pics and google hasnt helped.

This is why we get feedback :-)

Chaz
30-11-2015, 01:40 PM
Chaz,

You seem to have money to spend on this which is why I am keen to see what you come up with. My suggestion on the spindle arrangement is to make the box section adjustable so that you use two of the rails as data to bring the other two in to take up the tolerances.

Getting two carriages per rail should go without saying, but you may get away with two per rail on the back and one per rail on the front.

Do you have the powersupply for the 4KW spindle?

Seem to :-)

The idea was built out of getting something very rigid. I like the design and if you follow this thread (http://www.cnczone.com/forums/vertical-mill-lathe-project-log/259070-cnc-posts-16.html) that is the material type I wanted to use.

I believe that in most cases the machines that are made by Datron and Mori Seiki are either cast iron or some form of epoxy / cement type material.

For the spindle, yes, Im running a 32A feed on single phase. Plus, the likelyhood of using all 4KW is low.

Clive S
30-11-2015, 01:56 PM
Unfortunately I don't know, how I can add a pdf-file.I think you have to go the advanced tab and then click on attachments

uli12us
30-11-2015, 02:16 PM
16664
Ok, it works I hope. Here is the only flyer I found, where you can see, how the hermle machines were build. On page5 there is a gantry construction and page 7 shows the normal construction. These machines run since 15Years or so, without problems. The spindle for the Y-Axe is hidden behind the spindlehead in the middle of the X-axe. I don't know exaktly, but I think in this case the have a stationary spindle and a driven nut. The only obvious disadvantage of these construction is, you need at least the complete Y-way free space behind the machine.

The flyer is free downloadable on the hermle homepage, so I think its ok, I post it here, without any copyright issues.

Chaz
30-11-2015, 02:22 PM
16664
Ok, it works I hope. Here is the only flyer I found, where you can see, how the hermle machines were build. On page5 there is a gantry construction and page 7 shows the normal construction. These machines run since 15Years or so, without problems. The spindle for the Y-Axe is hidden behind the spindlehead in the middle of the X-axe. I don't know exaktly, but I think in this case the have a stationary spindle and a driven nut. The only obvious disadvantage of these construction is, you need at least the complete Y-way free space behind the machine.

The flyer is free downloadable on the hermle homepage, so I think its ok, I post it here, without any copyright issues.

Thanks. Will go through it in detail when I get some time.

Chaz
30-11-2015, 06:15 PM
Hi Chaz,

Some thoughts -
1.The plates which hang off the Y axis (and hold the Y bearings) look weak as they cantilevered off the side.
2. How will you clamp the spindle? Are you planning on using grub screws? Not sure that is a good idea.
3. Are you only using 1 bearing per rail on the Z axis? I would use 2 per rail.
4. I would use 2 screws on Y to get the most out of a machine like this.
5. 2 screws in Z is worth looking at but it takes up Y travel and I think one screw is still pretty good.

I've also been working on a new design over the past few weeks based on the DMG Mori - here is my attempt so far:

Iso metric:
16658

Side:
16659

Top:
16660

Y axis:
16661

Box Z axis:
16662

Underside of Z axis showing spindle clamped by front plate:
16663

1.The plates which hang off the Y axis (and hold the Y bearings) look weak as they cantilevered off the side.

They could be improved / widened, would that still be a concern?

2. How will you clamp the spindle? Are you planning on using grub screws? Not sure that is a good idea.

Agreed, not sure either. Problem is, how do you ensure that the square remains perfectly square (for the rail system) when you tighten to clamp the spindle?

3. Are you only using 1 bearing per rail on the Z axis? I would use 2 per rail.

Understood / agreed.

4. I would use 2 screws on Y to get the most out of a machine like this.

Understood / agreed.

5. 2 screws in Z is worth looking at but it takes up Y travel and I think one screw is still pretty good.

Fair point, I was looking at this to mitigate for only having 4 carriers. I dont really want to go to 8 and I like the idea of a square / Octogon shape for strength. Suppose a traditional design might be better.

routercnc
30-11-2015, 07:29 PM
Hi Chaz,

If you look at my design the side plates are braced back to the main Y axis assembly with effectively triangular structures. There is also a closer plate (see Y axis picture) to create a local box structure, which also provides stiffness for the Y ballscrew to connect to. On your design some triangular bracing pieces could work with a notch / clearance area at the inside corner to clear the ballscrew.

I thought alot about the spindle clamping and problems with keeping the rails aligned. In the end the spindle clamp is quite small, and the plate on the bottom surface is fixed (the transparent piece in the last picture), plus the front plate above the spindle is also fixed. This should keep the rails aligned whilst adjusting / clamping the spindle. Look at the last picture I posted for my thoughts on how to do it.

