SD - 3 - 20e Throughput For MT

Special print: SD 3/20e
Handling of serial products for the magnetic

particle crack detection
Special print from
October 2004
An information composed by
Dr. W. Weber, Dr. W. A. K. Deutsch, Dr. V. Schuster, F. Bartholomai
Summary
In the automobile and their supplier industry, high quantities of uniform workpieces have to be inspected,
partially piece for piece, under a high cost pressure. The cost pressure has not only influence on the used testing
technique, but also on the required throughput-rate.
In non-destructive testing, crack detection plays an important role. For testing high quantities of specimens with
simple geometry like rods, tubes or bolts, the ultrasound- or eddy current technique are preferably used.
However for ferromagnetic specimens with complicated geometry, magnetic particle inspection (MPI), is the
most suitable technique.
Based on practical examples, several MPI techniques with various handling procedures are shown in order to
minimise the testing cycle.
(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005
KARL DEUTSCH Prüf- und Messgerätebau GmbH + Co KG x Otto-Hausmann-Ring 101 x D-42115 Wuppertal x Tel. (+49 -202) 71 92 - 0 x Fax (+49 -202) 71 49 32 x e-mail: [email protected]
Seite 1 von 11
Handling of serial products for the magnetic particle crack detection
Dr. W. Weber, Dr. W. A. K. Deutsch, Dr. V. Schuster, F. Bartholomai
KARL DEUTSCH, Wuppertal, Germany
1 Introduction
In the automobile and their supplier industry, high quantities of uniform workpieces have to be inspected, partially
piece for piece, under a high cost pressure. The cost pressure has not only influence on the used testing technique,
but also on the required throughput-rate.
In non-destructive testing, crack detection plays an important role. For testing high quantities of specimens with
simple geometry like rods, tubes or bolts, the ultrasound- or eddy current technique are preferably used. However
for ferromagnetic specimens with complicated geometry, magnetic particle inspection (MPI), is the most suitable
technique.
Based on practical examples, several MPI techniques with various handling procedures are shown in order to
minimise the testing cycle.
The current state of the art to inspect for cracks in all directions during one cycle, is not specially discussed in this
paper. It is only mentioned, that this is achieved by applying a combination of two phase-shifted alternating
magnetic fields, which are perpendicular orientated to each other.
The procedure of MPI consists of several steps: the transportation of the specimens, the testing process in order to
generate the crack indications and finally the visual inspection.
2 Lowest level of automation

For an inspection system with a comparably low level of automation, the operator manually places the specimen
into the machine and the inspection process is started by activating a food or hand switch. Figure 1 shows such a
standard machine with 2 contacts. The inspection process has a fixed time frame and it runs automatically. This
process includes the clamping (approx. 0,5 sec) of the specimen, the magnetizing and the spraying of the inspection
medium (approx. 4 sec), the re-magnetizing (approx. 1,5 sec), the demagnetization (approx. 1,5 sec.) and the
release (approx. 0,5 sec) of the clamping device. Afterwards, the operator removes the specimen and performs the
visual evaluation by using ultraviolet light in order to check, if crack indications have appeared or not. Workpieces
without indications are put e.g. into a skeleton container.
In this case, the automation is given by the automated time frame of the inspection process. Including
demagnetization, the inspection process requires approximately 8 seconds. Additionally, the time for loading and
unloading of the specimen and the time required for the visual evaluation have to be considered. With such a
system, you cannot expect to achieve a high throughput-rate. Workpieces with a complex shape like steering
knuckles, have to be clamped several times in such a machine, if it is required to check the entire surface of the
specimen. This is an extremely time-consuming procedure. On the other hand it is possible to improve the
throughput-rate by testing more than one workpiece during one cycle, if the geometry of these workpieces are quite
simple (Figure 2).
Seite 2 von 11
Figure 1: Standard machine with 2 contacts Figure 2: Testing of 5 gear rim in one cycle.
3 Possibilities and influences

