584,817 active members*
4,660 visitors online*
Register for free
Login
IndustryArena Forum > MetalWorking Machines > CNC "do-it-yourself" > Small watchmaker CNC design idea
Page 2 of 2 12
Results 21 to 38 of 38
  1. #21
    Join Date
    Jul 2022
    Posts
    11

    Re: Small watchmaker CNC design idea

    Wow Craig you make these calculations seem easy, thanks for your time, cool to hear that 100w should work, I wanted to share this build too because I think is similar on what my mill idea looks like, unfortunately they don't say what servos or ballscrews they used but that spindle and rails are extremely good quality.

    ZealCNC : A new desktop mill

  2. #22
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,
    my first mini-mill was of similar dimensions with travels of 180 x 180 x 180mm.

    Over a period of time I used it all, and had bigger parts in jigs so I could shift them to fit within the envelope. I also made a 1.8kW servo based spindle so I could do
    a good job on steel. The bottom line is that no matter how big or rigid you build it you'll always find yourself wanting more.

    A 100w Delta servo is 23/24 size, as is the 400W Delta servo, its just longer. On the basis of always wanting more go for the most powerful servos you can afford and squeeze into
    the design. You'll never regret it......but always lament the fact that you bought too smaller servos in the first place.

    Craig

  3. #23
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,
    the little mill you linked to is a 'doozy'.

    Note though that it has just a single extra wide rail for both the Y and X axes. The potential for the car to rock from side to side is the preload of
    the car itself. Your design, like my own mini-mill,has two parallel rails a few hundred mm apart, and the propensity to rock from side to side is vastly reduced.
    By my reckoning much preferred.

    I did notice that they are using Granite drives, which rather suggests they have used servos rather than steppers.

    Craig

  4. #24
    Join Date
    Jan 2005
    Posts
    15362

    Re: Small watchmaker CNC design idea

    Quote Originally Posted by markgonz View Post
    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.
    The problem I see is the 100w motor does not match the granite machine and the 800w spindle you were going to build nothing else, I'm well a where of the parts you make, I make small parts too, there is nothing special about making small parts, and with a simple light machine build you could use 100w motors, the Granite machine would be a good build, a small light weight machine will do the same job for what you want to do and would be easy to move when you have too, if you are worried about noise from a light weight machine build, the spindle motor will make the most noise, there will be almost not cutting noise
    Mactec54

  5. #25
    Join Date
    Jul 2022
    Posts
    11

    Re: Small watchmaker CNC design idea

    Quote Originally Posted by mactec54 View Post
    The problem I see is the 100w motor does not match the granite machine and the 800w spindle you were going to build nothing else, I'm well a where of the parts you make, I make small parts too, there is nothing special about making small parts, and with a simple light machine build you could use 100w motors, the Granite machine would be a good build, a small light weight machine will do the same job for what you want to do and would be easy to move when you have too, if you are worried about noise from a light weight machine build, the spindle motor will make the most noise, there will be almost not cutting noise
    OK I get your point, is that with 100w motors I will not reach the limit of the frame or spindle anytime soon and makes no sense, like having a racing car with wood wheels sort of situation.

  6. #26
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,
    while the calculation shows that 100W servos would work fine both Matec and I suggest 400W, for the extra few dollars why wouldn't you?...but that was determined
    by doing the calculation with a particular ballscrew. If for instance you increase the diameter of the ballscrew to 20mm the calculation changes quite dramatically, it
    less sensitive to pitch but none the less changes the inertia calculation also.

    I would suggest you look very closely and make a decision about the ballscrews, you may have thought that 1000Euro for granite is steep, but just wait until you
    price C5 and C3 ballscrews! Ballscrews are at the heart of the motion system, once you have those many of the other design decisions are simple and just follow along,
    servo sizing for example.

    When I built my first mini-mill I found 4 Koruda 20mm C5 screws second hand for $600USD, I got them and built around them. With my new mill I found
    3 near new THK BNFN double nut 32mm C5 ballscrews and FK end supports for $1000USD delivered to New Zealand.....a steal. If I had paid for new ones I might have
    expected $1500 to $2000USD EACH.

    Given the price of ground (C5, C3) screws you're bound to ask about rolled C7 screws being 1/10th to 1/5th the price. You will point out, correctly, that they have an accuracy of
    50um over 300mm, so with the small parts you are making the error is vanishingly small.....but its not quite so simple.

