No announcement yet.

Chinese scale ?

  • Filter
  • Time
  • Show
Clear All
new posts

  • Chinese scale ?

    I'm building myself a little (desktop) CNC gantry mill ATM. Planning to use cheap DRO scales/calipers to close the loop. I've found the comms protocol and can read that with a PIC. Question is, does anyone know where I can get connectors for these critters? I can pull em apart and solder to the board but that's kinda messy.
    Any pointers would be much appreciated.
    Rgds, Lin

  • #2 I think 8.95 and they ship immediately.

    Remember, the chinese signal is less than 5vttl level coming out of the scale. It must be amplified even for it.

    The signal is actually the battery voltage. The connector/ to the pc supplies voltage, not sure how they are doing that unless they have a voltage divider circuit and feeding off the pins.

    Max232 makes a good chip. It uses capacitors to take low level signals and bump them up to pc-readable levels. It requires a max232 chip and about a half dozen small electrolitic capacitors.
    It'll run off 5vdc if I remember and output up to 7.5 to 10 in the signal lines.

    I can post you the vb code for reading it using MSCOMM for a communicator addin to take care of the port for you.

    I am fixin to put one on my shaper, one on the head ram and the other on the cross slide. It is not a real cnc with blinding speed just going to be a automatic cut control. I am having some trouble right now with it siezing up on the ways.. It is I am pretty sure a 1900 or so model.. From the looks it still has some original grease in there. HA..
    The top is the cable I bought.


    [This message has been edited by ibewgypsie (edited 12-05-2004).]


    • #3
      Thanks for the advice Dave, I'll go check em out. Re signal levels, I'll prob take care of that with op-amp comparators. I'm gonna use an old DOS box to control it, writing my code in turbo pascal. I want it to intepret HPGL files as well as nc drill files, g-code etc. Its main tasks will be simple 2D milling and engraving (front panels etc) and drilling PCBs


      • #4
        BTW, for anyone interested, the comms protocol can be found at
        Oh, and good luck with the shaper Dave. How you planning to close the loop there?


        • #5

          Turbocnc is wrote in Turbo Pascal 7. That'd save you about a million lines of code. Register it and he gives you the code.

          Really not a closed loop system on the shaper.. Kinda a monitor. I have to write my own because of the scale. Unless I can bum or buy the working edition from you.



          • #6
            Thanks again David. I have little experience with CNC machines but a fair bit with TP. I'll hafta check out turbo CNC, but as I understand it, that is for clock/direction control of steppers and servos. I'm using straight DC motors (24v wiper motors) to drive the lead screws. As soon as I have any serious code working, you'll be welcome to it. At this stage, I'm still working on the engineering of the table and axis drives etc. It aint easy in 1" aluminium with only a mitre saw and belt sander
            No mill here, just my little 9 x 20 lathe.


            • #7
              S&S, Of the two Chinese scales I have, one will occasionally go into erratic readings with no nearby cause (motors, magnetics, etc.) apparent.

              If you are coding your own control for the machine, it might be a good idea to test every increment on a running basis to see if it is consistent with the speed of motion and about what it really should be. This would give a little safety to the digital scale usage.


              • #8

                Think on this, Parallel port PWM controlled motors, encoder feedback. I worked out some software back in the spring to do just that. Putting the resolver (positional feedback) in the PC makes it work hard. Windows won't do it, it'll lose position and pulses. It must be single tasking. Pics are a good ideal. If you can figure out a way to USB them even better.

                Still just to save the time reading Gcode files and the subroutines he has in there. It is documented really good.

                I suck at Pascal. I remember just a tad, but too many brain cells have been reformatted since then.

                I got a copy of Steven's STepster program too in basic. Website says he passed away.
                I have here Borland turbo basic too. Not sure if the copyright has expired on it or not thou. It'll compile to a exe on it.
                Also, TP english help file.. it does not come with the code from

                The scales communicate so slowly, you can't use them as feedback, only a reference. If you looked at every distance, you could calculate a time-to-desired off them. I have been trying to do that here. Not got it worked out yet. It is in visual basic 6.

                I ran my mill a bit today, I have XP Pro on there as a OS. While running a program I noticed it staggered in speed. I looked up to see a balloon around the right lower, it said to hook up to MSN for "updates". I was quite angry. I felt like trashing the whole mess, formatting and going back to turbocnc. It never stutterered.. One track mind.



