Servo CNC conversion of Vertex BS0 dividing head


Results 1 to 17 of 17

Thread: Servo CNC conversion of Vertex BS0 dividing head

  1. #1
    Member
    Join Date
    Sep 2008
    Location
    UK
    Posts
    229
    Downloads
    0
    Uploads
    0

    Default Servo CNC conversion of Vertex BS0 dividing head

    Actually, it isn't a real Vertex, it seems to be a clone from Axminster.

    I have seen a few of these with steppers attached, typically retaining pretty much all the original dividing gear and worm shaft bearings etc.
    Typically the stepper-powered versions are not being used for cuts under power, so the original worm thrust bearing arrangement is adequate. I want to use mine for hobbing gears, so many of the moves will be under cutting forces. For this reason I decided to convert to angular-contact bearings adjustable for zero backlash.
    It took me a while to find suitable bearings, but eventually I settled on 3801 bearings. These are double-row but I couldn't find single-row with the right thickness ratio. I am basically using them as single-row bearings.

    I also decided that I wanted the motor to stick out as little distance as possible from the head, and after a lot of head-scratching and CAD-fiddling I settled on a scheme that uses a hexagonal socket in the end of the worm shaft, and a hexagonal coupler on the motor shaft. In this setup almost the entire motor shaft is inside the worm shaft.

    First I modified the worm shaft, shortening it and reducing the diameter to suit the bearings. It was hardened, but a CBN tip dealt with that, and by the time I was down to finish diameter I was through the hardening.



    I then needed to make a hexagonal hole in the end.
    , but pressing in a machined-down cap screw head would probably work.
    Doing it on the lathe meant that I could have an 8.5mm hex, however. (I was tempted to use a pentagon or heptagon just for fun)


    I made a new version of the eccentric housing. The adjusting arcs need to be concentric with the eccentric
    spigot on the reverse side, not the motor mounting bore, which is why things look a bit squiffy.



    This image shows the two mating hexes disassembled:



    The finished article can be seen doing its thing on a rather dull Youtube video:



    Similar Threads:
    Last edited by andypugh; 01-24-2013 at 04:01 PM.


  2. #2
    Member aarongough's Avatar
    Join Date
    Feb 2007
    Location
    Canada
    Posts
    473
    Downloads
    0
    Uploads
    0

    Default

    Very cool! Looks like you did a really nice job on it!

    Gough Custom - http://goughcustom.com/


  3. #3
    Member samco's Avatar
    Join Date
    Jul 2003
    Posts
    1753
    Downloads
    2
    Uploads
    0

    Default

    Very nice work!

    sam



  4. #4
    Member handlewanker's Avatar
    Join Date
    Sep 2006
    Location
    Australia
    Posts
    6463
    Downloads
    0
    Uploads
    0

    Default

    Hi, you are still faced with backlash in the worm/wormwheel part.......this has been chewed over at great length with rotary tables and so far no solution to the backlash has risen......spring loaded double worm designs lead to accelerated bronze wormwheel wear.

    If you're only going to go one way (hook milling) you have less of a problem, but you will get backlash the moment you reverse the drive (climb milling)....it can be accounted for in the program, but is not the ideal solution as the worm drive needs to be slack enough to allow driving but firm enough to prevent the cutter pulling the wormwheel round when you reverse the drive and start climb milling.

    The same occurs with Acme thread drives where you have steel screws sliding on bronze nuts and in that mode you JUST CAN'T DO CLIMB MILLING without a hydraulic buffer.

    Production horizontal mills with air feed and hydraulic buffers are always used in the climb milling mode.

    A lot of work has been done with epicyclic cycloidal worm drives amomgst others.
    Ian.



