[Tkamsker]

Hi people
there is a thread called "Show how to build a CNC machine from the very beginning to the end" where a chinese manufacturer attracted people how to build an CNC Machine (similar to Tormach, Novakon, Syll, Optimum machines ) some people (like me ) have invested and since long time nothing has happened ,..
So while waiting there is in that thread a discussion about VMC mill designs.

Do this thread will be in general about pro cons and designs stuff of this kind of machines. And i will make my requirements clear.
Hast to be able to do Steel , X axis 400mm y axis 200mm Z axis 200mm or bigger.
CNC is a must. It should not stall and small batches should be able as well. Later an ATC would be nice.

So sayed that there is the possibility to retrofit an "old" VMC mill (you can search for my Bridgeport 412 Retrofit thread so i know what i am talking about ) this option is not covered here.
And i will also do no look on mach3 , linuxcnc or similar i expect servos or steppers in that size of machine.

Machine Frame :
there was an discussion if it has to be casted or can be welded in steel.
I think it depends of the quality of steel if you have good quality steel why not weld it there is an given wall thickness you need for the needed strengt.
So in my opinion there is no need of cast for an VMC mill

Linear Rails or dovetails.
i have seen heavy and precise dovetail an machines (deckel or similar) i personally think that proper dimension linear rails make more sense for an cnc machine because the friction is less. so fast and small moves will work better.

Ballbearing Spindle.
I personally think that this is the hart of an cnc machine and has to be also properly sized and high quality.

Drives.
Depending on the requirements i think that in small sector overpowered steppers are acceptable. and in closed loop there is only the end speed the differiaenciator of servo and stepper.
If my intention is alu and high speed feeds then an servos system might be a good choice so i think that in that case the manufacturer should offer that option.

Spindle.
Here i guess it is complicated and depends on that requirement
for instance on my CNC Router i insist on ATC why because i do PCB´s with it. When i have to mill ALU i change the spindle and use the normal Kress with it.
But sayed that i think that the ISO30 Spindle is more expensive but has more pros that cons (i never worked with an R8 of tormach )
I guess that often used tools can be prepared and known by tool database
i also have the advance that the spindle motor force is mechanically transportet by this small blocks.
If the BDLC motors are really that big advantage i would like to hear your opinions

Fazit
If i look for instance on an optimum BF46 and the cnc optimum i can not see why this should be an bad optimum ?
Mr Handlewanker gave me the impression that there is something bad by using the frame of an stable VMC mill and use it as CNC
i have now looked all the dfengs pictures through and can not really see the big advantage.
it seems that cast is cheap compared to EU (i guess less regulations cheap energy ) but i would like to see where is really that big thing.
i almost bought an Syll machine and they are also straight forward.

so i am looking forward for an nice and constructive discuss

[G59]

Subscribed to this thread. I think it should be focused on quality and accuracy. I couldn't see what the big hoopla was about the defeng machine, other than 'cast'. Nothing new or special for me.
A well designed, welded steel plate machine is possible especially if you strategically place vibration dampening materials in the right areas..
Finally more into the DIY realm.

[LeeWay]

A few thoughts here.
If starting with a welded base, I don't think the type of steel for the tube is actually all that critical. I think the actual mounting pads or plates would be where you might want a higher quality steel.
Weld those parts onto the tube. Include thick gussets in the design where possible. That or incorporate then into the design. Like at the base of the column. Perhaps the head to Z plate mount, depending on the design.
I assume some sort of stress relief will need doing before the pads are machined? I also assume that you may want to fill the tubes with epoxy granite or something. I would leave a provision in the design to do this after receipt of the mill. Most of the aggregates can be sourced locally by all, so would not need to pay shipping charges on rocks.
Just a few thoughts. Look forward to this treads discussions and your progress.

[aarggh]

I agree with Lee, nothing beats sheer mass for a mill, so designing it so mass can be added after receipt by the customer would be a really good way of reducing costs and build time required.

cheers, Ian

[G59]

I also agree with Leeway, very ingenious and would save on shipping and possibly facilitate assembly. Pure mass is however not always the best. Good for the foundation, but inherently problematic when trying to achieve rapid moves along with accuracy. What we need is a balance of all the variables combined so we can ensure a top of the line machine that is durable, predictable and reliable, all the while being serviceable.
A good example is the Brother machines TC32. Incredible speeds but at what costs? Large motors and controlled by proprietary motion controllers that would throw this build into the stratosphere. A RoboDrill is another example of over the top.
We also must remember that with the new HSM toolpaths available, pure horsepower or massive assemblies are not always required. Sometimes a faster spindle and quicker motions that can compensate for the lack of torque, can achieve the same results if not quicker, than the old hogging mills. . Look at a Datron M8 for example. Looks like a router on steroids but is truly a machine with great finesse.

