In process grinding measurement - Page 3
Close
Login to Your Account
Page 3 of 3 FirstFirst 123
Results 41 to 52 of 52
  1. #41
    Join Date
    Aug 2005
    Location
    Raymond , CA
    Posts
    259
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    176

    Default

    Quote Originally Posted by CarbideBob View Post
    In reality given these controls why are you not in the couple of microns in test? The machine and the feedback are.
    Bob
    The rotary table has out of plane run out from the bearings and elastic deflection from the wheel head load. The continuous monitoring of the chuck height by the Marposs gage is necessary to keep track of the height of each of the 25 parts on the table. The software monitors the 25 heights and stops the grinding process when all 25 parts are within the tolerance window. If grinding where allowed to continue until spark out the thinnest parts of the set of 25 would be out of tolerance.

    Maintaining size on a batch of 25 parts simultaneously is difficult. That is why double disk grinders and double disk lapping machines were developed. The tolerances that Motion Guru's customer are trying to achieve would be easy on a rebuilt 50 year old double disk grinder. Piston ring manufacture is one example of this.

    The project is a attempt to use engineering, software, and automation to correct what was originally a management mistake. My experience is that this seldom ends well for the engineers involved in the project.

  2. Likes motion guru liked this post
  3. #42
    Join Date
    Dec 2003
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    4,673
    Post Thanks / Like
    Likes (Given)
    2187
    Likes (Received)
    3990

    Default

    Robert is spot on with his note of machine deflections - this machine is as floppy as a noodle, and yes the engineering / software / technology deployed is making up for what is really a poor machine design in the context of tolerance goals.

    As for second op grinding, the chuck is rinsed with high pressure clean coolant after degausing, then wiped down, and parts are cleaned after first op before being placed for second op. Its fussy, but that is what the customer is accustomed to doing.

    At one point, we fixtured a cross bridge between chucks with a large high quality industrial steering caster mounted to the bridge. The caster wheel was riding on the chuck where the parts are placed and a downward force was imparted on the bridge with the spindle carriage to simulate loading.

    We had dial indicators mounted everywhere including the underside of the chuck gearbox and carriage. The spinning chuck operated as normal and we took before, during, and after measurements with video running of the dynamic deflections of the chuck based on localized differences in deflection due to non-uniform chuck mechanical design.

    Head nodded back 0.003 inches, chuck deflected at the caster location - 0.007 inches, the underside of the chuck gearbox deflected down 0.0015 inches on the side nearest loading of the chuck and it lifted 0.001 inches on the opposite side. The grinding wheel flange deflected upward 0.002 inches. It was a mess!!

    That is what convinced us to go with the Marposs setup that measured the difference between the part and the chuck surface local to the part.

  4. Likes michiganbuck liked this post
  5. #43
    Join Date
    Aug 2007
    Country
    UNITED STATES
    State/Province
    Wisconsin
    Posts
    2,069
    Post Thanks / Like
    Likes (Given)
    1099
    Likes (Received)
    927

    Default

    Quote Originally Posted by motion guru View Post
    Robert is spot on with his note of machine deflections - this machine is as floppy as a noodle, and yes the engineering / software / technology deployed is making up for what is really a poor machine design in the context of tolerance goals.

    As for second op grinding, the chuck is rinsed with high pressure clean coolant after degausing, then wiped down, and parts are cleaned after first op before being placed for second op. Its fussy, but that is what the customer is accustomed to doing.

    At one point, we fixtured a cross bridge between chucks with a large high quality industrial steering caster mounted to the bridge. The caster wheel was riding on the chuck where the parts are placed and a downward force was imparted on the bridge with the spindle carriage to simulate loading.

    We had dial indicators mounted everywhere including the underside of the chuck gearbox and carriage. The spinning chuck operated as normal and we took before, during, and after measurements with video running of the dynamic deflections of the chuck based on localized differences in deflection due to non-uniform chuck mechanical design.

    Head nodded back 0.003 inches, chuck deflected at the caster location - 0.007 inches, the underside of the chuck gearbox deflected down 0.0015 inches on the side nearest loading of the chuck and it lifted 0.001 inches on the opposite side. The grinding wheel flange deflected upward 0.002 inches. It was a mess!!

    That is what convinced us to go with the Marposs setup that measured the difference between the part and the chuck surface local to the part.
    I have been keeping an eye on this thread. Sure am interested to see how production runs go. I know you have been at this project for a while, I sure hope it works out. The post here about all the machine deflection would concern me. Marposs will bring you to size but I don't know how it will make up for a weak/loose machine.

