Established Methods For Sub 5 Micron Positioning Accuracy? - Page 2
Close
Login to Your Account
Page 2 of 2 FirstFirst 12
Results 21 to 27 of 27
  1. #21
    Join Date
    Feb 2011
    Country
    UNITED STATES
    State/Province
    Indiana
    Posts
    61
    Post Thanks / Like
    Likes (Given)
    35
    Likes (Received)
    27

    Default

    Quote Originally Posted by cameraman View Post
    Along those ^^^ lines @apt403
    DMGMori Heidenhain scales and magnescales use a magnetic encoded moiré effect read resolution of the order of 0.01 micron (ten nano meter ) but with a control system and real stuff to move + just physicality of the scale itself, and numerous thermal compensations including the scales, not just the machine - over one meter +/- two tenths +/- 5 micron is not atypical.
    I don't think Heidenhain ever made any magnetic scales. Correct me if I am wrong. Their high resolution models (LIF/LIP) are all glass, using interferential scanning principle to achieve high accuracy within 1 grating period.

    DMDmori now owns Magnescale, and they do have so called "laserscale". Those are also optical, using interferential scanning principle as well, and the grating period on their glass is impressive 400nm vs 20um on typical encoders. Wondering if that is what is going into the latest DMGmori machines.

    Quote Originally Posted by cameraman View Post
    https://arnd-sauter.com/wp-content/c...ra/VX-1000.pdf

    PDF ^^^ for Matsuura 3 axis c frame vertical VX 1000
    Gotta love the fine print: "* The measurement results are actual values but not guaranteed values". Nevertheless, very impressive performance for a real life machine.

    Quote Originally Posted by cameraman View Post
    Tighter tolerance relative moves can be achieved if it's just unidirectional more localized moves over 100 mm or 200 mm or so (most machinists know that or have practical experience of that
    Right on point! 1um over 100mm unidirectional move is relatively easy. Take it to 1000mm plus add reversal, and its exponentially harder to achieve. At micron level accuracy, most encoders have backlash or reversal error as well, which has to be taken into account.

    Abbe error can screw thing up really bad too. Workpiece is typically quite far from the encoder, so microradians of tilt can easily cause several microns of displacement

  2. Likes apt403 liked this post
  3. #22
    Join Date
    May 2009
    Location
    San Jose, CA
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    258
    Likes (Received)
    233

    Default

    It is not particularly hard to do this if you take the right approach, depending on the size of the required motion. I used MANY motion control systems in my career with 0.1 micron of accuracy and repeatability or better. These were in some cases microstep motors and in others servos. What is important is that you CONTROL the motion with a closed loop system which references a GOOD linear scale. For our most precise applications these were "Glass" scale encoders. Probably fused silica. In those days we used Heidenhain scales, ball screws, and a dedicated motion control card. In those days, the card plugged into a PC. Now-days, you can buy standalone controllers which read the scale and control the motor to keep position. The systems today are pretty sophisticated in terms of setting servo parameters and compensating for all manner of errors. Having excellent lead screws and nuts, good linear bearings, rigid and square construction, watching out for cantilevered loads, etc, etc is all valuable, but it is the closed loop control of a measured displacement that is the key. Counting steps on a stepper is JUST NOT THE WAY TO GO if real accuracy is desired. For extreme precision, interferometers are necessary. There exist phase measuring interferometers (since the 1990s or even earlier- see Hewlett Packard) that measured to nanometers over long distances, and corrected for barometric pressure and temperature.

    Michael

  4. Likes apt403 liked this post
  5. #23
    Join Date
    May 2015
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    278
    Post Thanks / Like
    Likes (Given)
    126
    Likes (Received)
    149

    Default

    Much appreciated, everyone. So, the trend I'm seeing is, much past this 5 micron level, screws and scales, while possible to use, become increasingly impractical solutions to actually implement. By 1 micron, over any appreciable distance, you're looking at tightly controlling... Essentially every parameter that could have an effect.

    @jim rozen: Without giving away details that aren't mine to disclose - Think small, specialized VTL. The devil is in the details, as they say. But practically, it's a VTL that also has other bits and bobs bolted to it for different operations.