  5. #5
    Member
    Join Date
    Sep 2008
    Location
    UK
    Posts
    229
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by handlewanker View Post
    Hi, you are still faced with backlash in the worm/wormwheel part.......this has been chewed over at great length with rotary tables and so far no solution to the backlash has risen.
    True, though this will mainly be used for indexing (and I may add a pneumatic actuator to the spindle clamp, the machine has air for the toolchanger already)
    The other use will be for gear hobbing, time will tell how much the backlash matters there, but as freewheel hobbing seems to very nearly work some of the time for some people, I am optimistic. :-)



  6. #6
    Member handlewanker's Avatar
    Join Date
    Sep 2006
    Location
    Australia
    Posts
    6463
    Downloads
    0
    Uploads
    0

    Default

    Freewheel gear hobbing......ahhh, first gashing the blank and then driving it with a hob that generates the tooth form....the blank running freely driven by the hob.....that's standard practice on a milling machine....can be horizontal or a verticle mill.

    If you intend to cut teeth on the blank, without first gashing it, then you need to gear the hob spindle to the blank spindle so that you don't finish round the blank with half a tooth.....the purpose of gashing in the first place.....this is done on a dedicated gear hobbing machine.

    CNC'ing the index spindle so that the hob and the blank can rotate independently, but in synchronisation, would not be affected by backlash as long as you had some mechanism to apply some braking effect to prevent the blank from bouncing around due to the backlash between index spindle and blank spindle....the stepper motor on the indexer would always be driving in one direction throughout the complete cut, and you also need to have a stepper drive to the hob ( main) spindle to have the hob and blank in synchro too.....highly impractical.

    The big problem will be if you lose a step or two in the index drive or the hob drive, as the blank, being under index control, cannot rotate freely when a step is missed.

    We are talking about hobbing...... the process where the hob, being a spiral type cutter, needs the blank to rotate at a certain ratio of gear teeth to hob pitch.

    However, if the purpose of the CNC'ing is to move the blank only by the index mechanism and a stepper motor, (and clamp it at each tooth cut), that is not hobbing but simple gear cutting with a single form gear cutter, with the blank being moved to the next tooth by the CNC program after the slide comes back when the last tooth was cut.

    That operation is CNC gear cutting, not hobbing.

    The gear is cut, as in a manual operation, by indexing the dividing head for each gear tooth, the CNC part would occur when the table cames back to the start and the div head is indexed round by stepper motor under CNC control so saving the tedious moving of the fingers that show where the index arm and detent must go in which hole in the hole plate....the whole operation can be fully automated (CNC control) for each tooth until the gear was fully cut, but you will need a set of gear cutters for each range of teeth from 9 teeth up to a rack, approx 8 cutter in a set.

    Hobbing generates the ideal tooth form with one worm shaped gear cutter that covers all gear teeth from 9 to a rack, in one particular tooth size.....a different hob for each tooth size or DP, Module etc.

    BTW, the stepper motor on the indexer has to accomodtate all hole combinations that are used to index all the tooth numbers, and must be in exact synchronisation with the stepper motor resolution of steps.

    Perhaps a simpler method would be to remove the manual index worm completely and drive the head spindle directly by stepper motor, like a 4th axis.

    If you were cutting a 100 tooth gear and you were out 1 step while driving the index arm by stepper motor, you would be out 100 steps when you got round the gear to the last tooth.

    You can still have an error driving the spindle directly under CNC control, if you cannot match the stepper motor resolution steps to the number of teeth you want to cut on the gear, as they are accuulative errors per step X teeth number.

    The problem lies in the fact that a dividing head is variable (by the hole plate pattern of holes) whereas a stepper motor resolution is fixed.

    Some gear ratios cannot be indexed without going to differential indexing and/or compound indexing....best of British.

    I may be totally wrong on this, so correct if applicable....I'd love to see it happen with CNC.
    Ian.



  7. #7
    Member samco's Avatar
    Join Date
    Jul 2003
    Posts
    1753
    Downloads
    2
    Uploads
    0

    Default

    andy has done this before on a smaler scale



    The system he is creating now is a servo - closed loop setup. No losing steps. May get a following errror - but this will be constantly corrected.

