Best practice: vertical axis DRO scale on a surface grinder
Close
Login to Your Account
Results 1 to 8 of 8
  1. #1
    Join Date
    Sep 2011
    Country
    GERMANY
    Posts
    3,751
    Post Thanks / Like
    Likes (Given)
    2534
    Likes (Received)
    1538

    Default Best practice: vertical axis DRO scale on a surface grinder

    Dear Metrology group,

    I have a question about mounting a vertical axis scale/encoder on my surface grinder.

    I'm adding a digital readout (DRO) to the vertical axis of my 1986 Jones and Shipman 540APR surface grinder. I'm the second owner of the machine, which was apparently only used for a few months by the original owners, and then shoved in a corner and forgotten for the following three decades. The ways on the machine are pristine, and I've mapped the grinding footprint (150 x 450mm) comparing it to a slightly larger surface plate. Within this rectangle the total deviation from planar is at most 2-3 microns (and might be better, since that's comparable to the accuracy of the surface plate). I've had the machine for several years and really like it. The only shortcoming of the machine is that it doesn't have a DRO.

    So I'm adding a three axis DRO to the machine. The long and cross axes are easy enough. But for the vertical axis I'm trying hard to do it as well as possible. I'm more interested in repeatability and resolution than in absolute accuracy. Total range of vertical motion is 280mm. Total cross section available is about 12 x 60mm, which is not enough for a sealed encoder.

    I'm using Renishaw parts for an open non-contact system: an RGH25U optical read head, an RGB25Y interpolater, and RGS20-S scale tape. I've used this system in the past for the long axis of my cylindrical grinder, and it works very well. The tape is 6x0.2mm, with a 20 micron period, accurate to +- 3 microns/meter and within +-0.75 microns in any 60mm section. The interpolator provides 10 counts per micron, so a 0.1 micron resolution. That makes sense here: the machine hand wheel scales have 1 micron divisions, and with care, it is possible to work to micron precision. So another order of magnitude in the resolution is justified.

    I'm mounting the scale inside the machine, almost directly over the vertical lead screw, and close to halfway between the two vertical guide rails. The measuring tape is attached to a strip of annealed ground O2 steel (thermal coefficient ~11.2e-6/K) 50 x 10 x 360mm, which in turn is mounted to the cast-iron carriage (thermal coefficient ~10.7e-6/K) that carries the grinding spindle and motor. The 50x10x360 strip is attached only at the top and bottom ends, not along its length or in the middle. The optical read head is mounted to the cast-iron machine chassis with steel blocks, bolted in place over an epoxy "shim".

    Because of this mounting location, and the slow change in machine temperature in that area, I expect that the O2 carrier strip and RGS20-S measuring tape will be in good thermal contact, and will have a temperature similar to the cast-iron body of the machine. The RGS20-S tape is "mastered" to the O2 strip (glued down at both ends) so will expand and contract with that strip. However there will be some differential expansion between the O2 strip and the cast-iron spindle/motor carriage. Probably this will be less than 5 microns but might be as much as 20 microns if the carriage and scale differ by 5C in temperature.

    My question concerns the attachment of the O2 strip to the cast-iron carriage. I can either fix the strip (secure datum point) at one end, and leave the other end free to "float", or I can fix the strip at both ends. I am about 90% sure that the first choice would be best, in part because I have no good way to attach a 10 x 50 mm cross section strip well enough to ensure that the strip rather than the fasteners would deform. My question, how to decide between fixing the strip at the top or at the bottom? Is there a better choice? Mechanically, either is OK. The top of the strip is the region that is in use when the grinding head is near to the grinder table. This might argue for the top being the secure datum point, and the bottom of the strip being allowed to float. But I'm not sure.

    I can provide some photos if that would help.

    Cheers,
    Bruce
    Last edited by ballen; 10-22-2021 at 11:23 AM.

  2. #2
    Join Date
    Sep 2011
    Country
    GERMANY
    Posts
    3,751
    Post Thanks / Like
    Likes (Given)
    2534
    Likes (Received)
    1538

    Default

    I finished the hard part of the installation yesterday, and this is now working well. I got no feedback for my first post, so decided to fix the top of the scale (area that is used when grinding thin parts) and to leave the bottom of the scale "floating". Here are a few photos for the next person doing something similar.

    The parts I am using are from Renishaw:

    RGH25U non-contact optical read head (produces an analog signal)



    RGB25Y interpolator (produces an RS422 incremental digital quadrature signal). The output from this goes directly to the SINO DRO SDS6 display, which has RS422 inputs.



    RGS20-S scale material (gold-colored strip) which has a 20 micron period. I really like this stuff. It makes it possible to built an accurate open high-resolution optical encoder at reasonable cost, which occupies very little space. The only constraint is that you need to find a location which provides good protection from dust, dirt and coolant.