The achievable degree of automation (maximum reduction of the cycle time) depends on many
influences/conditions e.g.: weight, size and geometry of the workpiece. Also the required scope of inspection, the
consideration of additional inspection techniques, the quantity of available operators and also the given local
preconditions define the extent of the highest possible automation level.
In principle, a reduction of the cycle time can be achieved by the following factors:
1. Using special inspection concepts, e.g. machines with more than 2 magnetizing circuits or using a permanent
test cycle (that means: the time gap between two cycles is pre-set by the machine) or by performing the
demagnetization outside the machine.
2. Transport devices for loading and unloading the specimens.
3. Special machines, designed for a special test problem
4. Automated evaluation instead of operator evaluation
5. A combination of the above listed items
In the following, we will discuss some possibilities based on some practical examples.
4 External Demagnetization and visual evaluation

The demagnetization performed inside the machine during the testing process requires about 1,5 seconds. For the
procedure described above, this is about 17 % of the time needed for the entire testing process. The conclusion: An
external demagnetization (that means outside the machine) reduces the cycle time considerably. With external
demagnetization, the next workpiece can already be tested in the machine. The external demagnetization is mostly
performed by moving the specimen through a separate coil, fed by AC-current. The workpiece can be moved
through the coil on a conveyor belt (Figure 3), or the specimen slides through an inclined adjusted coil.
However, for the inspection of big or bulky pieces the operators very often use handling systems (crane etc.) in
order to load and unload the machine. In these cases, the demagnetization inside the machine is preferable, as it is
less time consuming as the external one. When it is required to demagnetize the specimen after the visual
inspection, the automated testing cycle does not include the demagnetization. In this case, the visual evaluation is
performed inside the machine and afterwards the demagnetization is activated by pressing a respective key on the
operator panel of the machine.
Seite 3 von 11
Figure 3: The workpiece moves through the coil on a conveyor belt.
In general, high through-put rates should not affect the inspection safety. Depending on the inspection result (clear
or doubtful), the required time for the visual evaluation process can vary considerably. It is therefore critical to
integrate the visual inspection process into the fixed inspection cycle. This is avoided by building a buffer in front
of the visual evaluation station. Many years of experience have shown, that the use of slowly moving turn tables
with one or several inspectors are favourable for the visual evaluation.
5 Multiple contacs
If specimens with complicated geometry such as steering knuckles, have to be inspected in a standard machine with
two contacts, the test procedure would be very time consuming. In order to find cracks in all directions on the
whole surface, the specimen has to be clamped several times, each time between respective two other ends of the
specimen and after each magnetizing process you have to make the respective visual inspection. Besides the
considerable inspection time, the crack identification deteriorates from one visual inspection to the next as the
fluorescent background on the specimen becomes stronger and stronger.
Figure 4: Machine with multiple contacts, here testing of a steering knuckel
In order to have the capability to inspect complicated workpieces during an acceptable time with MPI, special
machines with more than 2 contacts have been developed. Figure 4 shows a steering knuckle positioned in a
workpiece holder. The 3 laminated horizontal yokes conduct the field flow and also 3 independently adjustable
Seite 4 von 11
currents into the specimen. In addition, two independently adjustable currents are fed into the workpiece via two
vertical contacts coming from the top. The set-up of the current circuits can be freely adjusted with the machine.
With such a 5-contact-machine, the entire surface of workpieces with complicated geometry, such as steering
knuckles, can be inspected for cracks in all directions in one shot (Figure 5, left).
During the clamping process, each of the contacts can be pneumatically moved up to 100 mm until the workpiece is
properly clamped by the contacts. In order to avoid the jumping of the specimen out of its position during the
clamping process, the contact pressure of each contact and its initial position are adjustable. Also time delays to
coordinate the clamping, can be chosen (Figure 5, right).
H5, I 5
H1, I 1 H2, I 2
I6
H3, I 3
H4, I 4
Figure 5: Different contacts for field and current flux (left); Memory (right).
In order to test specimens with simple geometry, this versatile machine is also applicable. Two vertical currents and
one of the horizontal yokes on the left side have to be deactivated. In this case a standard 2 contact machine is
obtained.
6 Transportation system: workpiece ejector