    Can't wait to see another video!

    sam



  8. #8
    Member
    Join Date
    Sep 2008
    Location
    UK
    Posts
    229
    Downloads
    0
    Uploads
    0

    Default

    samco has already pointed out that I have a setup for hobbing gears already, and this is part of an upgraded system I am putting together on a much beefier machine.
    However...
    Quote Originally Posted by handlewanker View Post
    You can still have an error driving the spindle directly under CNC control, if you cannot match the stepper motor resolution steps to the number of teeth you want to cut on the gear, as they are accuulative errors per step X teeth number.
    This can certainly be a problem if the indexing system uses integer arithmetic (as I think some of the standalone indexing drives do) but shouldn't be an issue when using floating-point arithmetic in the software layer.



  9. #9
    Registered
    Join Date
    Jan 2009
    Location
    USA
    Posts
    123
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by andypugh View Post
    This can certainly be a problem if the indexing system uses integer arithmetic (as I think some of the standalone indexing drives do) but shouldn't be an issue when using floating-point arithmetic in the software layer.
    Integers can be used to for decimal numbers by simply using multiplication. For instance 1.000 can be represented as 1000. for three decimal places. Floating point has issues since its rounded off. A floating point number has a mantissa to store the non exponent value. and generally has limited accuracy. Its really intended for handling numbers with very large swings, ie (very small numbers to very very large numbers).



  10. #10
    Member
    Join Date
    Sep 2008
    Location
    UK
    Posts
    229
    Downloads
    0
    Uploads
    0

    Default

    After a request by PM, I am attaching a drawing (PDF) and the 3D model (STEP)

    Attached Files Attached Files
    Last edited by andypugh; 09-28-2013 at 12:13 PM. Reason: Added IGES to the ZIP file


  11. #11
    Member
    Join Date
    Sep 2008
    Location
    UK
    Posts
    229
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by TechGuy5002 View Post
    Integers can be used to for decimal numbers by simply using multiplication. For instance 1.000 can be represented as 1000.
    The most reliable way to handle this in the case of keeping machine axes in step is to keep a count of encoder counts (if you use 64 bits you will basically never overflow) and to calculate the position of the slave axis directly from this every time.



  12. #12
    Member handlewanker's Avatar
    Join Date
    Sep 2006
    Location
    Australia
    Posts
    6463
    Downloads
    0
    Uploads
    0

    Default

    Hi, that sounds like a very complicated process for a simple job.

    There has to be two points to take into consideration here, and the first would have to be.....how many gears of a certain pitch, DP etc do you need/want to cut in a number of years to make it payable......you'll need a different hob for each gear DP type.

    The second point is, how much does a bought in commercially made HSS hob cost compared to a set of gear cutters, at 9 in a set, ranging from 10 teeth to a rack, and buying just the separate cutters as needed.

    For a few gears now and then it does not pay to invest in a hob.........it is more cost effective to cut a gear with a single gear cutter of the correct one from a set and a 4th axis or straight dividing head under CNC or manual indexing, and that is completely possible by anyone with a bit of mill knowledge.

    I personally have cut gears with a ground to form fly cutter when the gear cutter from a set was not available, and I have cut a number of racks with a hand ground cutter shaped like an engraving bit, but with a 29 deg included angle.

    Cutting racks is simple in a vice with an end mill type gear cutter shaped to a 29 deg included angle, but very difficult with any other method, and certainly not by turning the mill head over on it's side and using a circular gear cutter......a horizontal mill makes this easy, or a right angle drive in a vertical mill.

    I would go as far to say that any gear can be accurately cut with a hand ground cutter shaped to 29 deg, which is the rack form, and using a 4th axis to cut the form for any tooth number from 10 to a rack.

    The cutter is ground from a suitable diam HSS blank, like an end mill and has two cutting edges on the flat form.

    I've cut gears in plastic, brass and steel with a hand ground cutter in this form.

    BTW, I've also cut spiral gears in steel with a hand ground fly cutter, but that was with a dividing head geared to the table screw, once only in 50 years.
    Ian.



