Re: [CAD_CAM_EDM_DRO] DRO intermitent fails

Sven Peter wrote:

> Dear Jon,
> your answer gives me very detailed search directions,
> although my post was very inexact. That means you are
> pretty deep involved in the matter.
> My respect!
>
> It is a Mitutoyo glass scale of an elder style (I have no idea
> of the model) The scale is mounted at an angle.
> 45 - 60 deg? I can't remember well because I was too excited.
> => Yesterday wasn't my best day!
> It doesn't happen too often that I crash equipment I have to fix.
> (+ the telephone running hot)
>
> The reader's basement is linked to the reader head by a
> connecting rod that is spring loaded and has on the ends
> two brass balls. These balls sit in spherical nests which are
> supposed to keep the balls tight as long as the spring load
> is strong enough. Well, I see the nests pretty worn compared to
> the ones of another broken reader head I could compare with.

I haven't run across this exact design, but it is generally similar
to how many of the scales are made. Being Mitutoyo, it makes sense
that it is a little better than some others. if the linkage can't move
freely, or there is backlash, then you have to fix that.

> The reader head itself is a pair of pcboards that is equipped
> with 5 roller bearings. The upper one has 3 (kind of) LED's
> the lower one has a glass surface with 3 gratings
> and below are 3 light sensors. I guess that the central
> one is the main and count grating and the side ones are
> 25% offset of a stepping pass.(90 degree talking in a
> sine wave comparison) These ones probably only give the direction
> for counting.

Well, 3 channels is pretty common, 2 channels in quadrature for
the standard encoder function, and an extra, index channel for
precise home position. The index channel may be obvious, as
it might have one mark every 100 mm or something like that.
The quadrature channels read the same part of the scale, but the
analyzer gratings are offset by 90 degrees. The 5 roller bearings
need to travel smoothly, with minimal friction, so the connecting link
doesn't change length.

> I will clean up once again thoroughly and try somehow to tighten
> the connecting link.

Make sure it is not really due to binding of one of the rollers, or
perhaps one of the rollers has been knocked out of alignment,
causing the head to roll crooked across the scale and bind.
I think this is likely to be the cause of the trouble.

Jon