I drew loads of single and twin Z axis ballscrew arrangements but could not get a compact package with twin screws, so after agonising over it went back to single which has the smallest offset from the centre line I could achieve.

As for the Z axis bearings on the dmg machine - I don't know this for a fact but I can't see any Z rails so my conclusion is that is runs on a finely ground flat surface on the hexagon faces - bit like the ways on a mill / lathe but flat and trapped on all sides. You can see some sort of pre-load screw(?) or oiling point so perhaps it also has a film of high pressure oil in there like an engine crank/big end. Really just guessing here so would welcome thoughts from others on this. Either way, probably not a DIY solution.

Chaz
30-11-2015, 07:45 PM
Hi Chaz,

If you look at my design the side plates are braced back to the main Y axis assembly with effectively triangular structures. There is also a closer plate (see Y axis picture) to create a local box structure, which also provides stiffness for the Y ballscrew to connect to. On your design some triangular bracing pieces could work with a notch / clearance area at the inside corner to clear the ballscrew.

I thought alot about the spindle clamping and problems with keeping the rails aligned. In the end the spindle clamp is quite small, and the plate on the bottom surface is fixed (the transparent piece in the last picture), plus the front plate above the spindle is also fixed. This should keep the rails aligned whilst adjusting / clamping the spindle. Look at the last picture I posted for my thoughts on how to do it.

I drew loads of single and twin Z axis ballscrew arrangements but could not get a compact package with twin screws, so after agonising over it went back to single which has the smallest offset from the centre line I could achieve.

As for the Z axis bearings on the dmg machine - I don't know this for a fact but I can't see any Z rails so my conclusion is that is runs on a finely ground flat surface on the hexagon faces - bit like the ways on a mill / lathe but flat and trapped on all sides. You can see some sort of pre-load screw(?) or oiling point so perhaps it also has a film of high pressure oil in there like an engine crank/big end. Really just guessing here so would welcome thoughts from others on this. Either way, probably not a DIY solution.

Thanks. I think you might be 100% correct, perhaps its a 'perfect' fit solution - would be lovely.

Taking some inspiration from this thread, gantry design with epoxy. (Dutch thread). http://www.cnczone.nl/viewtopic.php?f=8&t=8410

http://www.icecoldcomputing.com/misc/portaalfrees_screenshot_26okt2013.gif

Look at the speed on this - https://www.youtube.com/watch?v=rtJ3MGSM030 and it looks fairly rigid.



This also looks interesting - http://www.icecoldcomputing.com/misc/portaalfrees_geleiding3.jpg

Chaz
30-11-2015, 07:51 PM
This is the same machine cutting steel.

https://www.youtube.com/watch?v=0etcYg6S0dU#t=74

routercnc
30-11-2015, 08:37 PM
Hi uli12us,

I'm not sure I follow your description of the Hermle machine. It looks to me like it uses a SINGLE gantry beam with the spindle hanging off the front, like most of the DIY machines on this forum.

As I see it the spindle moves left/right across the gantry, and up/down, as per normal. It is not a fixed gantry with moving workpiece (ignoring the 4th&5th axis for now). The whole gantry can also move fore/aft. This is same as the dmg except that the dmg has a TWIN beam gantry with the spindle in the middle so is potentially stiffer than the Hermle.

The Hermle may look like the spindle is 'fixed' because of the pieces either side but I think these are protection bellows to keep the swarf out and can move as the spindle moves. Let me know if you are seeing something I am not !

All this aside, I think it is generally agreed that for small machines a fixed gantry (with spindle only moving up and down) and a moving X/Y bed is better, which might have been the point you were trying to make as a general comment.

Thanks

routercnc
30-11-2015, 08:44 PM
Interesting machine Chaz in that video.

Spindle speed looks strangely high for steel but it seems to be working at that feed rate.

Forgot to mention that I'd been looking at double ballscrews, or using back-to-back singles with belleville washers in my machine design to remove (up to a point) the 50 microns or so backlash. Something to think about if you are really trying to max out on performance. I'm undecided on whether to bother . . .

Boyan Silyavski
02-12-2015, 06:12 PM
Chaz,
have not been here long, just looked at your initial design .

2 major flaws pop at once:
-the gantry profiles
-the bearing blocks for the Z

Let me explain how i see it and what i mean. I would compare with my design / the violet one/ so its easy to understand, as mine is overbuild also.
-If you take a closer look on the reference machine from the pdf you would see that these are not profiles, the whole gantry is massive cast and braced piece. meaning that your is flimsy in comparison. The only way i see it is make it like my machine- solder a additional reinforcing plates at the gantry beams to tighten all up. Thats why in fact i did not went with that design, cause while on Johnatans red machine it works, for a bigger machine gantry needs serious reinforcing. So i designed my gantry/a typical style/ so its more rigid than the gantry you have in mind. The ideal super machine for me would be combined between both designs. meaning 3 profiles for the Z left right. An overkill. Say no more than 200Euro more on rail and profile


-For the Z i have no doubt. Carriages must be fixed and rails must be on Z , for ultimate strength. plus better 2 rails x2 carriages than 4 carriages and 4 rails. The arm of force, the lever. Now its good to have in mind that for that type of machine and 2 rails with Hiwin 20 x 2 carriages is overkill. So no need for 4.
I bet that the way i designed it /30cm lever arm and 20cm z travel is more rigid than not properly designed Z with 4 carriages.