    - Low power: 2HP spindle motor (10k RPM through timing belts).
    - Z axis travel is 800mm (32"), X axis is around 400mm (16").
    - Granite machine frame/base
    - Weight is constrained to 500kg (1100lb). I'd like more mass to play with, but hey, I just build 'em.
    - Accuracy requirements are +/- 15 micron (6 tenths) over 400mm (16").*

    *The 800mm Z travel is a clearance/ease of loading/unloading thing. Actual work will take place 0-400mm from the spindle mounting face.

    @Limy Sami: LOL. I narrowly avoided spilling coffee on my keyboard w/ that one.

    @CarbideBob: Abbe error! That's a solid point that I have not accounted for as of yet. Thank you.

    @zsinstruments: Re environmentals - the building this will be installed in is climate controlled, but not to a metrological degree of precision. Machine enclosure will have provisions for thermal isolation.

    @Halcohead: Thanks for that! I badly need to hunt down a copy of that book, apparently.

    Since we're on the subject of the actual project: To hit the +/- 15 micron requirement, my thinking is scales + screws. While the machine base is now a long lead time type situation, as long as UPS keeps delivering through this plague, I plan to bolt all the linear motion stuff to a surface plate for testing. Things like "Should I thermally isolate the scales?" "Should I be planning to keep the whole assembly above ambient?", making sure my motor size calculations are correct for the acceleration requirements, etc, etc I feel warrant some testing.

    Does anyone have any recommendations for practical info on machine damping? It seems to be a bit of a black art. Currently at the "well, granite is pretty damp. and 500kg of the stuff shouldn't ring too bad with this tiny spindle motor" level.

    Also, CONTROLLERS. So far, I've spent most of my day trying to assess options there. Galil's DMC-40x0 or DMC-41x3 controllers are very feature dense, and the price is certainly right, but I'm not interested in writing my own front end. Siemens and Bosch both have nice options, but I'm not sure about costs, as of yet. Requirements are:

    - Non-trapezoidal motion profiles (Jerk and jounce control is important)
    - Dual encoder inputs w/ a somewhat easy method of setting up the cascading PID loop
    - Some (nearly) out of the box solution for an operator front end
    - A working post existing for Inventor's CAM.
    - Friendly to a non-volume customer (I highly doubt more than two or three of these will be built)
    - Sub $5k <- how practical this is, I don't know.

    The non-trapezoidal motion is what's tripping me up. Honestly, a 'hobby' level package like LinuxCNC has all the flexibility to handle my application, but can't do anything but trapezoidal.

    There are some funky closed loop servos (ClearPath, etc) out there that handle the motion control internally (just needing step/dir or velocity input), but they're not provisioned for another encoder. In theory, I could use an external motion controller to handle the positioning part of the loop (where the linear encoder is providing feedback) and let something like a ClearPath handle the velocity part. There will be following error since the external controller isn't expecting an acceleration profile to be applied. It sounds janky, and probably is. But I mean, I like the project manager. I don't want the poor bastard to vomit blood when he sees an invoice from Rexroth for an Indramat.

  6. Likes barbter liked this post
  7. #24
    Join Date
    Jun 2015
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    526
    Post Thanks / Like
    Likes (Given)
    1090
    Likes (Received)
    183

    Default

    Quote Originally Posted by apt403 View Post
    - Sub $5k <- how practical this is, I don't know.
    Your project sounds fun. Yea, that will be difficult but I wish you the best of luck in the project and please keep us updated. I say that with the best intentions and hope you achieve your goals. You're at least asking the right questions!

    Quote Originally Posted by Robert R View Post
    An extreme example of what is possible can be seen at the Precitech or the Moore Nanotech websites. The products are lathes with nanometer accuracy. The machines at both companies were designed and built by the same person.
    I'd love to read/hear more about this person if you have any links.

  8. #25
    Join Date
    Aug 2005
    Location
    Raymond , CA
    Posts
    295
    Post Thanks / Like
    Likes (Given)
    0
    Likes (Received)
    197

    Default

    Quote Originally Posted by BugRobotics View Post

    I'd love to read/hear more about this person if you have any links.
    There are a number of articles written about the precision machine tool industry in Keene New Hampshire. There is one history that describes the sequence of events in detail. I cannot find it at the moment.