A very simple but also a very effective system for the reduction of the cycle time is the application of workpiece
ejectors. Figure 6 shows a machine with 3 contacts and a workpiece ejector suitable for transverse links. The
operator loads the machine with the transverse link and activates the testing process by pressing the foot switch.
After releasing the specimen, the workpiece ejector is pneumatically tilted backwards. The transverse link slides
out of the machine, down on a visual inspection table or a conveyor belt leading to the visual inspection station.
Afterwards the specimen is demagnetized by moving it through a coil. During the inspection, at least two operators
are involved, one operates the machine, one or more operators are responsible for the visual inspection. The ejector
tilting and the movement back into its loading position is integrated in the automatic test cycle.
Seite 5 von 11
Figure 6: Machine with 3 contacts and workpiece ejector, here testing of a transverse link
The ejector principally consists of two parts. The pneumatically tiltable ejector ground plate and a special
workpiece holder specially designed for the respective specimen. As it is possible to quickly mount other
workpiece holders on the ground plate, this set-up is suitable for a wide range of different products.
In order to meet the safety regulations, a two-hand-operation is often used. If this is not desired, light barriers or a
light curtain can be built-in in front of the machine.
Of course, the ejector can only be re-loaded with the next specimen, when the empty ejector has been moved back
to its loading position. The loading is integrated into the cycle, in other words, the ejectors can not be re-loaded
while the testing is performed.
Machines equipped with an ejector for workpieces with complex geometry (e.g. steering knuckles, transverse links
or pistons) offer a through-put of 8 seconds per specimen.
As the operator does not need to take the workpiece out of the machine during the testing procedure, he is able to
prepare the next workpiece or perform other tasks. Examples are a visual evaluation in order to find geometry
defects, an eddy current testing to check the structure, or straighten the workpiece in a press.
7 Transportation system: Chain conveyor with/without buffer

The chain conveyor enables the loading with specimens independent (or dependent) of the testing cycle, which
means that during the testing process the workpiece holders of the conveyor can be loaded. Figure 7 shows a chain
conveyor, which is mounted to a standard machine.
Fuel injection pump housings are inspected for cracks in all directions on the entire surface. The small size of these
parts allows for testing 2 pieces during one test cycle (Figure 8). During the clamping, both outer contacts can be
pneumatically moved up to 100 mm. Two specimens are pressed against an additional middle contact. Thus, the
current flows from one specimen through the middle contact to the other specimen. Besides the current flow, also
the field flow is lead into both workpieces, following the same path.
In general up to three specimens arranged in a row can be checked in one test cycle depending on the geometry. In
that case, special movable contacts between the specimens have to be provided.
In order to increase the through-put, the machine can be switched over to the permanent cycle. Thus, no
unnecessary intermittences will occur and the time parameters of the machine entered into the electrical control,
determine the through-put. After the declamping, the next workpiece holder filled with two specimens
automatically moves to the magnetizing station while two already tested specimens slide on a visual inspection
table. The demagnetizing is performed by putting them on a conveyor belt which moves first through a
demagnetizing coil. Afterwards this conveyor could lead through a second coil in order to test its structure with
eddy current.
Seite 6 von 11
H, I H, I
H, I H, I
Figure 7: Standard machine with 2 contacts, a chain conveyor and an additional middle contact, here testing of 2
fuel injection pump housings in one cycle and manual loading of the chain conveyor
Figure 8: Testing of 2 fuel injection pump housings in one cycle.
The chain conveyor contains a lot of identical workpiece holders. After the operator has loaded several holders he
can perform other tasks for a short time, e.g. to lay out skeleton containers with oil papers. When most of the
guidances are empty, the operator has to load them again in order to avoid that the testing is interrupted.
Comparable to the use of a workpiece ejector, the use of a chain conveyor reduces the cycle time by further
2 seconds. Typically, one operator loads the machine and two others are involved in the visual inspection process.
The cycle time is 6 seconds for one specimen in the special case of testing the fuel injection pump housings. From
the machine side, a test cycle of 3-4 seconds is possible.
When the workpiece geometry changes, the chain conveyor can be equipped with new guidances. It is also possible
to remove the entire chain conveyor. Again, a standard machine with standard properties is obtained.
As the length of the conveyor belt is usually limited for economical reasons, the number of loaded specimens is
also limited. To further automate the entire loading process and to give the operator more time to do other tasks, a
specimen buffer with a transfer station ca be used. The operator manually places high numbers of specimens into
the buffer. The transfer station loads them into the conveyor belt without operator assistance.
Seite 7 von 11
Figure 9: 8 contact mashine (Testing of 8 workpieces in one cycle; left side not visible).
8 Integration into the production line