    There are only three parts: the scale carrier strip, the metal T that carries the head, and the block that carries the metal T. All three are visible in the photo above, and are made from steel, to more-or-less match the temperature expansion coefficient of the machine itself. All of these are stiff and chunky to reduce motion from vibration.

    The block that carries the T is bolted to the machine with two M6 SHCS, with a metal-filled epoxy shim between it and the cast-iron chassis of the machine. Those bolts are not visible: they are underneath the metal T, and angled 10 degrees outwards so that I could drill and tap them with hand tools. If you look closely, you can see the fillet of gray J-B Weld epoxy to the left of the block, which squeezed out during installation.

    The scale is located almost directly over the vertical lead screw, which attaches to the lower part of the carriage, shown below. The two round vertical rails that guide the carriage and make the vertical ways are on the left and right of this carriage, so the scale is close to halfway between those.



    The readhead must ride 0.8mm+-0.1mm above the scale. By tuning the four M4 corner screws, I was able to get this within 10 microns over the entire scale length. Then I injected a metal-filled epoxy shim (J-B Weld) behind top of the scale, so that the M6 mounting screws could be torqued down. This shim is very important, because the cast-iron carriage beneath the scale carrier has a rough and uneven surface. The shim provides a uniform level support, so that I can torque down the bolts hard without distorting the scale carrier.



    The bottom of the scale carrier is retained with screws that have lockwashers under them. These screws are not torqued, so that the bottom of scale carrier can move with expansion and contraction.



    I was happy and surprised to see how stable the measurements are: even with all of the hydraulics and spindle running, the 0.1 micron digit (4 millionths of an inch, about 1/5 the wavelength of visible light!) is stable. The combination of the scale location and machine design help a lot with this. I can push down hard on the wheel housing without shifting that final digit!

    I still need to glue on the top and bottom "end clamps" that fix the RGS20-S scale material to the substrate. This ensure that the scale material expands and contracts with the substrate. I'll wait a few days before I do that, to ensure that the scale has settled after installation. I'll also attach some soft vibration-damping material (self-adhesive bitumen sheet, used for auto-body damping, about 3mm thick) to the back of the scale carrier, in the area around the pulley.

    Note: the entire metal scale carrier (with scale attached) can be removed and replace via the 4 screws that hold it to the carriage, so I can still access the pulley and spindle behind them. I coated the back of the scale carrier with mold release compound before injecting the epoxy shim behind it.

    Renishaw make a small magnetic reference mark indicator, which provides an electronic reference mark at maximum resolution (here 0.1um). In principle this could be added, but I decided that it didn't make sense here, since every time you dress the wheel, the zero changes.

  3. #3
    Join Date
    Dec 2017
    Country
    UNITED KINGDOM
    Posts
    76
    Post Thanks / Like
    Likes (Given)
    129
    Likes (Received)
    33

    Default

    Nice job. Seems you have it well thought out. Looks like a lovely machine in good condition.

  4. Likes ballen liked this post
  5. #4
    Join Date
    Sep 2011
    Country
    GERMANY
    Posts
    3,751
    Post Thanks / Like
    Likes (Given)
    2534
    Likes (Received)
    1538

    Default

    Quote Originally Posted by Renardiere View Post
    Nice job. Seems you have it well thought out. Looks like a lovely machine in good condition.
    Thanks. I have never tried to do a DRO with better than 1 micron resolution before, so I did research and thought about this carefully before proceeding. The machine is wonderful, here's a photo taken some time ago, after I had fixed an oil leak and was inspired to clean it up.



    One of these days I'll have to do another "cover shot" with the DRO.

    Since I wanted to ensure that the scale extended to the very bottom of the vertical axis travel, I had to take off the chuck. Fastest and easiest way was to remove the entire table and chuck together, which only requires taking off four screws for the cylinder rods, and unplugging the chuck.





    I bought this machine within seconds after seeing a photo of the ways (the table ways are equally good).

  6. Likes Mark Rand, ratbldr427, TNB, gregormarwick liked this post
  7. #5
    Join Date
    Sep 2011
    Country
    GERMANY
    Posts
    3,751
    Post Thanks / Like
    Likes (Given)
    2534
    Likes (Received)
    1538

    Default

    Here's how it looks with the DRO. Front view:



    and back view



    The interpolator and cabling are inside the column, to the left of the upper access panel. Behind that panel is the gold-colored optical scale and readhead.

    I'm planning to add a 1 micron resolution encoder to the long axis (490mm total travel). Because that environment is dirty (coolant spray and splash) I am going to use a magnetic scale made by RLS (an affiliate of Renishaw) called the HiLin. The scale is 6 x 18mm in cross section, and is completely encapsulated inside a welded stainless steel cover. The magnetic read head (rated IP67) is non-contact and "flies" 0.2mm over the surface of the scale. I should be able to fit it on the rear of the table, where there is a ~12mm gap between table and column. If you look at the photo just above, it will go to the right of the "90-degree elbow drain fitting" attached to the rear of the table. That area is not quite flat enough for the scale, so I'll pull off the table, bolt it to the mill, and machine a flat seat for the scale. I will probably epoxy a 2mm thick aluminium piece onto the table under the area where the scale will attach, and will then machine that flat. That way the rear side of the magnetic scale will be a little bit more distant from the (ferromagnetic) cast iron table. That should help maintain the magnetic field strength of the scale and its long-term accuracy.

