3 Attachment(s)
Small watchmaker CNC design idea
Hello fellow CNC nerds!
Im a watchmaker by profession and recently we got a CNC at work and I really enjoy working on it,
So I'd like to build a tiny one that I can have it in my flat ( I live in an apartment in a big city)
So I started to model my idea, the basic concept for the frame is to use metrology granite measuring plates, for the bottom use 400x250mm, and two triangular ones for Z, I will need to drill and glue inserts to make it all work and will be a challenge but then I will have an excellent frame material and excellent flat surfaces and geometry.
Other components I have fairly decided on are the following= hiwin 15mm rails, 100w servo motors by delta( nema 17), mechatron ER11 0.8kw motor, C3 ground ballscrews ( NSK?)
The model is not finished and has some mistakes, Y rails are too long, Y motor will be mounted on the front, all of them direct drive.
For reference I will normally use end mills 1mm diameter or under, so cutting forces are tiny. Also I do depth of cut of around 0.1mm for a 1mm end mill. Spindle speed at around 12.000 rpm
If you can recommend me some ball screws and what do you think of the design I appreciate it, I am a little lost on ballscrew pitch, resolution, etc. I was thinking 10mm diameter of the ballscrew.
Re: Small watchmaker CNC design idea
Hi,
my first mini-mill was of similar dimensions, but of a more normal bed mill type, made out of cast iron. I got some great big cast iron elevator weights (1000 x 200 x 100) from the scrap yard and milled the
axis beds and other parts out of them. Worked pretty well. I used it for 61/2 years before building my new bigger and more powerful mill.
While the machine was rigid enough for metals, the majority of my work is for PCB's. I too have a Mechatron 800W spindle, and I've used it, and used it, and used it.....for
eight years. The only time it EVER goes below 24000 rpm is if I'm cutting steel. For just about all materials with tools 1mm and under you will be rpm limited, run it a s fast as you can.
I used 20mm Koruda 5mm pitch C5 ballscrews, because I could get them second hand. Sourcing good quality ballscrews is about the most challenging part of the whole build.
You are right to choose ground screws, I've used C5's and had great results, C3 would be better but really only for skiting, C5's will be as accurate as you could possibly require.
If you use thin screws then they will tend to 'whip' at speed, and so you'll need coarse pitch screw to avoid having to spin them that fast. The minimum I would consider is 16mm diameter,
then at the lengths you are talking you can spin as fast as you want, say 3000rpm without whipping.
Fine pitch, say 5mm or less, allows great thrust from modest servos but slower rapids. Higher pitch, say 10mm or more, requires more torque from your axis motors but blinding fast
rapids. With such a small machine fast rapids are not really required, so I'd recommend 5mm pitch.
My new mill has 32mm C5 5mm pitch screws and I can still get 25m/min by spinning them at 5000rpm (Delta 750w servos), but I find to too scary fast and have detuned them to 15m/min,
and when making PCBs I reduce it further still. My recommendation is do not get too hung up on fast rapids, 5m/min or thereabouts is ample. In a production environment 50m/min is
more likely....but yours is not a production machine.
Acceleration is overall more important than speed, the higher the acceleration potential of your machine the better the machine can follow a toolpath. Acceleration is best with
small (5mm<) pitch screws.
I bought my ballscrews and linear rails from this outfit:
https://www.ebay.com/str/industrialp....m47492.l74602
I was able to get absolutely the top brands THK, NSK etc, new-old-stock for a small fraction of new price.
Are you sure you want granite? Its about 1/4 as stiff a steel, 1/2 as stiff as cast iron, costs way WAY more and is a PITA to drill and attach anything to.
Just as a matter of interest price a piece of steel say 32mm thick, 400x250 from a profile cutter, and find out how much it would cost to take it to a tool maker and have him throw it in the surface grinder.
I'll lay you odds its way cheaper and so SO much stiffer than granite.
Craig
Re: Small watchmaker CNC design idea
Hi,
these are beautiful, they're brand new, C3 grade no less:
https://www.ebay.com/itm/19516814125...YAAOSwp8Zit~1I
Not especially cheap at $269USD each and that's without end bearings but still way cheaper than new. Travel (290-42=248) 248mm, 5mm pitch. There's more than 10 of them.
These things would 'put an erection on a jellyfish'.
Craig
Re: Small watchmaker CNC design idea
Hi,
these to might be interesting, from Automation Overstock. Two (new) 15mm diameter, 10mm pitch C5 screws:
https://www.automation-overstock.com...F%2D3P%2DC5Z10
Just a bit cheaper. Note that only the support bearing end is machined, estimated travel 270mm.