Usually manufacturers with a relative large variety of specimen types, as for example forges, have no conveyor
belts which are equipped with special workpiece guidances. Therefore, the test pieces are transported for example
in skeleton containers to the test machine which usually requires an operator to manually load the machine.
In product lines, in which always the same product is produced over many years, in most cases conveyors with
special workpiece guidances are used. In this case, the degree of automation of the magnetic particle crack
detection process can be further improved by linking the chain conveyor or the workpiece ejector with the
production line. They are automatically loaded by a suitable transfer station such as a robot or a system consisting
of a combination of a separator and a stopper. It is also possible to design a special machine around an already
existing transport system.
If a MPI is machine linked to a production line, its cycle time determines the cycle time of the chain conveyer or
the workpiece ejector. In order to provide sufficient time for the visual inspection process, a sufficient quantity of
operators for the visual evaluation and a transport system which transports the good and bad workpieces out of the
viewing area must be available. Such a linked machine operates in a permanent cycle, the operators always have be
able to interrupt the test cycle. The actual cycle is always finished before the machine is stopped. If other previous
machines can not be stopped, a buffer must be integrated into the line.
Figure 10: Buffer for steering racks
Sometimes the usage of a buffer is necessary due to other reasons. For example during the production of steering
racks, the specimens leave a previous polishing machine without any constant cycle. In order to load the crack MPI
Seite 8 von 11
machine with a constant cycle, a buffer is inserted between both machines. Figure 10 shows such a buffer, a
transfer station combines the buffer with a chain conveyor.
9 Special machines
In the examples given so far, we have basically used standard machines, which were modified in order to provide
test cycles as short as possible. After removing the modifications, standard machines are obtained. An adjustment
for different specimens is then easily possible.
In Figure 11 a special machine for detecting cracks in wheel hubs with outer diameters from approximately 80 mm
to 160 mm is shown. 2 wheel hubs are tested in one cycle. A double gripper loads the machine which operates in a
permanent cycle. The magnetizing station consists of 2 vertical positioned magnetizing bars, which are
pneumatically moved through the wheel hub during the testing process. Each of the two operators - both
performing the visual evaluation - can interrupt the permanent cycle by pressing a key. After the current cycle is
finished, the entire machine stops. In this machine, two additional testing techniques are applied before the
magnetic crack detection is performed. First, a testing of geometry and afterwards a testing of the structure.
Therefore, 3 different testing techniques are involved in one test cycle resulting in a cycle time for one wheel hub
of 5 seconds. The wheel hubs slide directly onto the visual evaluation table. After the visual evaluation, they are
demagnetized by sliding through an inclined coil.
In this case the additional geometry- and structure inspection determine the arrangement and the design of the
workpiece holders. Only a special machine, suitable just for this application, would solve this complicated
inspection task. A modification of a standard machine would not be possible.
Figure 11: Special machine with integrated structure and geometry testing (see the middle extract above), 2 wheel
hubs in one cycle.
10 Robotic loading
Figure 7 shows a 1800 mm long and 45 kg heavy crank shaft while it is unloaded from a crack detection machine.
Because of the length of the crank shaft, the longitudinal magnetization is performed by a moving coil. The coil is
seen on the left side of the machine.
The spraying is performed by a spraying device, which is mounted to the coil. The coil velocity is about 100
mm/sec and can be increased up to 300 mm/sec. A conveyor belt with workpiece holders transports the crank shafts
Seite 9 von 11
to a linking station. There, a robot takes the crank shaft and puts it into the machine. The demagnetization is
directly performed after the re-magnetization inside the machine.
