When making the arms initially omit holes C and then assemble together with the arm
spacers and secure by riveting and filing flush the ends of the arm spacers (6).
Next place the cross bar (11) between the two arms, clamp and drill the 10mm hole
through all three, ideally, drill and ream. Small holes could though initially be
drilled in the three parts enabling them to be aligned using a pin making it easy
to position the parts for clamping.
Follow this, with the cross bar between the two arms, by measuring the width across
these and make the hinge pins (4) just larger, say + 0.02 to + 0.04mm. I have included
the third pin as a spacer as that is the way I have made mine and it worked well,
more about that later. However, the reader may like to omit this and place a nut
behind the countersunk plate to secure the screw, then another nut inside to support
the second plate with a third nut on the outside. The two can then be used to adjust
the space at the outer end of the side plates. In retrospect, I think perhaps this
is the best method.
At this stage I assembled the parts and found that there was still a little movement
at the ends of the arms, though very much less than previously. This was not that
surprising as any clearance between the hinge pin (4) and the cross bar (11) would
be multiplied by a factor of about 7 to 1. I decided therefore to take out the clearance
between the two parts by securing the hinge pin with a screw (H3). This improving
the situation considerably. To avoid the screw sliding the hinge pin sideways as
it started to grip, causing one of the side arms to become stiff, or even locked,
I placed a small ball of very fine copper wire between the screw and the pin.
With the tool reassembled I then put it to use producing a diamond knurl on a piece
of 20mm diameter steel, 230M07, maybe leaded, but it was only a scrap so was not
marked. Photograph 3 shows the setup and Photograph 4 the result, as can be seen
this is good. It is a pity that I did not carry out the same test prior to carrying
out the changes but my opinion is that it was an improvement, it was certainly more
satisfying to use.
This prompted me to measure the movement at the end of the arm, Photograph 5. Applying
a firm pressure with a finger, the result was about 0.0005” (imperial dimension as
I was using an imperial gauge as the photograph shows). Using a very firm pressure
the deflection was still only just above 0.001” and some of that may be due to other
factors. I declared the repair a sizeable success!