Craig
Re: Small watchmaker CNC design idea
Quote:
Originally Posted by
joeavaerage
Are you sure you want granite? Its about 1/4 as stiff a steel, 1/2 as stiff as cast iron, costs way WAY more and is a PITA to drill and attach anything to.
Why wouldn't he want granite, its already flat. Big problem gone from the get go. It's plenty stiff for the job.
Re: Small watchmaker CNC design idea
Hi,
Quote:
Why wouldn't he want granite, its already flat
That's easy, granite is expensive and hard to work with. For a fraction of the cost a slab of steel can be surface ground and be very much stiffer. If you want to attach something
then drill and tap, no diamond drill and bonding inserts etc. There is a very goosd reason steel and iron still dominate in the machine market....cost to achieve a given stiffness.
Craig
Re: Small watchmaker CNC design idea
Quote:
Originally Posted by
joeavaerage
Hi,
That's easy, granite is expensive and hard to work with. For a fraction of the cost a slab of steel can be surface ground and be very much stiffer. If you want to attach something
then drill and tap, no diamond drill and bonding inserts etc. There is a very good reason steel and iron still dominate in the machine market....cost to achieve a given stiffness.
Craig
granite is much cheaper compared to having steel ground at a company, you know those prices well. It's completely stress free too, having been there for millions of years. Sounds to me like it's perfect for micro machining.
Re: Small watchmaker CNC design idea
Hi,
Quote:
granite is much cheaper compared to having steel ground at a company,
Complete and utter rubbish, grinding and machining is not cheap but its not the end of the world either. If that were the case ALL machine would be made with granite...and are they...no.
Granite Youngs Modulus 30-60GPa
Epoxy Granite Youngs Modulus 15-30GPa
Steel Youngs Modulus 210GPa
Cast Iron (Grey 40kPsi) 110GPa
Craig
Re: Small watchmaker CNC design idea
Quote:
Originally Posted by
joeavaerage
Hi,
Complete and utter rubbish, grinding and machining is not cheap but its not the end of the world either. If that were the case ALL machine would be made with granite...and are they...no.
Granite Youngs Modulus 30-60GPa
Epoxy Granite Youngs Modulus 15-30GPa
Steel Youngs Modulus 210GPa
Cast Iron (Grey 40kPsi) 110GPa
Craig
I know my prices well. Building with precision granite is 3x cheaper than having things machined at a company, where I live. Machining/grinding prices are absolutely the end of the world.
Of course companies make with steel and iron because they have invested in their own massive double column centers, so they don't have to pay another company to do their machining, that's the only way it's viable financially. Which has nothing to do with diy building.
His frame is build from premade precision parts. All he has to do is drill holes and rough up some surfaces that he can glue together. It can't get any more convenient than that while also being precise.
Re: Small watchmaker CNC design idea
Hi,
Quote:
Building with precision granite is 3x cheaper than having things machined at a company, where I live
Delighted that you can get granite cheaply, given that its as stiff as a limp noodle you'll need it.
Quote:
Of course companies make with steel and iron because they have invested in their own massive double column centers,
Whos talking about putting parts in a machine like that? The parts that OP pictured can be done in a much smaller machine, but more importantly I suggested he
price just the 400x250 slab, that can go in a surface grinder. I had three pieces done at Christmas about the same size, had both surfaces done on each, cost me $90NZD
(approx $60USD), cheap as chips, flat and parallel within 1/10th thou.
Craig
Re: Small watchmaker CNC design idea
Hi,
Quote:
I'd like the name of that company. 1200x800x120 granite slab $1500, if I cast the same slab with inserts and have them machined at a company $6000-8000. That's even without grinding. You're very lucky to have access to such services.
That's a vastly different proposition, there are very few grinders that could handle 1200 x 800. A piece of that size would have to be milled flat, followed by grinding if desired, or even possible.
OP mentioned a slab 400 x 250, which does comfortably fit in a common toolroom grinder. The pieces that I had ground are 275 x275 x20. Width is usually the most constrained dimension
in a surface grinder, and the closest tool company to me has a grinder that can do 305mm width, so my 275mm wide piece could be done in one pass, and at that size did not require face milling
prior to grinding.
If you start welding steel, even material like the 20mm plate, then you might expect the residual stress from welding would cause the steel to move over time. If you do not weld then the only stress is that left over from
when the plate was hot rolled in the slab mill when it was made, and given its done hot the stresses are minimal. I've never detected 20mm plate to move unless some subsequent process causes it.
To attach parts to the steel plate just requires drilling and tapping. Easy.
