Announcement

Collapse
No announcement yet.

Parallel port encoder input for homemade DRO

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

  • Parallel port encoder input for homemade DRO

    A few years ago I saw a DRO made out of a old XT (cheap) using the parallel port for input for some claristat optical encoders.
    NOW I understand they used the 4 input bits in the second half of the port address, And I understand the sequential logic of encoder direction, but I am at a total loss on why not more of this has been done. Them DRO's are expensive. YOU have another 8 output bits to control anything else you want from the computer also. I am operating the steppers via this way. Just not using the input bits.
    IS all this common knowledge to all machinists? Or is the robotics field just now starting to merge, or I am more confused than before?
    (ohh the mechanical position indicators on my bridgeport pooped out triggering this episode) I currently have a magnet with attached ruler stuck to side of machine slide.

  • #2
    Computers have whats called "interrupts" to worry about. They will mess up the timing on any tight stepping pulses that may be required for step motors, etc. They can also cause a loss of counting pulses from your scales. There are a few programs that can overcome this.

    The best groop for homebrew CNC and DRO stuff is "CAD_CAM_EDM_DRO", (a yahoo group).

    http://groups.yahoo.com/group/CAD_CAM_EDM_DRO/

    Very helpful site, try them.

    Tom M.

    Comment


    • #3
      As the resolution of your encoder goes up the other thing that happens is that the speed of data transfer must go up to keep up with it. Soon, you are into the megabit update rate and need external counters. US Digital has some pretty neat chips (I forget the actual maker, the part is a 7266) at relatively low cost. They are multiple micro interfaced counters with quadrature inputs for encoders. There are some simple interfaces shown at their site www.usdigital.com as well as a few others in the net. There are parallel port, rs232, isa bus and opther interface types available or you can roll your own. They also have source code and demos for all their stuff.

      There are several other parallel port DRO interfaces on the net, one I believe counts directly from parallel port lines. The author mentions maximum rates he was able to attain with this method. Sorry but I don't have the link for it.

      Den

      Comment


      • #4
        Taking the weenie way out.

        I profess a micro knowledge of things electronical but I can't seem to justify the level of study and experimentation it takes to build a reliable, do-it-myself home brew DRO. All the building blocks are available in the form of ships, reader heads scales etc.

        Accurate, backlash-free measurement is the be-all end-all of the machinist's trade and is in the end is all that separates a DRO that works from a POS better reposing at the re-cyclers. There's no middle ground.

        Maybe I'm dumb and wasting money but when I want to know where my machine axes are I don't want to know almost within a couple of counts here and there with maybe a litle reversal error. Therefore I want a measuring system separate from the lead screws with a greater intrinsic resolution than the minimum tolerence I work with.

        So I spend the money and get pretty good stuff becaause some of my daily bread depends on it.

        I fervently wish someone would come up with a DIY multi-axis computer based DRO system reading to 0.0001" or better that requires a few low cost computer cards and measurement scales and reader heads costing less than $1 an inch.

        As it is you have to spend $300 to $1000 for the display and A base cost of $500 plus $10 per inch for the scales plus $200 for the reader.

        My Sony DRO for my lathe cost over $2700 and I don't begrudge it but it's three times too expensive considering the decline in the cost of other industrial electronics over the years.

        Comment


        • #5
          Forrest,

          I'm in the electronic industry, and I can't agree more with you about everything you said. I have research this over and over for the past year or so, and I'm unable to source inexpensive parts which will produce the level of repeatability and accuracy that a DRO needs to have. Almost always, the DRO manufacturers use proprietary parts whether it be optical, magnetic, or capacitive.

          I've explored many home made solutions, but I'm unable to find one that will yield the necessary accuracy over time ( at least 1/2 thou). As you said, DRO which leaves you second guessing is of no value.

          Albert


          [This message has been edited by Rotate (edited 02-24-2003).]

          Comment


          • #6
            Rotate is correct, I think. The issue is the scales/reader. Everything else is "easy".

            Ones that just read the rotation of the screws cannot have any knowlege of actual position, they know what it SHOULD be with everything perfect. News flash: everything is NOT perfect. So forget that.

            Now, how to create a home-made scale accurate to tenths (so you get good thousandths) is a problem.

            Making it heat-stable so it doesn't vary out of tenths while being used is another issue.

            I don't see why this should be easy, or else I must have failed to notice the $50 DROs for sale new.

            Comment


            • #7
              Forrest:

              $2700 for a Sony DRO?
              You did!! pay three times too much! :-)

              Tom M.

              Comment


              • #8
                Forrest:

                Likewise, I could not agree with you more. During my posts on the "mouse smarts" topic and before posts here, I turned around and bought 3 new Heidenhain scales on ebay as well as a new Acu-rite turning DRO to go with 2 of them.