[Tkamsker]

Hi,
I agree with Leeway and G59 there is to be a balance of table weight and your intended pieces like
Attachment 253736 Here you see the table of an Bridgeport 412I which allows Pieces to be milled of 350kg ! in the back you see the SEM Servo which can do 8-17 NM and draws 75A !!
but even that servo is belt driven , or another example Attachment 253738 This is an BF 46 CNC kit but dovetails . And here dfengs solution Attachment 253740
I personally think that the basement is cast it makes it verry heavy but i don't see the advance in that configuration (Size and force ) i also have the feeling that the base where x and y axis are connected it is too high.
So if you really have a heavy piece of iron to mill it will stress the rails. where probably dovetails would have a benefit.
Now we come to size and connection of servos/steppers.
Attachment 253742 Here defends solution and to compare the solution on my machine Attachment 253744

I now think if you have the need of speed (alu copper and long traveling of x and y axis ) it may be senseful to have an oversized servo on direct drive ,.. (there i would be interested on your opinion because of pid there may be markings on the goods ) but if i do steel and a lot of 3d work which is mainly little steps and slow speed (what my config is intended to do ) the stepper with an belt drive is an ok sultion i will have force enough but less high speed.
I am now calculating the costs of closed Stepper (nanotec ) and german control (triple beast ) and yes it should work with mach 3 - against an taiwan servo combination using granite devices VSD in dual mode
to see how big the price difference is.

[asuratman]

Hi,
This will be interesting discussion. I will subscribe to this thread. Does this thread will discuss on designing a good cnc mill from ground up? Not retrofit?

[LeeWay]

Another thing to mention. I just finished up tramming my Pulsar last week. Of course, it required shims. I keep plenty of it on hand. However, with some forethought in design as is being done here, the column could have shim pads on each corner. Might not work as well on a large machine, but it would make short work of tramming on a machine this size. I don't think you would loose any strength with them.

The head also needs attention with regard to tramming. I know that good accurate machining of the assemblies will negate most of the need for tramming, but there is always that bit that may need adjusted. They make set screws with swivel pads on the bottom. Those might be perfect for such an application.

I also think the Y motor must be at the rear. The motor in the front would take up some very valuable real estate.

One other thing that I really like about the Pulsar is the SS way covers on Y. Having them on Y and Z make a lot of sense to me. They have to be easily removable though. That goes along with easy service.

[Tkamsker]

Leeway sorry i am not an native speaker (although i work in english on an daily base ,.. ) how do i translate tramming in your context did you build an manual gear ?
About fotos no problem

the design is based on an existing machine
so here is an closeup of x and y Attachment 254646 the manual wheels are still there but sling loaded so they don't turn when the cnc motor is in action.
the other side of X axis is driven in my test setup by an well sized stepper (nanotec ) Attachment 254648 for now i keep the setup because it works maybe in servo setup a direct drive is key,
same for y axis this is well covered under the x and y table where it really make sense. The Z axis is covered by steel plates (like the bridgeport) the back side of y axis is covered by an industry rubber sheet (i have to see how this works by using it ) the front side of y axis only the spindle is covered i designed an sheet metal cover which works like the z axis cover.
now the Motor the motor is an 2,3HP Toshiba VFD Driven one which has belt drive i did an setup to test the spindle orientation (for my Bridgeport 412 Retrofit https://www.youtube.com/watch?v=yzQXqge-NLs )
looks like Attachment 254650 here you see the top side of the spindle (left side) the spindle is an ISO 30 M12 Manual one. (i tinker on an draw bar for that but not for now)
the right side is the driving side of the motor. I think in that setup it makes sense to get an encoder on the spindle (AMT101 or similar ) which then helps at tapping and gear cutting if needed for orientation and speed only one is needed. (I have an mod bus board for the Toshiba where i can really direct control the motor but no time for exzessive testing) my idea is to make the spindle modular.
That when you order an machine you say what flavor you like
so i have a chance to keep the costs down i am awaiting the ATC Spindle which will then lead to test in an electronic wise setup like i did on the bridgeport means i can drive the machine till physical limits !!
Now electronic setup i played around with an touchscreen box (an ugly fast hack but ok for testing ) Attachment 254652 and the inside Attachment 254654 here you see the
din rail for the relays module (general mechatronics) and the benezahn bob board for VFD control on the left you see the backside of an 15" Touchscreen (i intend to use gmoccapy as on my lathe )
bevor that i did experiment with an kind of external electronic box Attachment 254656 here you see my test setup the box is also an proof of concept where
the tripplebeast is inside all switches and on top there is an NCBOX (which runs linuxcnc and Mach3 if needed !) and beside that i tested an ITX board with linuxcnc because the servo require FPGA board and i use same setup in the bridgeport so the black box is an standard ITX pc with the Generalmechatronics FPGA board inside, with this setup i can drive servos and steppers only physical limitations ,..
But the setup was mechanically not good it shaked and because of cooling i need an propper splashguard.
so i did splashguard design i did the machine in scale on 3d printer then i started to tinker Attachment 254658 in the right corner there is the electronics closet.
other angle to get the picture ( i had 8 different iterations to find that solution ) Attachment 254660
Good think is having the factory just 25km from home. I made an agreement with the owner that they can use my design for their customers and here pictures before
painting where you see also the machine base. Attachment 254662 and Attachment 254664 and electronics closet Attachment 254666 backside view.
the design also has an steel holder where the front side (haas like ) is place for the 15" Touchscreen and the keyboard knobs mouse and so on to have it nice.
so from the backside of that cover there is just an LPT Cable extension for the knobs , an VGA and usb extension and an extra line for special knobs like Emergency halt. cooling whatever
i found an company here which would laser cut and bend the sheet metal so i will draw next weekend the LCD part and see when i get it.
hope you like my findings so far

[G59]

That's what I'm hoping for. Something new.