    There are two pitch errors, the absolute error over 300mm but also cyclic error. A C7 screw can vary as much as 30um over just one revolution, and that's significant.
    A C5 screw has an absolute error of 18um or less and a cyclic error of 8um or less, both specs are better, but they sure do cost!!! I regard C5 as the entry level
    for an accurate machine. C7s would be fine for a plasma table or a wood working CNC router table, but for engineering parts C5 at least is required.

    Craig

  7. #27
    Join Date
    Jul 2022
    Posts
    11

    Re: Small watchmaker CNC design idea

    Quote Originally Posted by joeavaerage View Post
    Hi,
    while the calculation shows that 100W servos would work fine both Matec and I suggest 400W, for the extra few dollars why wouldn't you?...but that was determined
    by doing the calculation with a particular ballscrew. If for instance you increase the diameter of the ballscrew to 20mm the calculation changes quite dramatically, it
    less sensitive to pitch but none the less changes the inertia calculation also.

    I would suggest you look very closely and make a decision about the ballscrews, you may have thought that 1000Euro for granite is steep, but just wait until you
    price C5 and C3 ballscrews! Ballscrews are at the heart of the motion system, once you have those many of the other design decisions are simple and just follow along,
    servo sizing for example.

    When I built my first mini-mill I found 4 Koruda 20mm C5 screws second hand for $600USD, I got them and built around them. With my new mill I found
    3 near new THK BNFN double nut 32mm C5 ballscrews and FK end supports for $1000USD delivered to New Zealand.....a steal. If I had paid for new ones I might have
    expected $1500 to $2000USD EACH.

    Given the price of ground (C5, C3) screws you're bound to ask about rolled C7 screws being 1/10th to 1/5th the price. You will point out, correctly, that they have an accuracy of
    50um over 300mm, so with the small parts you are making the error is vanishingly small.....but its not quite so simple.

    There are two pitch errors, the absolute error over 300mm but also cyclic error. A C7 screw can vary as much as 30um over just one revolution, and that's significant.
    A C5 screw has an absolute error of 18um or less and a cyclic error of 8um or less, both specs are better, but they sure do cost!!! I regard C5 as the entry level
    for an accurate machine. C7s would be fine for a plasma table or a wood working CNC router table, but for engineering parts C5 at least is required.

    Craig
    Hi Craig, thanks again for your help.

    So if I understand correctly the smaller the leadscrew the better the 100W servo should work right? Right now if really the price difference is so small I may go for 400w

    Yes about ball screws I have heard other people tell me so, is much better idea to source eBay deals on really good ballscrews and go from there,I really would like to get c3 ground and preloaded ball screws, the problem I also have is that with the current granite squares my Z travel is on the small side at 90mm,I need to double check but if I get ball screws bigger than 12mm in diameter I need to raise the table to clear it and loose Z travel, in my current design I have 12mm ball screws witch seem a good size, my travels area small so a short 12mm ballscrew should be stiff enough.

  8. #28
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,

    Yes about ball screws I have heard other people tell me so, is much better idea to source eBay deals on really good ballscrews and go from there,I really would like to get c3 ground and preloaded ball screws, the problem I also have is that with the current granite squares my Z travel is on the small side at 90mm,
    I know exactly how you feel, you have one idea but its dimensionally incompatible with another idea, I've suffered the exact conundrum.

    The granite squares or triangles are what causing you the grief, and they are bloody expensive , right? Then ditch them, get some 30mmm, or 40mm or 50mm steel plate profile cut and have two edges
    machined square to each other. Extremely simple machining proposition that I would not expect to cost more than $100-$200, and the profile cut material for $100 or less.
    Then you can determine the dimensions of the squares/triangles to suit your ballscrew choice, not the other way around. Additionally when you drill and tap the edges of the steel
    to accept the bolts through the granite base, you'll be able to tighten then until you risk crushing the granite plate, not pulling the inserts out.

    I'm sorry but I think your 12mm ballscrews are just too whimpy, they'll whip at speed. Screws 15mm or 16mm diameter and 300mm long you should be able to spin at any reasonable speed.
    If you have not jumped on those 15mm C3 NSK's I linked to earlier....then why not?. They are just absolutely perfect for your build, and at $270USD each a steal.