Afterwards, a second robot takes it out (Figure 12) and puts it into a workpiece holder, which is mounted to a turn
table. A second workpiece holder is mounted at the other end of this table. The viewing area and the operator are
lockated at the back side of the turn table (not visible in Figure 12). Two visual evaluation stations are present, each
equipped with a turn table to increase the through-put. If the workpiece holders on the both tables are occupied, the
robot automatically positions the tested crank shafts into a buffer.
The achievable cycle time is determined by the time required for the visual evaluation which is about 30 seconds.
This relatively short cycle time would not be possible without robots.
Figure 12: Robot unloading of a crank shaft from a machine with moving coil (coil here at the left contact side).
11 Automated evaluation
In order to minimise the cycle time and to reduce the labour cost, the automation of the entire evaluation process is
of large interest. Systems with automated evaluation have been presented in the past. The idea is to capture the
specimen surface with cameras which produce a digital image of the specimen surface. Digital image processing is
used to automatically judge about surface indications. An algorithm is implemented which imposes conditions on
the crack geometry.
The edges of the specimen always lead to crack-like indications because the magnetic stray-fields are produced at
those locations. Therefore, such indications have to be suppressed in the flaw evaluation.
The automated inspection is very sensitive to the overall contrast of the digitised image. The test conditions such as
the background illumination and the properties of the magnetic particle agent have to be well controlled.
Two FLUXAMAT systems were delivered to two well-known automotive companies in Germany (Figure 13). It
has been proven that the systems were more reliable than the human inspector. This was true not only for the
detection of various flaw types but also for the percentage of pseudo-indications.
Seite 10 von 11
Figure 13: Automated evaluation with a camera system.
Especially for high throughput rates, the automated systems becomes more economical. On the other hand, the
operator personnel for a conventional system does not need high qualification. For an automated system, at least
one operator needs good background knowledge about the inspection method and the operation of the system. The
future will show, if the automated MPI is the way to go.
12 Literature
[1] J. Thomas Schmidt, Kermit Skeie, Paul McIntire: “Nondestructive testing handbook: Magnetic particle
testing”, American society for nondestructive testing, Inc., 1989
[2] V. Deutsch, W. Morgner, M. Vogt: „Magnetpulver-Rissprüfung - Grundlagen und Praxis“, VDI-Verlag,
Düsseldorf, 1993
[3] V. Deutsch, M. Vogt, M. Platte, V. Schuster, W.A.K. Deutsch: „The Magnetic Particle Crack Detection“,
Castell Publication Inc., Wuppertal, 2002
[4] G. Wahl, V. Deutsch, M. Platte: „Automatisierte Risserkennung“, Kontrolle, Heft 7/8, 1995
[5] V. Deutsch: „Magnetpulver-Rissprüfung sichert Qualität an komplexen Schmiedeteilen“, Maschinenmarkt
98, Heft 25, S. 25-33, 1992
Seite 11 von 11

SD - 3 - 20e Throughput For MT

Uploaded by

Copyright:

Available Formats

SD - 3 - 20e Throughput For MT

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SD - 3 - 20e Throughput For MT

Uploaded by

Copyright:

Available Formats

Special print: SD 3/20e

Handling of serial products for the magnetic

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

2 Lowest level of automation

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

3 Possibilities and influences

4 External Demagnetization and visual evaluation

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

Figure 3: The workpiece moves through the coil on a conveyor belt.

Figure 4: Machine with multiple contacts, here testing of a steering knuckel

6 Transportation system: workpiece ejector

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

7 Transportation system: Chain conveyor with/without buffer

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

Figure 8: Testing of 2 fuel injection pump housings in one cycle.

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

8 Integration into the production line

Figure 10: Buffer for steering racks

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

Figure 13: Automated evaluation with a camera system.

(SV) sd_3_20e Throughput for MT.doc 08.Sep-2005

You might also like