                • #9
                  Thanks again for the constructive comments guys. That's what this site is all about.
                  I may have to re-adjust my head on this one. I'm looking for a cheap way of absolute position sensing so I can PWM control the DC motors and not have to worry about backlash, lost steps etc.
                  The axis drives will be slow as a wet week anyway, so maybe there's still hope. I don't care how slow it is, as long as I can push a button and walk away.
                  Some food for thought here while I proceed with the mechanics.
                  Best rgds, Lin.


                  • #10

                    Wave type stepper drive, 4 bits from the parallel port = 4 pins on 3900 nand gates w/inputs tied together. Output goes to tip120 transistors, they switch on and off coils inside the steppers. A few more parts than that but that is mostly it. Turbocnc has wave drive as part of it's control scheme.
                    The current is controled via inline ceramic resistors. Not as fast as a gecko, not 2/3rds as fast but you can build the whole 3 axis control board for about $20.

                    I have drawings somewhere.. probably on the old 98 pc..

                    And them new servo chips that are out, gee whiz batman.. One large 28 pin chip w 1" long tab.. gee.. never saw nuttin like it but it is good for 27 amps. Resolver is onboard. By the time you buy a pic, and "that" chip in onesies twosies thou you are within $10 of a rutex 990h. And geckos are on sale right now, $99.each. Servo or stepper.



                    • #11
                      Thanks again Dave, but by the time I build/buy drivers and steppers with enough torque, I might as well buy a Taig or similar.
                      Stepper motors in that range are an obscene price in Oz and the freight pretty much precludes offshore purchase. I have built my own logic and H-bridge stepper drives for past projects but that was just little type 23s.
                      I'll persevere for a while and save my pennies for a real mill, even if its a real chiwanese mill no cheap used Bridgies in these parts


                      • #12
                        Rutex is Ozzie, (down unda) australian manufacturer.

                        You should be able to cut a deal. I got motors laying everywhere. Some of the servo deals I have found are old-er tape drive motors about 600 oz torque for $48. Canadian electronic surplus.

                        I got this 3hp INdex cnc w/servos rusting quietly, I'll build a box large enough for me/bike to climb in next to it and ship it to you. I been wanting to see the land down under anyways.

                        Sorry I am not more help. Been up all night again here. (Darn doctors anyways) Should've took the green pill last night. I did find and download the documentation on the two reliance SP500 2hp inverter drives I came up with. Now what to use them FOR? Probably one will go on a belt sander polisher out under the shed roof.

                        Lemme know which way you are going. SOunds like you are heading the way I started about a year or more back. I started collecting Electronics to do a similar project, I got about twenty pounds of "stuff" including HC68 Motorola chips and other goodies.



                        • #13
                          Re: The level conversion thing.

                          Almost any TTL chip will switch between Hi and Lo at about 0.7 volts. If you have a 5 volt circuit just use a hex buffer or inverter and it should work without any extra components. This will not work with CMOS logic which switches states at half the supply voltage (2.5 V for a 5 V supply). But a TTL chip can be used as a level converter for a circuit that uses CMOS components.

                          If you are using a higher supply voltage in your circuitry, then almost any small signal transistor will do it. Use an NPN with a base resistor calculated to limit the current to a 1 or 2 ma. (1K) and a collector resistor of 10K or more. Again, no more parts are needed. The 2N3904 and 2N3906 come to mind but I can't remember which is NPN and PNP.

                          Heck, your logic circuit may have a TTL input and not need any further conversion. Try it that way first.

                          Either way, you should be able to convert the levels for less than $1.00 in parts for several axies.

                          Paul A.
                          Paul A.
                          SE Texas

                          Make it fit.
                          You can't win and there is a penalty for trying!


                          • #14

                            I would like to tell you about my "signal polarity convertor" I built using 3904 transistors. It swapped polarity coming out of the parallel port (12vdc) into 5vdc ttl levels. Switching them on and off for the step and direction pins on the drives.
                            My machine ran away. Right to the end stops on the ball screws. I forgot that it is also a amplfying transistor too. It amplified the rf noise from the drive stepper leads into a signal for the stepper pulse inputs. And with it self sustaining it ran away. OF course all my machine was in the breadbox testing stage and I did not have the end limits or the estop hooked up. YEP, had to pull the main switch.
                            I was redfaced on that one considering my experience level. I knew better.



                            • #15
                              Thanks for that Paul. My mind was on another prob when I mentioned the op-amps (was the profile digitizer I think, good for 5 micron without traffic vibration)
                              Sure all here are familiar with the 'fifty-leven projects' syndrome. You're quite right, most of the time the PICs will have no prob deciding between 0.7v and 2v. If any prob, a schmitt buffer or inverter will square em up.
                              Hehehe, cut my teeth on 74XX