    A previous post showed the parts on a table with 2 probes. Is this all the customer wants to load the table up with parts? I see that the constant on and off parts will be very hard on the probes. I know this was an issue years ago with marposs. Hopefully they have made some improvements.

    Maybe that is just for testing and in production they will fully load the table.

    Keep the posts coming and good luck!

  6. Likes michiganbuck liked this post
  7. #44
    Join Date
    Jun 2012
    Country
    UNITED STATES
    State/Province
    Michigan
    Posts
    10,050
    Post Thanks / Like
    Likes (Given)
    3387
    Likes (Received)
    3592

    Default

    We had deflection due to a super hard wheel needed on certain a job.. Is that the case here.

  8. #45
    Join Date
    Dec 2003
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    4,673
    Post Thanks / Like
    Likes (Given)
    2187
    Likes (Received)
    3990

    Default

    Hard parts, interrupted loading/probing. Again, Marposs has been involved since the end of last year and has reportedly done this kind of interrupted gauging as a standard part of their business.

    Grind aesthetics and part thickness are #1 priority with the customer. They do high mix low volume parts so batch to batch of the same part of more than 50 parts (2 - 4 complete grind cycles depending on parts) is rare. The machine is parameterized by part SKU from right chuck to left chuck as well so that when it is grinding on one side, the operator can be setting up on the other side for a completely different part seamlessly with production.

    Right side uses an acoustical sensor to get within +/- 0.001 inches for first side grind (again with all the deflections), and Left side uses the Marposs sensor for final side grind.

  9. Likes michiganbuck, cash liked this post
  10. #46
    Join Date
    Aug 2005
    Location
    Raymond , CA
    Posts
    259
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    176

    Default

    The variation in height of the work pieces coming off the grinder will follow a bell shaped curve. The center location of the curve will be controlled by the Marposs program. The width of the curve will be determined by the design and condition of the grinding machine. It may be that the least expensive option will be to discard the out of tolerance parts rather than attempt to narrow the width of the curve by mechanical retrofits to the grinder.



    For those PM readers who have not grown tired of my off topic responses I have included a Ebay listing of a old and most likely worn out horizontal table (vertical spindle) double disk grinder of about the right size for Motion Guru's customer. This machine is uncovered and illustrates the basic design of the disk grinders..

    This is a good example of a stiff machine design . The photographs show a machine with heavy castings with no slides to wear out. The grinding wheel feed is along the spindle axis. The spindles are supported on quills in the upper and lower castings.The quill positions are controlled by lead screws driven by a reduction gear train. The part feed table is non magnetic and is used only for funneling the parts into and out of the grinding wheels. The wheel head motors are at the back of the machine two feet away from the spindles The design keeps the largest source of heat away from the machine frame. This machine does require a full time operator. That would not be a problem if part batches are small and would avoid the cost of making a custom part feeder required for a production machine. Many of the used, for sale, double disk grinders in the U.S. are old Gardner machines some of which have extensive upgrades including ball screws, servos, encoders and CNC controls On the older machines the acme lead screw is the first component to wear out and need replacement.


    DISKUS WERKE DDS457R Verical Double Disc Opposed Face Production Grinder | eBay

    Or if cost is not a concern one can buy a newer but still old (1995) Gardner grinder.
    The marketing article describes material removal rates, accuracy, surface finish, typical parts, and the application of the vertical and horizontal machine geometries. There is a a mention of the use of CBN wheels for fast metal removal rates and the use of polishing wheels for producing mirror finishes.


    Double-Disc Grinding On The Move :


    Modern Machine Shop



    Gardner was purchased by Western Atlas which was purchased by Landis which was purchased by Fives Group.
    This is what happens when managers have nothing useful to do.
    Last edited by Robert R; 03-26-2019 at 03:09 PM.

  11. #47
    Join Date
    Sep 2010
    Location
    Oklahoma City, OK
    Posts
    4,299
    Post Thanks / Like
    Likes (Given)
    746
    Likes (Received)
    1792

    Default

    Looks familiar.

    Sent from my SM-G930R4 using Tapatalk
    Last edited by Cole2534; 03-27-2019 at 09:30 AM.

  12. #48
    Join Date
    Dec 2003
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    4,673
    Post Thanks / Like
    Likes (Given)
    2187
    Likes (Received)
    3990

    Default

    After a week of testing and refinement - and a few additional problems solved we have the following.