To join granite requires drilling with a diamond drill and epoxy bonding inserts. In many circumstances, OPs proposed model included, require that the inserts come under tension to prevent wracking
of the two granite parts. Epoxy is about 2 GPa, even filled epoxies are not much more than 5GPa, so the inserts will move under load, and consequently the granite parts will wrack.
A drilled and tapped hole can in steel be torqued until the bolt starts to yield, a far better circumstance than the inserts yielding.
Craig
Re: Small watchmaker CNC design idea
Hey guys I appreciate your talk on frame material, however after some research seems that granite is the winner, in most ultra precision cnc mills the frame or base is granite or granite epoxy, the cost of the 3 plates I propose come at around 1000 euros and for me that's resonable if I only need to bolt them together, sure the same build can be done in steel and cheaper maybe if you can access a big shop or have friends with that kind of machines but I live in a city centre so no machine shops here.. in the end my cutting forces from 1 mm end mills and 0.1mm depth of cut is nothing, so steel or granite will be overkill either way, drilling and glueing inserts with the right tooling will be fairly easy and is well documented.
Check out this insane build: https://www.usinages.com/threads/pro....114107/page-7
Frame material is the least of my worries right now, ballscrew choice, calculations to know if a 100w servo is enough or what preload I should order rails are the kind of things that Im quite lost at the moment.
Re: Small watchmaker CNC design idea
Hi,
Quote:
in the end my cutting forces from 1 mm end mills and 0.1mm depth of cut is nothing, so steel or granite will be overkill either way, drilling and glueing inserts with the right tooling will be fairly easy and is well documented.
Hi the cutting forces are minimal but the inertial acceleration remains the same, accelerate any mass and force is required.
Quote:
the cost of the 3 plates I propose come at around 1000 euros and for me that's resonable if I only need to bolt them together
You will require inserts that come under tension, and they will move, and your machine will not be stiff.
Your choice of ballscrews will determine the size and weight of the machine, which will determine servo size etc. Start with the ballscrews, they are the most expensive, hard to find
and yet absolutely machine critical components.
Craig
Re: Small watchmaker CNC design idea
Quote:
Originally Posted by
markgonz
Hey guys I appreciate your talk on frame material, however after some research seems that granite is the winner, in most ultra precision cnc mills the frame or base is granite or granite epoxy, the cost of the 3 plates I propose come at around 1000 euros and for me that's resonable if I only need to bolt them together, sure the same build can be done in steel and cheaper maybe if you can access a big shop or have friends with that kind of machines but I live in a city centre so no machine shops here.. in the end my cutting forces from 1 mm end mills and 0.1mm depth of cut is nothing, so steel or granite will be overkill either way, drilling and glueing inserts with the right tooling will be fairly easy and is well documented.
Check out this insane build:
https://www.usinages.com/threads/pro....114107/page-7
Frame material is the least of my worries right now, ballscrew choice, calculations to know if a 100w servo is enough or what preload I should order rails are the kind of things that Im quite lost at the moment.
The idea to use the granite squares and base plate is a good choice, as they can be any accuracy grade you want them to be
As for 100w direct drive servos don't even consider them, nothing below 400w for your machine, for Ballscrews C3 / C5 minimum 16mm to 20mm diameter.
Mount the Y axes servo out the back as you will have room back there.
Re: Small watchmaker CNC design idea
Quote:
Originally Posted by
mactec54
The idea to use the granite squares and base plate is a good choice, as they can be any accuracy grade you want them to be
As for 100w direct drive servos don't even consider them, nothing below 400w for your machine, for Ballscrews C3 / C5 minimum 16mm to 20mm diameter.
Mount the Y axes servo out the back as you will have room back there.
Thanks for the answer, why do you think 100w servos are too little? my reasoning is that I've seen bigger machines, like 200kg and 2kw spindles use 400w servos, and I have seen a similar power machine as the one I want to build use 100w servos, Im talking about this one: https://build-its-inprogress.blogspo...axis-mill.html, I would like to go with the nema 17 size if possible because suits better the machine I believe
Re: Small watchmaker CNC design idea
Hi,
Delta servos, a Taiwanese brand made in China, or DMM, a Canadian brand made in China are highly recommended. Both brands perform well, good quality, support,
documentation and most importantly manufacturer supplied set-up and tuning software at fair prices.
There are cheaper Chinese brands, and most perform well but the documentation is poor, support questionable and no set up and tuning software. If this is your first foray into servos
I would not recommend them despite being cheap.
If you look at the prices of what's offered the price of a 400W servo is only marginally more than a 100W servo, so it makes economic sense to get the 400W unit.
Just because the cutting forces are low does not lessen the need for torque to accelerate the axes, the more power, the more torque, the better the acceleration.