    Per my earlier post finding good ballscrews (at a price) is the key to building a cost effective machine. The 32mm THK C5's at $1000USD made my new mill possible, I'd never be able to afford
    new, and would not be able to build at all.

    Craig

  9. #29
    Join Date
    Jul 2022
    Posts
    11

    Re: Small watchmaker CNC design idea

    Quote Originally Posted by joeavaerage View Post
    Hi,



    I know exactly how you feel, you have one idea but its dimensionally incompatible with another idea, I've suffered the exact conundrum.

    The granite squares or triangles are what causing you the grief, and they are bloody expensive , right? Then ditch them, get some 30mmm, or 40mm or 50mm steel plate profile cut and have two edges
    machined square to each other. Extremely simple machining proposition that I would not expect to cost more than $100-$200, and the profile cut material for $100 or less.
    Then you can determine the dimensions of the squares/triangles to suit your ballscrew choice, not the other way around. Additionally when you drill and tap the edges of the steel
    to accept the bolts through the granite base, you'll be able to tighten then until you risk crushing the granite plate, not pulling the inserts out.

    I'm sorry but I think your 12mm ballscrews are just too whimpy, they'll whip at speed. Screws 15mm or 16mm diameter and 300mm long you should be able to spin at any reasonable speed.
    If you have not jumped on those 15mm C3 NSK's I linked to earlier....then why not?. They are just absolutely perfect for your build, and at $270USD each a steal.

    Per my earlier post finding good ballscrews (at a price) is the key to building a cost effective machine. The 32mm THK C5's at $1000USD made my new mill possible, I'd never be able to afford
    new, and would not be able to build at all.

    Craig
    Yeah maybe this is the way to go, also the granite squares in my design take a lot of space of the bottom plate and I end up with less Y travel, the only thing that worries me is that cutting steel plates to be exactly 90 deegres and then be able to install without having to shim may be a challenge but on the other hand the parts I want to make are really small and short so its not critical, like I don´t actually need micron squareness over 300mm

    I will think about it and may jump on the NSK you linked after I am really sure they would work, I think they are a tad too long but pretty much the size I need.

    If i ditch the squares and get a deal on ballscrews this build may be cheaper than expected and that would be pretty cool, If I can keep it under 5k euro would be awesome

  10. #30
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,

    the only thing that worries me is that cutting steel plates to be exactly 90 deegres and then be able to install without having to shim may be a challenge
    You might be surprised how good profile cutting is but in any event you will need to machine two faces square to each other. Even if you just clamped it down and
    ran around the outside with an endmill in a VMC you will have faces as square as the machine that made them. Any decent, modern VMC in any toolroom in
    good order should be within a few um per 300mm at least.

    When I had my cast iron axis beds machined the guy who did it did so on a 1500mm travel Okuma VMC about five years old. He did the works including drilling
    and tapping the holes for the linear rails. It was always my intention to get them ground after milling, but they were better than I could measure straight off the mill so
    I didn't waste my money. When I offered the rails up to the line of drilled and tapped holes I thought I'd have to clamp/straighten/re-clamp etc......but no. The holes
    were in such a straight line that fitting the rails was so easy. You've heard the saying 'straight as a die', well this is the machine that the die was made on. It only stands
    to reason that it will make a great job.

    It's easy to fall into the trap that granite squares are the last word in 'squareness therefore I must use granite'. BS any decent machinist with a decent machine will get
    a part that you or I would have difficulty in even measuring for square. I have spent many hours/week/months down that rabbit hole only to come to the realisation that
    a much cheaper, easier, direct solution is as close as my local machine shop, and that they consistantly produce highly acceptable parts, perhaps not at the same
    precision as granite, but good enough that I cant measure the differnce.

    How about this for an idea. If you want to use a component that has two very square faces what about a cast iron angle plate?. Here in New Zealand you can buy Chinese
    made cast iron angle plates reasonably cheaply, some of the bigger ones, say 300 x300 x200 are up around $300NZD ($200USD). Why not use a power hacksaw to slice
    one in two? Even a Chinese made square should be good to 10um over 300mm....good enough for a hobby machine while trying to conserve money.