  13. #13
    Member
    Join Date
    Sep 2008
    Location
    UK
    Posts
    229
    Downloads
    0
    Uploads
    0

    Default

    Quote Originally Posted by handlewanker View Post
    Hi, that sounds like a very complicated process for a simple job.
    If you already have a spindle encoder and a rotary axis then the hobbing setup is just a few lines of code in the HAL file. (literally a few).
    The second point is, how much does a bought in commercially made HSS hob cost compared to a set of gear cutters, at 9 in a set, ranging from 10 teeth to a rack, and buying just the separate cutters as needed.
    I get the hobs second-hand from eBay. But to buy them new they cost about the same as a set of involute cutters.
    Gear Hobs - Arc Euro Trade



  14. #14
    Registered
    Join Date
    Sep 2007
    Location
    USA
    Posts
    300
    Downloads
    0
    Uploads
    0

    Default Re: Servo CNC conversion of Vertex BS0 dividing head

    Andy, I know this is a old thread but hopefully you still get notifications. I have the identical BS-0 head that I plan on cnc'ing. What size servo did you use? I have been wondering what size servo motor would be suitable for cutting while rotating. Also, now that you have had it a while, how did the backlash end up? Any other comments now that you have had it a while would be appreciated.

    George



  15. #15
    Member handlewanker's Avatar
    Join Date
    Sep 2006
    Location
    Australia
    Posts
    6463
    Downloads
    0
    Uploads
    0

    Default Re: Servo CNC conversion of Vertex BS0 dividing head

    Hi, one thing that has apparently escaped the thread watchers......if you cut gears it is most important to have the gear teeth running true to it's bore........one reason why cutting gears with a manual mill and dedicated single gear cutters is done with the blank mounted on a mandrel in a dividing head that is dialed to run true otherwise the gear will run unevenly as it rotates.

    The video previously screened shows a blank being cut without a hole to subsequently mount it onto a spindle.....whatever.........boring a gear after teeth cutting and to have it run concentrically with the pitch circle is going to be very difficult.

    While you may cut teeth of some description on a gear blank with this set-up, the outcome is not practical........it's purely cutting a gear by CNC for CNC's sake.

    BTW.....when cutting gears you need a gear tooth vernier to ascertain when the tooth thickness at the pitch circle is correct, otherwise the gears just fit where they touch........with that you may as well buy accurately cut gears from the experts who do gear cutting as a business.

    To ensure the gear comes out right the tooth thickness at the pitch circle has to be measured and the setting re-set for the cutter depth/tooth width............doing this with CNC is not CNC per se.
    Ian.



  16. #16
    Member
    Join Date
    Sep 2008
    Location
    UK
    Posts
    229
    Downloads
    0
    Uploads
    0

    Default Re: Servo CNC conversion of Vertex BS0 dividing head

    Quote Originally Posted by Sparky_NY View Post
    What size servo did you use? I have been wondering what size servo motor would be suitable for cutting while rotating.
    I used a NEMA23 sized servo, 24V and rated at something like 0.8Nm. I used it because I had it. It works OK for indexing, gear hobbing and engraving, but I think that it would stall with moderately heavy rotary cutting. If you expect to do a lot if that then I think you would eant something bigger.
    Also, now that you have had it a while, how did the backlash end up?
    Backlash is higher than I would ideally prefer. The limitation is how low I can get it in the tightest spot compared to the backlash in the loosest spot, because the work wheel isn't perfect. Your backlash in CNC-operation should end up the same as in manual. Get it as good as you can, then see if that it good enough.

    If you can find a decently big harmonic drive on eBay then that would probably be a better starting point, though I don't think that there is any way you could ever fit one to a BS0. You would have to think in terms of building from scratch.
    Here is one I designed earlier (but never made).



  17. #17
    Registered
    Join Date
    Sep 2007
    Location
    USA
    Posts
    300
    Downloads
    0
    Uploads
    0

    Default Re: Servo CNC conversion of Vertex BS0 dividing head

    Thanks much for the info Andy. Big harmonic drives would not be easy to come by or inexpensive unfortunately, even on ebay. Another route I am considering is a nema34 servo with a planetary reducer directly attached and then a robust timing belt drive to the spindle shaft. I would think that both the timing belt and the planetary drive would yield very low backlash.

    George



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
  •  


About CNCzone.com

    We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!

Follow us on


Our Brands

Servo CNC conversion of Vertex BS0 dividing head

Servo CNC conversion of Vertex BS0 dividing head