• Final thoughts

    Igor • 08/30/2017 at 10:56 • 0 comments

    Well, it took some time to complete, but it has been in happy service for a few years now.

    The 1 micron resolution is a bit overkill as the flexibility of the rest of the machine is the weak link. However, if I ever bother to make a fancy toolpost grinder, the accuracy might come in handy.

    I did make it work with Yuri's android DRO, but actually never bother to bring the tablet. If my eyesight goes bad when I get old I'll dig it up again.

    Things to do different if you ever get around to undertaking something like this:

    - Use a push button to switch it on and have an auto shut-off. I often forget to switch if off and come back to an empty battery. They're rechargeable but it's a nuisance.

    - Make the housing much stiffer. Plastics warp over time under pressure so taking this more into account will improve your experience

    - Don't use an LED to light the LCD. Nowadays there are cheap OLED displays which will make life better.

    - Use something with more than 2k of program flash.......

    Anyway, hope you enjoyed it.

    Igor

  • Adding Bluetooth

    Igor • 08/30/2017 at 10:41 • 0 comments

    The display is a bit on the small side and as I came across Yuri's site here I thought it would be good to integrate a bluetooth link to show the numbers on a tablet.

    I made a small board so I could mount an RN42 bluetooth module onto the programming port.


    I didn't pay enough attention and found out later I only had one pin available. So after cutting those traces off the power control and LED control couldn't work anymore.

    With the UART pins already occupied by something else, the serial stream is bit banged through a single pin on the programming port. Timing was a bit tricky as the serial stream shouldn't affect the reading frequency of the magnetic sensors.

    Now the 2k of flash was really starting to hurt and I had to optimise the code for the lcd severely.

    But, in the end it worked. Bonus feature was that it also displays the battery voltage on the Y channel so you know when it's going to quit on you.

  • The enclosure

    Igor • 08/30/2017 at 10:31 • 0 comments

    At the time (in 2012) Shapeways was just starting their 3D printing services, so I happily made use of them. Colored options were not yet available which means I had to paint them if I wanted another color but it was not much of a success. This 3D printing process at the time still used wax supports or something like that which didn't come off properly and I didn't want to use any harsh solvents. As a result the paint doesn't adhere well.

    The design consist of two main parts; the main housing with the longitudinal sensor plus a small satellite housing for the cross slide sensor. It was the first CAD part I ever designed so it's a bit crude, but it's functional.


    If you remove the satellite connector on the side the cover can rotate open to give access to the battery compartment. When closed the connector keeps it securely in place.


    Things that went wrong:
    - small pins to hold pcb were not strong enough and are breaking off. Splitting them four ways to snap into the hole didn't help. Should have made them solid or better, used some screws.
    - the material warps, so now a few years later it's not straight anymore. The only issue is that it doesn't snap closed anymore but the connector hold everything together anyway.
    - the painting obviously, but now you can just get colored plastic straight from the printer.....
    - the holes for the buttons are fine when the cover is closed, but they slightly interfere during opening

    The main sensor sits in a pocket behind the pcb, I added a conector later so I could take it apart again:


    The satellite just holds a pcb with the sensor IC. The sensors are "potted" with hotmelt, which still works just fine five years later.

    So, normally you don't mount the sensor on this side of the cross slide, but on the other side it would interfere with the tailstock. The magnetic target only adds 2mm to this side and the sensor is rather small, so it's not really in the way.

  • The electronics

    Igor • 08/30/2017 at 10:12 • 0 comments

    This project started out already quite some years ago and has been more or less finished for a while now. Projects don't go very fast when you have kids.

    However, I thought it would be nice to show how it was done and perhaps inspire someone else.

    As mentioned, the idea came when I saw the magnetic encoder sensors from AMS and I thought it would be nice to whip up a DRO out of them.

    Regular DRO's have quite some functionality but also require a large keypad. I figured I would need only reset inputs for the scales. It's a bit more work than with a keyboard, but it's fine really.

    The design started with an ATTiny2313, as I only required an I2C bus and not much programming. This came back to bite me as the 2k of flash required quite some squeezing when I added a Bluetooth interface later. But in the end it all fit.


    To keep everything small I wanted to use only one AA battery, so there's a small boost converter to create 3.3V for the MCU. The lcd display required 5V so a voltage doubler is connected to a pwm output and regulated by a 5V zener.

    The on/off switch also lights an LED for the display in one of the positions to enable working in the dark. Overall the LED is pretty useless though.