    • Chuck stylus is now a brass foot with 0.030 inch wide grooves cut in the surface to prevent hydroplaning
    • All Marposs data is now routed to the Grinder control and HMI so that actual part thickness as measured by the Marposs system is available to the controller in real time. Prior to this, we were getting digital outputs that indicated end of rough grind and end of finish grind transition points.
    • Marposs setup parameters are now stored in the grinder control and automatically downloaded to the Marposs unit with part flow through the machine.
    • Material removal rate is tracked and sudden increases in part thickness while grinding halt the process and put an alarm up on the HMI instructing the operator to clean the chuck stylus.


    So far the system is showing promising results - our batch to batch thickness consistency is +/- 0.0001 inches by chuck location and our in-batch consistency is +/- 0.0004 inches. We need to now experiment with our spark out duration to see what effect that might have on in-batch consistency.

    Despite efforts to get the chuck ground as flat as possible, the chuck surface deflects more or less depending on the location on the chuck surface where the part is located. This appears to correspond to internal chuck casting rib location and internal passage ways for magnet wiring.

  13. #49
    Join Date
    Apr 2018
    Country
    UNITED KINGDOM
    Posts
    2,157
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    991

    Default

    Quote Originally Posted by Robert R View Post
    Gardner was purchased by Western Atlas which was purchased by Landis which was purchased by Fives Group.

    This is what happens when managers have nothing useful to do.
    That made me smile

    But to their credit, a friend with a Landis cam grinder says that Fives has been very good with service and support. Landis was also but at a million five, you'd expect that. For something ten years old, a lot of takeover specialsts are not. Fives has been good.

  14. #50
    Join Date
    Aug 2007
    Country
    UNITED STATES
    State/Province
    Wisconsin
    Posts
    2,069
    Post Thanks / Like
    Likes (Given)
    1099
    Likes (Received)
    927

    Default

    Quote Originally Posted by motion guru View Post
    After a week of testing and refinement - and a few additional problems solved we have the following.

    • Chuck stylus is now a brass foot with 0.030 inch wide grooves cut in the surface to prevent hydroplaning
    • All Marposs data is now routed to the Grinder control and HMI so that actual part thickness as measured by the Marposs system is available to the controller in real time. Prior to this, we were getting digital outputs that indicated end of rough grind and end of finish grind transition points.
    • Marposs setup parameters are now stored in the grinder control and automatically downloaded to the Marposs unit with part flow through the machine.
    • Material removal rate is tracked and sudden increases in part thickness while grinding halt the process and put an alarm up on the HMI instructing the operator to clean the chuck stylus.


    So far the system is showing promising results - our batch to batch thickness consistency is +/- 0.0001 inches by chuck location and our in-batch consistency is +/- 0.0004 inches. We need to now experiment with our spark out duration to see what effect that might have on in-batch consistency.

    Despite efforts to get the chuck ground as flat as possible, the chuck surface deflects more or less depending on the location on the chuck surface where the part is located. This appears to correspond to internal chuck casting rib location and internal passage ways for magnet wiring.
    To help with in-batch consistency- you could try when you are at or very close to size and when ready to go to sparkout, back your head off to relieve grind pressure, then start feed again but at a slower rate down to size then begin spark out.

  15. Likes michiganbuck, motion guru liked this post
  16. #51
    Join Date
    Dec 2003
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    4,673
    Post Thanks / Like
    Likes (Given)
    2187
    Likes (Received)
    3990

    Default

    Just to close the loop on this - we wound up @ hitting +/- 0.0003 inches with Cpk of just over 1.1 . . . we lost our butts on this project to a significant degree but the customer is happy and considering more machines which we will work with them on pricing to see if we can come out whole in the long run.

    Next time we rebuild a 30 year old grinder with a tight tolerance requirement, I will spend more time looking at the specifications that we must hit and the actual performance of the machine in its clapped out condition as well as what they were capable of when new. We learned a lot and developed a lot of software and mechanical design innovation on this machine - we will need to do 4 or 5 more before it becomes a profitable venture.

  17. Likes cnctoolcat, memphisjed, bryan_machine liked this post
  18. #52
    Join Date
    Sep 2010
    Location
    Oklahoma City, OK
    Posts
    4,299
    Post Thanks / Like
    Likes (Given)
    746
    Likes (Received)
    1792

    Default

    Thanks for the followup. I was looking at one of their products this week and it made me wonder what became of it.

    Sent from my SM-G973U using Tapatalk


Tags for this Thread

Bookmarks

Posting Permissions

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