    The history goes something like this:

    The founder of the industry is Donald Brehm. He graduated from Rensselaer U. in 1957 and went to work for Allied Signal. Allied Signal was not interested in marketing the precision air bearing spindle he developed while employed there. He left in 1962 and started Pneumo Precision. The company's financial backers eventually forced him out. They wanted to sell the company and take the profits. Allied signal then bought Pneumo Precision from the backers in 1984. The company name changed to Rank, then to Taylor Hobson. In 1997 Schroder Ventures buys this company along with the Precitech company described below.

    Donald Brehm started a second company called Toolroom Craftsmen. There was a non compete agreement with Allied Signal. He was sued for manufacturing and selling a lathe that was designed for a different industry. Allied Signals lawyers were more expensive that his and the business was put on hold until the non compete agreement expired. The Toolroom Craftmen company was renamed Precitech in 1992. Precitech was sold to Schroder Ventures in 1997, then to American Capital in 2002 and then sold again in 2007 to Ametek. The sale was driven by the needs of the investors in the company.

    The Moore Nano Technology company was founded in 1997. The connection is through one of Brehm's former employees.

    These are the first documents that showed up on the internet search. A number of details have been left out.

    De/Reindustrialization of Keene – Pneumo Precision

    Brehm’s Pneumo to today’s Precitech – Industrial Changes and its Impact on a Community

    If you have the interest and time a search using some of the keywords above should provide the detailed story.

    The moral of the story is this:
    If you need to finance a new business you are better off robbing a bank rather than dealing with venture capitalists.

  9. #26
    Join Date
    Jun 2015
    Country
    UNITED STATES
    State/Province
    Colorado
    Posts
    526
    Post Thanks / Like
    Likes (Given)
    1090
    Likes (Received)
    183

    Default

    Quote Originally Posted by Robert R View Post
    There are a number of articles written about the precision machine tool industry in Keene New Hampshire. There is one history that describes the sequence of events in detail. I cannot find it at the moment.

    The history goes something like this:

    The founder of the industry is Donald Brehm. He graduated from Rensselaer U. in 1957 and went to work for Allied Signal. Allied Signal was not interested in marketing the precision air bearing spindle he developed while employed there. He left in 1962 and started Pneumo Precision. This company expanded to include the production of diamond turning lathes. The company's financial backers eventually forced him out. They wanted to sell the company and take the profits. Allied signal then bought Pneumo Precision from the backers in 1984. They in turn sold the company to Ametek in 2007.


    Donald Brehm then started a second company called Toolroom Craftsmen. There was a non compete agreement with Allied Signal. He was sued for manufacturing and selling a lathe that was designed for a different industry. Allied Signals lawyers were more expensive that his and the business was put on hold until the non compete agreement expired.

    The Moore Nano Technology company was founded in 1997. The connection is through one of Brehm's former employees.

    This was the first document that showed up on the internet search. A number of details have been left out.

    De/Reindustrialization of Keene – Pneumo Precision

    If you have the interest and time a search using some of the keywords above should provide the detailed story.
    Thanks for the lead.

  10. #27
    Join Date
    Dec 2003
    Country
    UNITED STATES
    State/Province
    Washington
    Posts
    4,751
    Post Thanks / Like
    Likes (Given)
    2284
    Likes (Received)
    4214

    Default

    APT - we have done linear positioning with system resolution / accuracy to +/- 10 picometers using a very expensive scale from Heidenhain that is not sold to the general public (and the foundation, isolation, temperature control, and done in a vacuum) all added to the complexity - in comparison, what you are trying to accomplish is fairly easy.

    We have also done large gantry mills that would hit those specifications (moving X-Axis varied from 65,000 lbs to over 135,000 lbs) with from 60 - 80 feet of travel using Reneshaw laser interferometers . . . then created 5-Axis compensation tables to correct for machine sag as a function of axis position.

    What you use for a controller is also important - if this machine is intended to use G-Code - I would suggest a Siemens 840D sl - it can also provide notch filters to dampen any mechanical resonances in addition to managing cross axis compensation for any mechanical non orthogonal axes due to mechanical issues.

    The built in controller based dampening features allow each axis current loop to be tuned to avoid the natural frequency associated with that axis, then the controller analyzes the entire machine for any remaining resonances as a function of axis position and machine variations in stiffness and it applies a jerk filter to the trajectory generator to ensure that the controller does not excite those frequency domains while in operation.

    I would be happy to help guide your journey off line but I can't do it for free.

  11. Likes BugRobotics, Jmaks, sfriedberg liked this post

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
  •