euspen Int. Topical Conference, Aachen, Germany, May 2003. Extended abstract.

High precision, low force 3D touch probe for

measurements on small objects
F. Meli, M. Fracheboud*, S. Bottinelli**, M. Bieri, R. Thalmann, J-M.
Breguet*, R. Clavel*
Swiss Federal Office of Metrology and Accreditation (METAS), 3003 Bern-Wabern,
Switzerland (www.metas.ch).
* Institut de Production et Robotique (IPR-LSRO), EPFL, 1015 Lausanne,
Switzerland (isro.epfl.ch).
** MECARTEX, Z.I. Zandone, 6616 Losone, Switzerland (www.mecartex.ch).

Abstract
A new 3D touch probe for coordinate measuring machines (CMM) with exchangeable
probes with spheres in the diameter range of 0.1 mm to 0.3 mm and probing forces
below 0.5 mN was developed. The device is based on parallel kinematics with flexure
hinges and is manufactured out of a single piece of aluminium. All rotational
movements of the probing sphere are blocked The remaining possible translational
motion is separated into its xyz-components, which are each measured by an
inductive sensor. All axes have the same orientation with respect to gravity. The
stiffness is isotropic with a value of only 20 mN/mm, while the effective moving mass
is about 7 g. First experiments with the 3D touch probe were performed on a linear
measuring machine equipped with a laser interferometer. The standard deviation of
repeated measurements, e.g. the difference between left and right probing on a
5 mm gauge block, was in the order of 5 nm.

Introduction
In the last years CMMs have become versatile and widespread metrology tools.
Today's CMMs can efficiently perform very complex measurement tasks. However,
with the ongoing miniaturisation in the mechanical and optical production there is a
new demand for highly accurate geometrical measurements on small parts. New
instruments should have low uncertainties and probe the objects with very small
spheres using very low contact forces. Up to now, the limiting factor for the
application of CMMs on small parts was mainly the probe head and therefore new

Figure 1: Measuring principle of a CMM (left) and probing data used to determine
the force free contact point of the probing sphere (right).
euspen Int. Topical Conference, Aachen, Germany, May 2003. Extended abstract.
developments are needed [1, 2].
The point of contact between the probing sphere and the sample surface should be
obtained by measuring probe deflections at several CMM positions. This allows
extrapolation to zero deflection where also the contact force is zero (Fig. 1 right). In
the context of a Swiss TOP NANO 21 research project the EPFL, the industrial
partner MECARTEX and METAS developed a new 3D touch probe for
measurements on small objects.

Probing force and surface damage

The relatively large probing forces of conventional probe systems may damage the
surface of the work piece and thus falsify the measurement result. Figure 2 shows on
the left an atomic force microscope image of such a plastic surface deformation on
an aluminium test piece. It was probed with a conventional 3D touch probe with a
sphere of 0.6 mm diameter resulting in an indent with 330 nm depth.
If a probe sphere with a diameter in the order of 0.1 mm is to be used without leaving
any permanent surface indentation, the force needs to be about 100 times smaller
than with today’s commercial probe systems (Fig. 2 right). This means the stiffness of
the probing system has to be very small e.g. 20 mN/mm to allow still some probe
deflection at these low forces (typ. 20 µm). The situation is even more critical with
respect to dynamic forces that act when the sphere hits the surface upon first
contact. Therefore the effective moving mass needs also to be as small as possible
to allow reasonable approach speeds.

Construction
Based on parallelograms and flexure hinges a new kinematic structure was designed
for the probe head. This structure leaves the probing sphere exactly three degrees of
freedom. The rotational movements are blocked and the translational motion is
separated in its xyz-components, which are each measured by an inductive sensor
(Fig 3 left). All axes have the same orientation with respect to gravity and provide the
same probing force in all directions. The main part of the structure is manufactured
out of a single piece of aluminium using electro discharge machining. Therefore the
most critical part does not need to be assembled. The flexure hinges have a
thickness of only 60 µm resulting in a stiffness of 20 mN/mm. The effective moving
mass is 7 g. Due to the low stiffness, the deformation caused by gravity needs to be

Figure 2: Plastic deformation on an aluminium surface after probing with a

conventional probe (left) and maximal admissible probing forces for
various sphere diameters at 1000 MPa contact pressure (right) [3].
euspen Int. Topical Conference, Aachen, Germany, May 2003. Extended abstract.

