1988–1994

Studies and doctorate

Diploma thesis at Tecmath AG in Kaiserslautern on data reduction methods in fatigue life analysis. Then a doctorate at the University of Bremen, on non-destructive testing, digital image processing and holographic interferometry; the thesis appeared in 1994 with VDI-Verlag, Fortschritt-Berichte series 8, no. 398.

1989–1995

Research associate at BIAS, Bremen

At the Bremen Institute of Applied Beam Technology, coupling numerical methods with measured results — hybrid working long before the term became common. Concretely: reconciling finite element calculations with displacement measurements from holographic interferometry. The model predicts a displacement field, the interferogram measures it, and wherever the two part company the model is wrong. More than fifteen publications on it, among them a chapter on holographic interferometry of adhesive bonds in the Handbuch Fertigungstechnologie Kleben.

after the doctorate

Research stay at Worcester Polytechnic Institute

A post-doc at WPI in Worcester, Massachusetts. The subject remained the combination of holographic interferometry and finite element calculation — the same pairing of measurement and model that would later carry the gearbox acoustics work.

1995–2022

Manager NVH & Dynamics, ZF Friedrichshafen

Twenty-six years of gearbox and drivetrain acoustics: head of the group-wide acoustics expert team, responsible for fundamental research in NVH. The work was NVH method development — automated evaluation, new evaluation methods, reconciling simulation with measurement, and mapping the limits of both: where a model still holds, and where an evaluation method stops carrying. Alongside it, transferring that knowledge into the business units, from designing the training courses to teaching them. The applications ran from the noise optimisation of commercial-vehicle gearboxes through the acoustic quality of wind-turbine gearboxes to system acoustics across component and vehicle boundaries.

from 2002

AutoPlotter — automated evaluation as a product

Alongside the employment I have run my own engineering office since 2002, at first as a second occupation. Out of it came AutoPlotter: a software product for the automated evaluation of noise measurements, built on PAK, the measurement software from Müller-BBM. Its sources have since been sold to ZF.

2008–2009

Lectureship in machine acoustics

Lecturer for machine acoustics at Ravensburg-Weingarten University of Applied Sciences.

2011–2015

End-of-line testing and order-based evaluation

Locating acoustically suspicious gearbox bearings on the end-of-line bench, and diagnosing them within the seconds a production line allows. In 2015, at the 22nd ICSV in Florence, deriving the relative transfer function and the relative tooth mesh excitation from ramp measurements — the direct predecessor of today’s evaluation.

2015–2022

Chair of the FVA working group on NVH

Seven years chairing the NVH working group of the Forschungsvereinigung Antriebstechnik in Frankfurt — gearbox acoustics not as one company’s problem but as the industry’s shared question.

2016–2021

Patents on automated cause attribution

Among them a method that automatically attributes possible causer sources to an acoustic anomaly of a rotation-synchronous noise source, a method for detecting defective bearings, and structure-borne and motion sensing built from electroactive composite material.

since 2022

Ingenieurbüro Dr. Thomas Bischof, full-time

Since 2022 the office in Graben-Neudorf has been my full-time work: tooling that automates evaluation work, automated assessment of acoustic measurements, root cause analysis of suspicious machine noise. The method now sits in s2m, which holds the measurement data, the kinematic models and every analysis run, and makes a result reproducible months later.

Understand the complaint

What is heard, where, under which operating conditions — and who decided it is a problem. A noise that only appears in coast-down is a different investigation from one that scales with load.

Check what the data can carry

Before analysing anything: is shaft angle available, is the run repeatable, does the speed range cover the suspected orders. If not, that is the first deliverable.

Model the gear train

The kinematics define which orders each element can produce. This is what turns a spectrum into a list of testable hypotheses instead of a picture to stare at.

Resolve and separate

Shaft-synchronous analysis per load-carrying mesh, sidebands included, so excitations that overlap in frequency stop overlapping in the result.

State the conclusion with its limits

Named cause, operating range it holds for, separation quality behind it — and an explicit note where the data does not decide.