[Tkamsker]

Depends when you mean working on things to show later is dead ??

i am now choosing the right servos integrate. (waiting for price infos )
i plan and design for the Toolchanger part in the head https://www.youtube.com/watch?v=-KKs...0kqdd9Z2XcIcxQ
where the stuff is to increase the y travel which needs proper thoughts
and therefore i have to decide if i go for an encoder similar to Lathe Spindle Encoder or i connect an standard encoder
but when i connect the standard encoder what is the best way ? Modular | Absolute | Encoders | CUI Inc
do i use an gear ? how to mount it best ?
do i stay with 3 speed belt system ?

so this are the actual topics i am working on but probably people have thoughts on them and i should have shared them
so i would be interested if people think that having an small gear inside the head where the encoder is running in same speed as spindle is an good idea ?
if there is an better one i am interested +i also want to keep the 3 speed transmission
not knowing what people what to cut

hard steel to alu ,..
thomas

[G59]

Hast to be able to do Steel , X axis 400mm y axis 200mm Z axis 200mm or bigger.
CNC is a must. It should not stall and small batches should be able as well. Later an ATC would be nice.

Seems reasonable? Is this build going to be all metric?
Can we agree on 10,000 rpm ISO or BT30 spindle? 2.2kw power(3HP).
Single Phase 230-240Volts?
AC or DC servo drives.
Our own software or another Mach3 make it work project?

Many questions, but interesting to say the least.

[Tkamsker]

Hi,
G59

My world is metric so it will be metric in my case
my spindle is 2,3kw at 230 V but can do 400 as well
I aim for 230 240 Volts because not everybody has 400V avail
About servor or steppers will be my next post i basically i have to work it out now i have steppers on it
i will only work on windows if i get heavily payed for so in my case it is linuxcnc but everyone can choose his control how it suits them if a small factory has all in fanuc i guess they insist on fanuc -

i already have ISO 30 in my shop so i stay with it but about 10.000 rpm i have to see
i ordered an ISO 30 ATC Spindle from an german i have to check what the ball bearings are supposed to survive -

but i am here really more about to understand the average pro / con and requirements and see what is key aspect of dfengs design ,..

[asuratman]

Tkamsker,
Will you start with CAD drawing?

[LeeWay]

I hope we might plan to use Mach 4. Surely it's getting closer now.

I would say we would need a Cad drawing once some sizing is figured out. I'll start by saying that 3/8" thick tube is probably plenty for the skeleton. Thoughts?

We would need a width and height dimensions and lengths can be determined later. I haven't actually drawn anything in SolidWorks yet, but it manipulates that kind of change easy in a drawing.
I have to erase the screen in TurboCad.

[G59]

My problem with Mach4, is it's relatively new and exactly how will it communicate with the controller or drives. I'm unfamiliar with it. Does it still use a BOB of some sort? Ethernet maybe?
As for ground up software, I think that's just too big of a bite.

Definitely need some CAD drawings.

3/8" tube has enough meat to be welded on and very rigid for that size machine.

[LeeWay]

Mach 4 isn't released yet. It should be left as an option I think. That is unless he sets on something like Flash Cut. Maybe Kflop, though I know nothing about either of those. I don't know anything about Mach 4 for that matter.

Drives and motors can be sorted out easily enough I think. Digital drives would be nice. AC steppers would work okay in this size machine, but could spring for Servo's. I would go AC over DC.
Defeng got caught up in making enclosures as well. I would think that to keep shipping costs reasonable, enclosures and stands should be sourced locally.
Perhaps provide some plans so a local guy could fab them up.
What you would save in shipping will easily make up for that cost. You can have it made like you want it then. Keep that hassle off the supplier. He will have his plate full with just the machines.

[frigger]

The control software shouldn't even discussed, it should be left to the end user to decide what electronics and control he wants to use to run his machine. Too many options available, I might prefer Linuxcnc and servos to your steppers and Mach 4, kflop, edging cnc, etc, etc. Just build a good base machine and let people accessories on their own with maybe belt drive or direct drive mount option for servos or steppers.

[LeeWay]

The control software shouldn't even discussed, it should be left to the end user to decide what electronics and control he wants to use to run his machine. Too many options available, I might prefer Linuxcnc and servos to your steppers and Mach 4, kflop, edging cnc, etc, etc. Just build a good base machine and let people accessories on their own with maybe belt drive or direct drive mount option for servos or steppers.

Actually, that is not a bad idea either. Bare bones so to speak.
A CNC ready mill. I like it.

[inov8r]

I think you guys are pretty much describing this build from 2007! I am diggin the solid base
Attachment 253700
http://www.cnczone.com/forums/bencht...htop-mill.html