Dro

With the help of many members of the
http://groups.yahoo.com/group/CAD_CAM_EDM_DRO mail list and
the rec.crafts.metalworking newsgroup with special thanks to
Steve Lindsay of Lindsay Engraving for making his DRO software
available to the public, I now have a DRO up and working. By
researching the Digest Archives of the DRO list, I happened upon a
link to http://www.goldmine-elec.com/ I was able to find an
inexpensive source of motors that have 1000 line quadrature
encoders using the HP encoder (Sorry, NO longer available). The
only problem I ran into with these units is that the shaft diameter
is 5mm not 7/32" as advertised.
By working with the parts, I found for me that the continuous loop
idea over two ball bearing axles was going to work better than the

system used by Steve in his design. I felt this was better as the shaft the encoder is mounted on
has two ball bearings instead of the bronze bearing I believe he used. This along with the additional
ball bearing mounted idler axle, allowed me to have higher tension on the axles than allowed by
the 2# limit imposed by the plain bearings. This will give better friction coefficients than previously
allowed. Also, as all movement is on ball bearings, this will give less torque requirements than
needed by the plain bearings. Another change was to use cable that is used on parallel bars for
drafting tables rather than the fishing line used in Steve's design. The reason for this was simple. I
had it in my junk box.
Another change was to the interface to the parallel port. I used a 74HCT14 IC from Radio Shack
with pull-up resistors on the inputs instead of the discrete transistors and this was wire wrapped
board that is fitted into the computer case. Power is pulled from a spare connector in the
computer. Below, are some pictures of my setup along with a picture of the design of my blocks.

An exploded view of the Z axis encoder block. The others are
similar. The blank plate is shown on the right side of the block.
Originally, I tried to make a block that was solid on this side. But
after spending three hours trying to thread the 0.035" drafting
board cable around the axle and back out, it was decided it would
be quicker and easier to bore the slot all the way through the
block and put a blank cover on the other side. The two counter
bored holes located above and below the main bearing hole are for
two 6-32 screws that go into the plate for the encoder wheel
cover. The first time, I drilled the holes out for the screws, but
found the head of the screw interfered with the encoder block.
Also when assembling the plate, you must be sure to have
clearance for the encoder hold down screw to clear the block.

This shows the X axis mounting. The block for the encoder end is
wider as I tighten the screws from the encoder side instead of the
off side. The screws run into "nuts" in the dovetail where the table
stops used to reside. The anchor for the cable is in the middle and is
two pieces of 1/8 x 3/4 aluminum to sandwich the cable. A 6-32 hole
was tapped in the middle of one of the original bolts for the stop
bracket and the anchor is slotted to be sure the cable is properly
aligned. Please ignore the bucket used for a coolant supply!

Connection of the encoder to the interface is quite simple. Using shielded cable, connect the Red wire
from the encoder to +5v, Black to ground, and the Blue and Yellow wires go to the A and B quadrature
inputs. I also grounded the shield at the interface end of the board only. Power was derived from the
computer using one of the spare power connectors and is by-passed on the interface board with a
0.01uF capacitor .

Calibration of V5.2 is really easy thanks to Roberts "Calibration" routine that is built into the program.
You simply follow the directions and move the table a known distance. To do this in the X & Y axis, a
scrap block of aluminum was placed in the vice and two slots were milled to hold a 12" scale for each
axis. A hold down screw and washer was used to secure the scale and a sewing needle was put in the
drill chuck for a pointer (I first used the tip of my scribe, but it looked like Mount Everest under a loupe:-})
For the Z axis, a dial indicator was setup in a collet and the distance measured this way. For the best
accuracy, the distance measured should be as large as practical to gain overall accuracy.

This is a close up shot of the block that was milled to hold the
scale in the X and Y axis direction. By Milling and NOT removing
the block, the scale will be parallel to the axis being measured
even if the vise is not perfectly square.

Here is a shot showing the scale in the X axis for
calibration. The needle is held in the chuck and we
are not worried about it being concentric as the
spindle is not turned on during calibration. It was
found the gradient was easier to read with the loupe when the
needle was located at the top of the mark on the scale. The
screw and washer "locks" the scale in place so you don't
accidentally move it while making the calibration.

Picture of the jigs used to measure hold the scale and the
dial indicator (1" travel in this case). The bar for the dial
indicator is simply a piece of rod that was cross drilled
and tapped to receive the bolt on the indicator. This was
then held in an appropriate collet for set up and
measuring. Be sure to have the plunger parallel to the
spindle so you do not introduce any sine error in your
measurement.

UpDate!
Robert Duncan has gone to the effort to re-write DRO as version 5.4 that includes new features. The
program will now attempt to select the proper type of parallel port AUTOMATICALLY, has a very nice
new GRAPHICAL INTERFACE so us Ole' people can see the screen, and a self calibration system. To
use the graphical font, you must use an interface
designed for IRQ mode of operation. He has also furnished a schematic using ICs to handle this (Total
of 4 ICs).
If you want Steve Linday's original background information, please get the 4.1 file
information..