Magnet

Sensor

Probe

Figure 3: Kinematic structure of the 3D touch probe (left), magnetic holding of the
probing element (centre) and mechanical filter (right).

compensated. For this purpose an adjustable system with permanent magnets was
developed.
The measurement range is ±0.2 mm while the mechanical limits allow a range of
±0.5 mm. The probing element is magnetically attached to the head and positioned
by means of three balls in three grooves (Fig. 3 centre). The magnetic holding of the
probing element allows an easy replacement and acts also as a mechanical fuse.
Therefore the handling of this highly sensitive device remains quite easy.
A small moving mass is important to keep the dynamic contact forces low while
maintaining reasonable approach speeds. Model calculations showed that the
effective mass (7g) of the probe is still too high. A small additional mechanical filter
element was developed to reduce the effect of dynamic forces. Its stiffness is almost
equal in all directions and roughly several times higher than that of the probe head
but it has a very low effective mass (Fig. 3 right). Typical approach velocity is thus
1 mm/s.

Experimental Results
The full performance of the new probe
can only be measured on a CMM with
equal or better performance than the
probe itself. However we had no such
ultra precision CMM available.
Therefore the first experiments with the
new 3D touch probe were performed on
a linear measuring machine (LMM)
equipped with a laser interferometer
(Fig 4). In this way only probings in a
horizontal plane could be made.
However, due to the special orientation
of the probe coordinate system all 3
sensors of the probe head were
Figure 4: Experimental arrangement for involved in these test measurements.
the probe test on the LMM.
euspen Int. Topical Conference, Aachen, Germany, May 2003. Extended abstract.

Figure 5: Deviation from linearity of 5 touch probe measurements for the left and
right side of the gauge block.

The test consisted of probing a known 5 mm gauge block on the left and right side.
The difference between the two points minus the gauge block length is the probe
constant, which is essentially the sphere diameter. The repeatability of the probe
constant is an important parameter for a 3D touch probe. In our experiments the
standard deviation of 5 such measurements was always in the order of 5 nm. This
value even includes interferometer noise and machine instability like vibration and
drift. Also for large probe deflections up to 150 µm the linearity of the probe signal
remains very good (Fig. 5).

Conclusions
A new 3D touch probe for CMMs with exchangeable probes and low probing forces
was developed and a patent has been filed. The innovative design is based on a
parallel kinematics with flexure hinges. First test measurements were very
successful. With probing forces smaller than 0.5 mN the repeatability was in the
order of 5 nm.
However the full characterisation of the probe needs a CMM with comparable
performance. METAS plans therefore to incorporate the probe head into a new ultra
precision CMM [4]. The goal is to offer calibration and measurement services for
small parts up to a size of about 50 mm in the near future.

References
[1] U. Brand, S. Cao, W. Hoffmann, T. Kleine-Besten, P.Pornnoppadol, S.
Büttgenbach, Proc. of 2nd euspen Int. Conf., p266-269, May 2001, Turin, Italy.
[2] H. Haitjema, J.K. van Seggelen, P.H.J. Schellekens, W.O. Priel, E. Puik, Proc. of
Memstand Int. Conf., p218-228, February 2003, Barcelona, Spain.
[3] W.P. van Vliet, P.H. Schellekens; Annals of the CIRP 45, 483-487, 1996.
[4] T. Ruijl, J. Franse, J. van Eijk, "Ultra precision CMM aiming for the ultimate
concept", Proc. of 2nd euspen Int. Conf., p234-237, May 2001, Turin, Italy.