Announcement

Collapse
No announcement yet.

Reed switch repeatability test for CNC

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Reed switch repeatability test for CNC

    We discussed limit and home switches recently and I suggested that magnetic reed switches would be a suitable alternative because of the long lifetime and sealed switch contacts. It was questioned if the repeatability of actuation would be sufficient to serve as a home switch and I opined that it would given the use of a very small but powerful magnet.

    This is so because the strength of the magnetic field surrounding a magnets falls off in strength by the inverse third power of distance from the magnet as measured in radii of the magnet. So, the smaller the magnet the more quickly the field diminishes with distance. This is what is needed to provide a sharp and repeatable actuation of the switch.

    I decided to quantify this tonight. The setup is a very small rare earth magnet from the optical tracking mechanism in a CD Rom drive glued to the end of an aluminum rod held in the mill spindle. This is adjusted to positively actuate the garden variety burglar alarm magnetic reed switch sold by Radio Shack.

    The measuring device is an Interapid .01mm per division dial indicator. The protocol is to use the home axis function in Mach 3 to have it home the Y axis, taking note of the indicator reading. Then the magnet is moved well off several inches for perhaps 30 seconds and the process is repeated.

    The result was very good with a repeatability error band of +0.01/ -0.03 mm over many cycles. It has a hysteresis of 1.1mm which is just right to deal with any backlash that may be present. The backlash compensation at the time of the test was disabled and the measured backlash of the axis was ~.01mm.

    The measured backlash isn't due to looseness in the leadscrew but is due to the total amount of system compliance (flex) present.

    I didn't bother to test it with a button sized supermagnet because it would actuate it from a distance of several inches which wouldn't be at all useful.

    The numbers translate to +.0004/-.0012 inches. If the home operation is repeated one second after the initial homing the error band reduces to +0.0/-0.01mm or -0.0004 inches. The measured error band may well be due entirely to system compliance instead of variability of the switch actuation point.

    Conclusion: If used properly with a very small but strong magnet a magnetic reed switch is a viable option for system home and limit switches.

    This is the setup I used.



    These are the magnets. Two are used in every optical drive read/write head assembly.

    Free software for calculating bolt circles and similar: Click Here

  • #2
    Evan, your techno-exploits are purest gold. You've written a simple, easy to understand, easy to replicate experiment demonstrating something useful to CNC people: reliability of reed switches as limit/homng switches in CNC axis drives. You've written others that I've found fascinating, emminently sensible, well illustrated, and apt. Speaking as a technical writer of limited credentials I decclare you a resource for your constituency. Long may you flourish.

    I hope you're keeping copies of all this. I foresee a book when sufficient material accumulates: "Secrets from a Basement Workshop."

    Comment


    • #3
      Coming from you Forrest that is indeed high praise. Thank you.
      Free software for calculating bolt circles and similar: Click Here

      Comment


      • #4
        What about machines that are working with ferrous materials all the while ?
        I have reed switches on my router but this only cuts wood and plastic.

        .
        .

        Sir John , Earl of Bligeport & Sudspumpwater. MBE [ Motor Bike Engineer ] Nottingham England.



        Comment


        • #5
          Problem with magnets is they attract metal chips which in turn become magnets. one chip and your home operation is screwed. No longer accurrate.
          Excuse me, I farted.

          Comment


          • #6
            Hi Evan,
            I agree with Forrest. Your contributions have been greatly appreciated by me. Often I just ingest without responding.

            I also can say I have magnet reeds on my CNC router and they maintain the highest repeatability. BUT, there are no ferrous metal chips to collect and alter the consistant results.
            John M...your (un)usual basement dweller

            Comment


            • #7
              In the previous discussion I suggested a couple of ways to deal with that. Because the magnetic field of these magnets drops off so quickly it is not difficult. If a cover is provided that simply keeps the chips away from the magnet a short distance they won't stick. Or, the magnet can be shielded inside a short length of steel tube covered by another length of aluminum or brass tube. When it is actuated it is extended into the aluminum tube where it can exert influnece on the switch. When it is deactuated it is withdrawn, perhaps by it's own magnetism, into the steel tube where it cannot attract chips. [edit] It's the same principle as a magnetic gauge base. Any chips it attracted during the actuation fall off the outer tube.

              I am going to try a couple of designs when I have time. Right now I have to do house work. My wife will be coming home today from the saw filers convention.
              Last edited by Evan; 04-26-2008, 12:39 PM.
              Free software for calculating bolt circles and similar: Click Here

              Comment


              • #8
                Good work Evan. I conducted much the same tests for an application inside an automatic transmission with similar results. I found improved performance by using two small samarium cobalt magnets approaching the reed switch broadside, one magnet at each end of the switch with the magnets at opposite polarities. North on one end of the reed, south on the other. This gave a nice crisp switch point that was more accurate than a single magnet. the very small magnets, buried in plastic, attracted some of the magnetic contaminents in the transmission, but dispersed over the surface of the plastic, had miniscule effect on the switch point.

                Edited to add:
                I had to do considerable testing and research to convince myself, and then make a Presentation to a wad of mid-level Big Shots and functionaries from one of the Detroit automakers. It was a hard sell but the facts stood on their own merit.
                Last edited by Weston Bye; 04-26-2008, 01:42 PM.
                Weston Bye - Author, The Mechatronist column, Digital Machinist magazine
                ~Practitioner of the Electromechanical Arts~

                Comment


                • #9
                  I always enjoy your posts. Don't always understand or care but they are always interesting. Fred

                  Comment


                  • #10
                    Limit switches and "home" or "reference" switches.