                I found that resolution was my problem. I can hold 0.0005 or better on metal and the Heidenhains give me that easily plus 1 micron resolution for some other optical stuff I do (positioning, not cutting).

                The whole shebang (spelling ?) cost around $900, I have 1 micron resolution and very nice turning oriented displays, referencing, etc. Shortly after, a brand new Acu-rite (milling DRO) appeared on ebay that could have saved hundreds more.

                I do have to admit sheepishly (baaaahhh !) that they have not been mounted yet.

                Den

                Comment


                • #9
                  The problem with the leadscrew backlash I have seen overcome.The solution was to mount the encoder on a mandrel at one end of the table and an idler at the other then with a stianless steel aircraft cable streched around these pullies and attached to the saddle it done away with the backlash the critical part of this system was the diameter of the encoder pulley.However I personaly would not build one because of two reasons#1 dros are too cheap.#2 why renvent the wheel?
                  I just need one more tool,just one!

                  Comment


                  • #10
                    The scales seem to be the limiting factor. Creating something to read the scales doesn't seem that hard. Old PC's might not be fast enough though for high resolutions. My idea was to use a dedicated microcontroller, something like a Microchip PIC, which is probably what the comercial units are using. A microcontroller, a display and some interface cicuitry should to it. Since it's dedicated to reading 2 ports, outputing 1 or 2 and calculations, the controller should keep up.

                    I currently have a 2500 PPR encoder, a PIC controller and an LCD display with a serial input. If I had time, it could easily be cobbled together into a rough DRO.

                    Figure the encoder geared to give 2500 pulses per inch and double that for using the A and B phases to determine direction. How fast can you move a machine axis? An inch a second? That gives roughly a 5khz signal for a resolution of .0002. The simplest PIC runs at about 4 MHz, so I should be able to catch the pulses and have a little time for calculations. Question would be, can I update the display fast enough? I figure less than $80 per channel on reader and display parts. Now if that works, I should be able to connect a true glass scale and have something.

                    I've probably missed or over simplified something here, but i've had my Acurite and scales completely apart and the only hurdle I see is that @#$ glass scale, that's where the accuracy is.

                    Now if someone just knows where I can buy some free time....

                    Comment


                    • #11
                      My research pretty much agrees with Forrest Addy's. If you want precision (real precision), you gotta pay for it. I can't see any point in building a DRO that's probably going to end up being no more accurate (probably less accurate) than my handwheel dials, and certainly not guaranteed. Can you guarantee your homebrew DRO will be accurate to 0.0002" at EVERY point along its travel?

                      But if you happen to enjoy playing with that sort of thing, have at it!
                      ----------
                      Try to make a living, not a killing. -- Utah Phillips
                      Don't believe everything you know. -- Bumper sticker
                      Everybody is ignorant, only on different subjects. -- Will Rogers
                      There are lots of people who mistake their imagination for their memory. - Josh Billings
                      Law of Logical Argument - Anything is possible if you don't know what you are talking about.
                      Don't own anything you have to feed or paint. - Hood River Blackie

                      Comment


                      • #12
                        You're right SGW, without a commercial scale connected to my homebrew reader, I wouldn't guarantee squat! For anyone making money with their machines, it's not even worth thinking about.

                        Even if it's a hobby, figure the time it will take you to cobble together at your hourly pay rate plus parts and you probably could have bought a decent used DRO.

                        I'd still like to attempt it though, just to see what I can come up with. I've already got the parts and the project is along the lines of my profession, so the experience might be worth it.

                        Comment


                        • #13
                          I'm curious why no one is going the "digital caliper" DRO route and then interpreting the output for a larger readout. If I weren't so busy making parts, I might try it. In the mean time, my small display on the side of the table will have to do. Mine cost under $100 each and I haven't been able to jump them yet.
                          Suggestions?
                          Mark

                          Comment


                          • #14
                            I have tried extract that using one of the imported digital caliper. The problem with those calipers is that they are repeatable to just under 1 thou, but their accuracy is about 2 thou over an inch (verified using Mitutoyo micrometer and block guage). Not quite good enough for a DRO. I guess one could write a correcting algorithm in the display unit so that it corrects for the inaccuracy, but if the unit exhibits any non-linearity, then that would be harder to compensate.

                            I guess a quality digital calipers wouldn't have this problem, but then you're talking about a price that's in the range of a real DRO.

                            Albert

                            Comment


                            • #15
                              Using a parralax Bs2 processor, Using the 128 ppr encoder, I was able to spin it with a drill (approx 1800 rpm) without losing count, The problem with that system was it was vibration sensitive and getting on the leading edge and vibrating would give false movement indication.
                              I continue to tinker, it solves boredom.
                              I did once make a binary encoder from some photographic film and cue cards. It worked great for that application. Lots of strip encoders are used on DRO, are they quadrature or binary?

                              Comment

                              Working...
                              X