So why they made that original monster like that. Cause for sure the spindle is much bigger power and weight than what you will use, the gantry is cast and generally is made for hard metal, not aluminum.


I will be happy to explain better if i not clear what i say. Soon i will be able to prove what i am saying, when chips start flying from mine machine.

Chaz
02-12-2015, 07:13 PM
Chaz,
have not been here long, just looked at your initial design .

2 major flaws pop at once:
-the gantry profiles
-the bearing blocks for the Z

Let me explain how i see it and what i mean. I would compare with my design / the violet one/ so its easy to understand, as mine is overbuild also.
-If you take a closer look on the reference machine from the pdf you would see that these are not profiles, the whole gantry is massive cast and braced piece. meaning that your is flimsy in comparison. The only way i see it is make it like my machine- solder a additional reinforcing plates at the gantry beams to tighten all up. Thats why in fact i did not went with that design, cause while on Johnatans red machine it works, for a bigger machine gantry needs serious reinforcing. So i designed my gantry/a typical style/ so its more rigid than the gantry you have in mind. The ideal super machine for me would be combined between both designs. meaning 3 profiles for the Z left right. An overkill. Say no more than 200Euro more on rail and profile


-For the Z i have no doubt. Carriages must be fixed and rails must be on Z , for ultimate strength. plus better 2 rails x2 carriages than 4 carriages and 4 rails. The arm of force, the lever. Now its good to have in mind that for that type of machine and 2 rails with Hiwin 20 x 2 carriages is overkill. So no need for 4.
I bet that the way i designed it /30cm lever arm and 20cm z travel is more rigid than not properly designed Z with 4 carriages.


So why they made that original monster like that. Cause for sure the spindle is much bigger power and weight than what you will use, the gantry is cast and generally is made for hard metal, not aluminum.


I will be happy to explain better if i not clear what i say. Soon i will be able to prove what i am saying, when chips start flying from mine machine.


Thanks, appreciate the feedback. Agreed on the Z design, it was an idea that wont work. In terms of 'reinforcing' the gantry, this can be done easily. Your purple machine is a good example and the reason why I am staying away from steel is due to me not owning a MIG welder (that can be changed, I used to Arc Weld a bit years ago), but that steel looks to be a pain to get flat based on welding etc. That said, it clearly is and there are other 'things' you can place on top of a steel frame that help with something being flat.

I also wanted to 'play' with something exotic and interesting like something that I can cast or make myself. I researched about casting iron, the epoxy options and even looking at granite.

So ye, I might not be making the 'best', at least it might be interesting (and perhaps risky). That said, I dont have money to piss away, it needs to work.

Noplace
03-12-2015, 07:35 PM
Hi this looks really cool, I checked the original DMG Mori machine and the design blew my mind! any idea if the whole frame is one piece from cast iron or something of that sort? I wonder if us mere mortals would be capable of building something like that.

anyway it's nice that you are following their design, it seems it is one of the latest advanced technologies. and maybe you can add the B and C axis as well hehehehe

Robin Hewitt
03-12-2015, 09:36 PM
Profile rails come in 2 flavours, those that screw from above, those that screw from behind.

If you screw from above, swarf and grit can sneak in passed the grease wipers on the ends.

Chaz
03-12-2015, 09:39 PM
Profile rails come in 2 flavours, those that screw from above, those that screw from behind.

If you screw from above, swarf and grit can sneak in passed the grease wipers on the ends.

Had this on a previous machine. pain.

Chaz
03-12-2015, 09:40 PM
Hi this looks really cool, I checked the original DMG Mori machine and the design blew my mind! any idea if the whole frame is one piece from cast iron or something of that sort? I wonder if us mere mortals would be capable of building something like that.

anyway it's nice that you are following their design, it seems it is one of the latest advanced technologies. and maybe you can add the B and C axis as well hehehehe

indeed. i agree, still like the design.

JAZZCNC
03-12-2015, 11:31 PM
Profile rails come in 2 flavours, those that screw from above, those that screw from behind.

If you screw from above, swarf and grit can sneak in passed the grease wipers on the ends.

Depending on make you can buy top mounting with covers to stop this. Don't mean the Green caps like Hi-win but full metal strip that clips into the rail.
Not cheap put that's what you'll find on many industrial VMC's that use Linear rails.