    Edit: I was asked offline about the planned location for the RLS magnetic scale. This is shown below, in red:

    Last edited by ballen; 11-02-2021 at 08:09 AM.

  8. #6
    Join Date
    Dec 2014
    Country
    ESTONIA
    Posts
    82
    Post Thanks / Like
    Likes (Given)
    34
    Likes (Received)
    15

    Default

    I am skeptical regarding all kind of tape scales. Not sure, how much they stretch during gluing. I have some experience with installation lika and sony tapes, they seem too soft and unstable to be precise near 1 micron.

  9. #7
    Join Date
    Sep 2011
    Country
    GERMANY
    Posts
    3,751
    Post Thanks / Like
    Likes (Given)
    2534
    Likes (Received)
    1538

    Default

    Hi Priitm,

    Quote Originally Posted by priitm View Post
    I am skeptical regarding all kind of tape scales. Not sure, how much they stretch during gluing. I have some experience with installation lika and sony tapes, they seem too soft and unstable to be precise near 1 micron.
    I was concerned about exactly the same thing, so I was surprised that the Renishaw tape is specified at the micron level, because Renishaw is a metrology company with a good reputation for honesty in their specifications. For the RGS-20S tape that I am using here, the spec is +-0.75 microns in 60mm and +-3 microns/meter.

    Like you, I was very skeptical, so I worked out the numbers myself, because it's simple physics. My piece of tape is 300mm long, so the question is, how much force is needed to stretch that by 1 micron? How does that compare with the forces applied during installation? Let me repeat the numbers for you, because it interesting, and shows that Renishaw is not deceiving us.

    In my case, I was applying the tape using their application tool, which replaces the read head so it lays the tape down in exactly the correct spot. Since I was working alone, I did this in short cycles. I raised the grinding head by about 10mm, then walked behind the machine, and used my fingertip to smooth down the tape, which was hanging from the applicator under its own weight. Then I pulled off another 10mm of the backing paper from the adhesive, and repeated 30 times. The tape was never under any force greater than its own weight, even when I was doing the smoothing. The stainless steel tape has a cross section of 6mm x 0.2mm, so a cross sectional area of 1.2 mm^2. The total mass of the piece that I was applying (length 300mm, volume 360 cubic mm of stainless) was about 2.7 grams, so the force of its own weight is about 0.03 Newtons.

    Let's compare this with the force needed to stretch my L=300mm piece of tape by dL=1 micron. That is a fractional deformation dL/L = 3e-6, and since Young's modulus for stainless is about Y=2e5 N/mm^2, and the cross-sectional area is A=1.2 mm^2, the force required is
    F = Y A dL/L = 2e5 x 1.2 x 3e-6 N = 0.72N
    which corresponds to a weight of about 70 grams (3 ounces for those in the USA). I never applied that much force to the tape, or anything near it.

    My conclusion: if the Renishaw tape is applied carefully, it will not be stretched enough to break the spec.

    Cheers,
    Bruce

    PS: for the record, the adhesive that Renishaw uses for these scales is made by 3M. If you apply the scale to a clean and smooth metal surface, it's remarkably difficult to break the bond.
    Last edited by ballen; 11-14-2021 at 04:38 AM.

  10. #8
    Join Date
    Sep 2011
    Country
    GERMANY
    Posts
    3,751
    Post Thanks / Like
    Likes (Given)
    2534
    Likes (Received)
    1538

    Default

    Here is the encoder installation for the cross feed. This is a KA-200 ML180 mm 1 micron resolution scale from Sino Germany, cross section of the scale is 16 x 16mm and the read head 14 x 16mm, so the scale and head fit into a 16 x 32mm opening. The scale came with a calibration chart, showing maximum error of +-1.3 microns over the range.

    The scale is mounted on the right side of the saddle, next to the ways, under the saddle. It's mounted to a piece of 20x20mm aluminium L extrusion, 3mm wall thickness, which is epoxied to the bottom of the saddle. If you look closely you can see that I'm milled a drip-lip onto the machine side of the L extrusion, so that any oil or coolant will drip off the extrusion rather than running onto the scale.





    The scale head is mounted to a blued steel bar, screwed onto the machine chassis in two spots:



    Then I added a cover, and reinstalled the trips that flip the feed direction:







    This is a well-shielded location, on the non-spray side of the grinder, so I expect the scale to work trouble free for many years.


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
  •