    Craig

  11. #31
    Join Date
    Jul 2022
    Posts
    11

    Re: Small watchmaker CNC design idea

    Quote Originally Posted by joeavaerage View Post
    Hi,



    You might be surprised how good profile cutting is but in any event you will need to machine two faces square to each other. Even if you just clamped it down and
    ran around the outside with an endmill in a VMC you will have faces as square as the machine that made them. Any decent, modern VMC in any toolroom in
    good order should be within a few um per 300mm at least.

    When I had my cast iron axis beds machined the guy who did it did so on a 1500mm travel Okuma VMC about five years old. He did the works including drilling
    and tapping the holes for the linear rails. It was always my intention to get them ground after milling, but they were better than I could measure straight off the mill so
    I didn't waste my money. When I offered the rails up to the line of drilled and tapped holes I thought I'd have to clamp/straighten/re-clamp etc......but no. The holes
    were in such a straight line that fitting the rails was so easy. You've heard the saying 'straight as a die', well this is the machine that the die was made on. It only stands
    to reason that it will make a great job.

    It's easy to fall into the trap that granite squares are the last word in 'squareness therefore I must use granite'. BS any decent machinist with a decent machine will get
    a part that you or I would have difficulty in even measuring for square. I have spent many hours/week/months down that rabbit hole only to come to the realisation that
    a much cheaper, easier, direct solution is as close as my local machine shop, and that they consistantly produce highly acceptable parts, perhaps not at the same
    precision as granite, but good enough that I cant measure the differnce.

    How about this for an idea. If you want to use a component that has two very square faces what about a cast iron angle plate?. Here in New Zealand you can buy Chinese
    made cast iron angle plates reasonably cheaply, some of the bigger ones, say 300 x300 x200 are up around $300NZD ($200USD). Why not use a power hacksaw to slice
    one in two? Even a Chinese made square should be good to 10um over 300mm....good enough for a hobby machine while trying to conserve money.

    Craig
    Yes I think this would be great since cast iron has even better vibration dampening than steel, I think if I search I will be able to get exactly two thin ones at the size I need, the ones in the pic are Misumi so really good quality, cutting a bigger one into two would be more work but still easy.
    Attached Thumbnails Attached Thumbnails Screenshot 2022-07-09 at 13.14.01.png  

  12. #32
    Join Date
    Jul 2022
    Posts
    11

    Re: Small watchmaker CNC design idea

    I think with this Vertex angle plate and some risers I win more space and is much more simple to build, just for the frame with the granite base and big angle plate we are looking around 30 kg, I think is going to be really solid

  13. #33
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,

    Yes I think this would be great since cast iron has even better vibration dampening than steel,
    Don't get too hung up about damping. Materials like cast iron and granite have nice damping properties but what you need is STIFFNESS, it trumps everything else.
    What matter if a structure is well damped if it flexes under cutting forces or even its own inertial acceleration?

    Remember mild steel is four to six times stiffer than granite and double that of cast iron. The stiffness of the material just cant be beat.

    The only issue I see with the last sketch you posted is that the workpiece/vice cannot pass between the columns and therefore you will require a long cantilevered extension
    from the Z axis plane to the spindle. They are problematic at best. I have quite a long cantilevered extension and despite being 100 x 100 x 9 with welded 20mm flanges,
    heat treated and machined, its still the weakest part of my mill. I do like the great flexibility to handle large and odd shaped parts by having clearance to the column but
    it places extreme demands on the extension. Its a tradeoff, I think I made my extension longer than I should have. If and/or when I have the budget I'll get the I'll get an
    extension cast, and not in just grey cast iron, it will be at least SG iron, but more likely cast steel, it just has to be rigid, to hell with damping.

    Craig

  14. #34
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,
    I thought I would price getting plate steel laser cut per attached diagram;

    In 32mm medium tensile plate for the base and uprights and 20mm medium tensile for the gantry the price (including local tax) is $504NZD or 307.50Euro.

    I would expect to get the 20mm gantry faced, be it ground OR face milled, two faces of the 32mm uprights milled square and flat to each other, and the 32mm base to be face milled on the upper face only
    for $500NZD or less.

    That would make the total 615Euro. Weight of the parts is 48kg.

    Craig
    Attached Thumbnails Attached Thumbnails PlateCuttingSketch.png   PlateCuttingQuote.png  

  15. #35
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,
    I have compared 32mm steel to granite. The test piece is 400mm x 200 mm, cantilevered with a 100N vertical force.