                    Magnetically operated switches work well as limit switches. They are, however, strictly pilot-circuit devices. As I'm sure you're aware, the design of the system that actually does the necessary work needs to be carefully analyzed to be sure it will "fail safe." Unless, that is, you enjoy fixing broken machinery! Been there, done that. More than once. Once it was even on one of my own designs.

                    For "home" or precision switches I've had much better long-term reliability using a vane to interrupt a light beam between a light source, usually a LED, and a photo-detector. I have found that using a wide vane such that there is only one transition over the whole range of operation works best with the least complication of logic. This insures that, for example, moving in a left to right direction will give one high-to-low transition as the "home point is passed while moving in the opposite direction will give one low-to-high transition as the "home" point is passed.

                    While using a narrow flag and deciding which transition you're interested in depending on the direction of travel can be done, I've found that madness lies at the end of the path!

                    I understand that you're designing a machine for your own use and that it's quite unlikely anyone else will ever need to maintain it. Perhaps your memory is much better than mine. That would not be impossible by any stretch!

                    Just my 2 centavos worth. YMMV!

                    Comment


                    • #11
                      Originally posted by David E Cofer
                      Problem with magnets is they attract metal chips which in turn become magnets. one chip and your home operation is screwed. No longer accurrate.
                      Chips can be a problem for any type of switch. Even a sealed microswitch can have a chip stick on the paddle or on the surface that the paddle hits. This will have the same effect on accuracy as a chip on a magnet. Worse yet, with other types of switches both magnetic and non magnetic chips can have the same effect so the magnetic types may actually have a distinct advantage.

                      The only way I can imagine to completely eliminate chips from any type of switch in this situation would be to place the switch at the end of a tube that is longer than the movement of that machine axis. A long rod would enter through a tight fitting seal and it's working end would never leave the sealed enviroment inside the tube. But that is so arkward that I can't really imagine anyone actually using it. Unless human life was at stake.

                      Evan, nice work. Very useful.
                      Paul A.
                      SE Texas

                      And if you look REAL close at an analog signal,
                      You will find that it has discrete steps.

                      Comment


                      • #12
                        Thanks all, but this is just a simple switch test.

                        GearedLoco,

                        During the time I worked with Xerox I saw the equipment designs progress from machines that were fully electromechanical with perhaps some vacuum tube electronics to fully computerized and networked multifunction devices. Some of those machines are more complex than the most complicated multi spindle machining center and use multi computer distributed CNC control systems for mechanical, pneumatic and hydraulic systems all in the same machine. The internal environment is just as bad as any metal machining equipment even to including accumulations of fine magnetic metallic powders that closely resemble cast iron swarf as well as the worst silicone based heated migrating oils that become mixed with micron sized particles of magnetic black dust, all a part of the process.

                        For the clean areas optical switches are the usual choice and work very well with good reliability. For the dirty areas they are very problematic and difficult to seal. Where total reliability is required reed switches are often used even though the designs are normally more complex mechanically.

                        Paul,

                        Remoting the switch from the object that produces the actuation is a common practice when it is possible. Long plungers or levers are often used.

                        ------------

                        Even in clean areas reed switches may be used when ultimate reliability is a concern. The key pad below was salvaged from a machine that weighed about a ton and ran on 230/50 amp service. When testing it the key pad was used to actuate many of the major subsystems including the main drives. Intermittent operation wasn't an option so it uses reed switches.

                        Free software for calculating bolt circles and similar: Click Here

                        Comment


                        • #13
                          One thing to bear in mind is that non of this stuff is new or very, very little.
                          What it often is, is a new slant on an old idea.
                          CNCZone has nearly 87,000 members and has been running for about 5 years.
                          Most of what we read on this board by various authors, me included, has been rehashed before.

                          Industry in it's search for the $$ has also been this route.

                          A quick search usually finds you the answer or very near to it.

                          .
                          .

                          Sir John , Earl of Bligeport & Sudspumpwater. MBE [ Motor Bike Engineer ] Nottingham England.



                          Comment


                          • #14
                            Industry in it's search for the $$ has also been this route.
                            One of the things that continued to surprise me when I worked for Xerox was the astounding lack of corporate memory. We technicians would constantly see new designs come out that clearly showed they were still looking for a way to do something that had been solved years ago. The same old problems cropped up again and again. Problems with materials, overly complex designs with overly complex adjustments, and worst of all designs where frequently serviced components required substantial disassemble to access the component.

                            The turnover of people in the high tech industry tends to be quite high for most companies. When an experienced designer leaves so does his expertise and and accumulated knowledge of the product lines. Even if the problems and solutions are well documented in some company database it is obvious that unless somebody else remembers the problem the new engineer will tend to make the same mistakes all over again rather than refer to the historical record of problems and solutions.

                            Xerox finally attacked that problem by requiring new engineers to work in the field for the first six months of their tenure alongside long term experienced service reps who provided them with valuable insight as to what works in the real world and what doesn't. It wasn't long after that program started that we began to see adjustments go away in favor of predetermined hard point alignment of assemblies. Covers became less of a puzzle box affair and frequently serviced assemblies could be removed without tools.
                            Free software for calculating bolt circles and similar: Click Here

                            Comment


                            • #15
                              Originally posted by John Stevenson
                              One thing to bear in mind is that non of this stuff is new or very, very little.
                              What it often is, is a new slant on an old idea.
                              CNCZone has nearly 87,000 members and has been running for about 5 years.
                              Most of what we read on this board by various authors, me included, has been rehashed before.

                              Industry in it's search for the $$ has also been this route.

                              A quick search usually finds you the answer or very near to it.

                              .
                              Good answer! Now that's pumping the suds out of the water.
                              John M...your (un)usual basement dweller

                              Comment

                              Working...
                              X