All the 400W units are 23/24 size. The next step up is 750W but they are 34 size, but only another $30 or $40 over a 400W unit...still probably outsize for your machine.
Craig
Re: Small watchmaker CNC design idea
Quote:
Originally Posted by
markgonz
Thanks for the answer, why do you think 100w servos are too little? my reasoning is that I've seen bigger machines, like 200kg and 2kw spindles use 400w servos, and I have seen a similar power machine as the one I want to build use 100w servos, Im talking about this one:
https://build-its-inprogress.blogspo...axis-mill.html, I would like to go with the nema 17 size if possible because suits better the machine I believe
If you only want a wimpy machine like that example then you need to rethink your whole machine, the granite structure you were planning on, would not be of much use if you were to use only 100w motors if you have ever seen or used an Emco cnc mill like the PC50 you will soon know the value of having more power, as the small motors on those machines, which are not direct drive, stall out very easy with wimpy cuts, so you want to use direct drive, which requires even more power
1 Attachment(s)
Re: Small watchmaker CNC design idea
Quote:
Originally Posted by
mactec54
If you only want a wimpy machine like that example then you need to rethink your whole machine, the granite structure you were planning on, would not be of much use if you were to use only 100w motors if you have ever seen or used an Emco cnc mill like the PC50 you will soon know the value of having more power, as the small motors on those machines, which are not direct drive, stall out very easy with wimpy cuts, so you want to use direct drive, which requires even more power
I get your point and understand that is always better to have enough power, but for the intended use of my machine witch is for watchmaking parts I think is good, the good way to do it of course would be to calculate the cutting forces involved, etc, but this is way out of my league.
I think its relative and you don't realise how small watchmaking parts are, a wimpy cut on a EMCO pc mill may be a monster cut for a watchmaking part, I attached an image of an example part I'd like to do, a watch dial is 0.4mm thick and 30mm in diameter, the cuts to make the engravings of the numbers with good finish would be around 0.1mm depth and take a while.
Re: Small watchmaker CNC design idea
Hi,
then do the calculation.
Lets assume a 16mm diameter 5mm pitch ballscrew 300mm long:
mass= (0.008)2x PI x 0.3 x 8000 (8000=density of steel)
=0.5kg
The first moment of inertia of the ballscrew:
Jballscrew= 1/2 x mass x r2 (mass is the mass of the ballscrew and r is ballscrew radius)
=1/2 x 0.5 x(0.008)2
=0.16 x10-4 kg.m2
The first moment equivalent of the linear accelerating axis mass:
Jlinear=mass. p2/ 4.PI2 (where mass is the axis mass and p is the pitch )
=10 x (0.005)2 /4 .PI2
=0.063 x 10-4 kg.m2
Note I assumed an axis mass here of 10kg. The Z axis and spindle maybe a little more while the Y axis maybe a little less. Note however that the moment of inertia of the ballscrew alone is three
times the moment of the linear accelerating axis mass, and so the calculation changes very little with axis mass, ie 10kg is a perfectly acceptable ball park figure.
The first moment of inertia of a 100w Delta servo aramature is 0.037 x 10-4 kg.m2, and has a torque of 0.32Nm (rated to 3000rpm)
Angular acceleration is:
d2w/dt2= 0.32 / (0.16 + 0.063 +0.037 )x10-4
=12307 rad/s2
or equivalently:
d2x/dt2= 12307/ (p x 2 x PI)
=9.8m/s2 or 1g, which is pretty impressive for a little servo.
The first moment of inertia of a Delta 400W servo is 0.277 x 10-4 kg.m2 and the torque is 1.27Nm (rated to 3000rpm):
Angular acceleration is:
d2w/dt2= 1.27 / (0.16 + 0.063 + 0.277 )x10-4
=25400 rad/s2
or equivalently:
d2x/dt2= 25400/ (p x 2 x PI)
=20.2m/s2 or 2g, which is outstanding!
Note this calculation does not make any attempt to allow for friction, ballscrew efficiency (typically 95%) or cutting forces, but with either the 100W servo or the 400W servo the acceleration
is still really high and allowances for those forces would still indicate that either would be more than adequate.
I personally would still go for 400W servos on the basis of economics and potential future use.
Craig
Re: Small watchmaker CNC design idea
Hi,
note that the rotational inertia dominates the acceleration equations, and is in fact quite common, even the norm.
In the case of the 100W servo the linear momentum of the axis is 32% of the total inertia, and in the case of the 400W servo the linear momentum is only
14% of the total inertia.
The inertia is very sensitive to ballscrew diameter, to the fourth power, and moderately sensitive to the ballscrew pitch, to the second power, but only linearly dependent on axis mass.
Craig