    Steel deflected by 0.0178mm whereas a 50mm thick granite plate deflected by 0.0184mm, so very very similar. A 75mm granite plate however deflected by 0.0056mm, so about one third the
    deflection of either 50mm granite or 32mm steel.

    Simply put if your proposed granite plate is 50mm thick it will match 32mm steel but exceed steel if greater than 50mm.

    Craig
    Attached Thumbnails Attached Thumbnails Steel400x200x32.jpg   Granite400x200x50.jpg   Granite400x200x75.jpg  

  16. #36
    Join Date
    Jul 2022
    Posts
    11

    Re: Small watchmaker CNC design idea

    Quote Originally Posted by joeavaerage View Post
    Hi,
    I have compared 32mm steel to granite. The test piece is 400mm x 200 mm, cantilevered with a 100N vertical force.

    Steel deflected by 0.0178mm whereas a 50mm thick granite plate deflected by 0.0184mm, so very very similar. A 75mm granite plate however deflected by 0.0056mm, so about one third the
    deflection of either 50mm granite or 32mm steel.

    Simply put if your proposed granite plate is 50mm thick it will match 32mm steel but exceed steel if greater than 50mm.

    Craig
    Hi Craig, thanks again for your help and time, the laser cut steel seems like a good option and could make it exactly the size I need, however I don't have access easily to a big metal workshop or have the tools to deal with it myself, so I give priority to ease of assembly than ultimate performance sort to speak, right now I think Im pretty set on the idea of granite base, that I just found a brand that does 60mm thickness instead of 50mm, for the size 400x250mm, so I will go with that and add some cast iron risers and then the big iron angle plate, in terms of cost Im looking at 250 euro for the granite surface plate in grade 00, and around 220 for the big angle plate, so around 500 euros for a 30kg frame that I think will perform well and I can bolt it togheter in a weekend or less.

    Here is a quick model of how it looks with 400w servos.
    Attached Thumbnails Attached Thumbnails Screenshot 2022-07-13 at 02.19.25.png  

  17. #37
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,
    the trick now will be to maximise the travels with a dimensionally constrained base. As large an X and Y travel as possible is desirable but do not fall into the trap of
    not enough Z axis travel.

    In some cases you'll have flat materials mounted directly on the bed which will require the nose of the spindle go down to the table and in other cases you'll
    have parts in a vice, itself 50mm or so high, and the part another 75mm or so above that, so the nose of the spindle needs at least to be 125 plus the length of the tool
    off the table. The biggest limitation of my mini-mill was that if the part was too tall (with or without the vice) then you couldn't do the job. Don't skimp on the Z axis travel, you'll regret it otherwise.

    Craig

  18. #38
    Join Date
    Nov 2013
    Posts
    4280

    Re: Small watchmaker CNC design idea

    Hi,
    just as a quick guide if you place the two cars on one rail are 100 mm apart on centres, and the length of one car is 60mm (approx standard for 15mm linear guide cars).

    If the width of the gantry is 250mm then the linear rails can themselves only be 250mm long therefore the travel is:
    travel= 250 -(100+60)
    =90mm which is really pitiful.

    The ballscrews I linked to have a approx travel of 250mm, I'd imagine you'd want to exploit it all. Additionally the amount that the table 'tips under load' is reduced
    as the distance between the cars goes up. I'd suggest 150mm as a minimum. So the length of the rails would be:
    length of rails= 250 (desired travel)+ 150 (distance between car centres) + 60 (length of one car)
    =460mm

    Even reducing the distance between cars down to 100 mm would still require rails 400mm long to get 250mm of travel.
    The only other alternative is to reduce the travel....an unpalatable choice.

    Craig

Page 2 of 2 12

Similar Threads

  1. Fair strong DIY CNC machine (Wood and Aluminum) - Design Idea
    By Rabso in forum DIY CNC Router Table Machines
    Replies: 6
    Last Post: 11-21-2021, 12:55 AM
  2. Looking for an idea for a solution. Small space small travel
    By boydage in forum Stepper Motors / Drives
    Replies: 3
    Last Post: 08-31-2021, 08:38 AM
  3. RFQ: Estimate Needed On Small Steel Adapter WatchMaker?
    By mabmojo in forum Employment Opportunity
    Replies: 0
    Last Post: 12-26-2006, 08:34 PM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •