short explanation of some projects:

For further informations (literature or schematics) or in the case of suggestions/ questions send me an email .

A3-truecolor-drum-scanner with parallel-port-interface

It was developed from a defect 9-needle-printer (possibly a NEC P7) and the idea to make something interesting with the new blue LED which just became available. The basis of the circuit is an interface for the parallel port which is emulating the PC's ISA-bus and comes with an object-oriented software-interface. A true-color-LED (red+green+blue in one package) in the place of the printer-head illuminates the object. The reflected light is guarded to the electronics via glass-fibre. The data (8 bit per color) are converted from truecolor to 256-color pseudo-color (VGA) on the fly. The responsible object was developed by me, because I had no hardware to view truecolor at that time.
The resolution is 60dpi, caused by the "single-pixel-scan" and the stepsize of the head-motor and the paper-drum-motor. A sample picture is the Orang-Utan from a A5-book, another one is the pair of pheasants from a A3-calendar.

solar power station

This is my oldest project. It consists of 15W a-Si-solar cells (approx. 1/3qm, nearly 10 years old) and a 12V-100Ah solar-batterie Varta 82000. It provides power for all of my experiments, electrical tools and some of the lightning. Especially popular in the case of a power failure :)
The electronics implements some tricks. One of them is the active diode (MOSFET), causing only a few mV of voltage drop. At clouds or fog, an automatically activated step-up-converter squeezes the last bit of current out of the cells - even if this can only compensate the requirements of the circuit and the losses of the battery in january. Since summer '97 the cells are turning with the sun to boost the efficiency. A 180°-turn and back-turn only needs 2A-minutes. The gain is considerably greater, because the sun is often shining in the morning or afternoon only - when the light normally is falling flat on the cells. For some hours the gain is a factor of 3 or 4 (e.g. from 50mA to 200mA), and then the requirements of the circuit are compensated in only 30 minutes.
The solar power station is in the stable state of a makeshift - this will change at the earliest with a redesign with microprocessor-control.


This new project intends to do something against the bad lifetime of battery-lights, the usual light source for racing cycles. In the summer with high temperatures and short power-on-periods the lifetime is acceptable, but in spring or autumn with falling temperatures and early darkness you can experience a worst-case lifetime of 15 minutes with cheap batteries, which is in unacceptable.
The use of other light sources should allow better efficiency and higher lifetime. Another aspect is the desired "turbo-operation" with higher optical output for driving in difficult regions. Actually I experiment with miniature mercury-vapour-lamps with power ratings from 1.5W up to 10W (nominal power of a cycle-lamp is 2.4W).
Difficulties lie in the estimation of the optical efficiency, the spectrum and the lamp-lifetime. Other problems are operation under high humidity and operation over a wide range of operating voltages (battery 4-6V, burning voltage 200V, ignition voltage 6kV). Beside this I have to control the discharge, which has several options. Probably a microcontroller will handle the operating point in future versions. Last but not least I need a new reflector, because the lamp really is not a point-light-source.

car alarm system

The funny thing on this is that it seems to be a "pseudo-alarm-system". All you can see are two different coloured LED and a key-switch. The status of the system is reported by different blinking sequences. The switching is not done by simply toggling the switch (surprised ?), but by serially entering a code synchronous to the blinking sequence. This has the major advantage, that a drunken driver is not able to start his car - because he cannot coordinate his motions in the required manner :-)
The circuit consists of a NEC V20, a CMOS-variant of the Intel-8088. Beside RAM, ROM, I/O-chip, overvoltage-protection, buffer-capacitor and piezo-siren is nothing more needed to build a small "XT"-type computer. The circuit-board is tailor-made for the dashboard of a Camaro. It has two layers, vias are manually made, 0.8mm diameter with 0.5mm drilling.
The software is written in 100% assembler - including miniature-BIOS with selftest and shadow-RAM, going further with testing the alarm circles (identifying short-cut, open ends and destroyed armatures), ending with the generation of the blinking sequences and the recognition of the serial codes.
By the way, the alarm system is blinking in the off-mode, too (only somehow different). It can be activated irrevocably even by the silliest thief by simply actuating the switch (hint: let the key stick in the switch).


The air-condition was developed because there are no shady places to park at the university. The car is only heated up to 70°C because the inner air is standing. So I modified the circurity for the fan and rear-window-heating. The fan is now controlled by pulse-width-modulation in 16 steps. With ignition turned off, only the steps with below 25% opeating power are used. A temperature control is activating the fan, after it is released by the optional timer. The result is a temperature below 30°C (programmable). Even with outer temperatures above this, I didn't measure temperatures above 40°C.
While driving you can control the fan via push buttons in 16 steps - or it is controlled automatically. The exchange of the rear-window-heating-switch has the advantage, that it is deactivated automatically by turning the ignition off, so you can safely forget it.
The actual realization (with a ispLSI 1016) is a bit big. The PWM-control for the motor (max. 14A,12V) cannot be operated without cooling metal. By the use of a microcontroller and perhaps with limitation to the usual power steps it should be possible to realize an exchange element for the armatures. Whether I realize it or not is a question of time - if my new car with real air-contition comes early ...

XYZ-table for controlling the neck of the picture and focus

In my CCD-camera is a 1/2"-sensor behind a standard photo-object lens. The object lens produces a picture much bigger than the sensor. By moving the sensor in the XY-plane you can control the visual scene without moving the camera itself. In the same way you can control the focus by moving the sensor in Z-direction. The mechanic is self made with modest tools, because I didn't find usable and cheap and precise mechanics. The special trick is the bearing method, which is simple, cheap and has low tolerance. The control range is approximately 5.5cm for each axis.
Maybe I develop a compacted version of the mechanics, but I suppose that I will need some 3D-CAD for it :-)

"the last ISA-card"

No, I didn't manage the jump to PCI yet. But I'm working on it.
In the meantime I'm developing a general-purpose-ISA-card. Beside normal I/O-operations it will provide data channels with up to 12.5MB/s throughput. Because this is impossible following the ISA-standard, there is some work to do. The first application to use it is my CCD-camera. There is no way to use a standard framegrabber with this non-standard-camera. A specialized card does not make sense at this state of development, either. And standard-grabber does not provide the quality and flexibility I need, and the manufacturer probably doesn't know how or why his driver works, and he won't tell me.

CCD-camera and picture processing software

Prototypes of the camera produce prototypes of control-software. After doing some DOS-programs, the target platform now is OS/2 because of its higher throughput and great possibilities implied by multithreading.
Because the hardware must be ready first, the first public software version will come in a few years. I still have problems in writing drivers. I capitulate with DMA at the moment. Life is hard ...
Beside a picture of me in front of my monitor, operating the camera, there is one of a garage in 300m distance slanting under me. Those who think that I have a tit can watch it eating in front of my balcony window. And for fun there is a picture of a light meter at darkness (approx. 0.1 Lux).
color ? coming soon.

Parking aid

my new car made this necessary...details are confidential at the moment, I'm sorry :-/

Tesla transformer

Well, some time in 1998 my friend Reimund asked me to help him with the construction of a Tesla transformer. Since then I'm hardly doing anything different... A Tesla transformer is a resonant transformer which produces high voltages with high frequencies. We target one mega-volt (1MV=1000000V), the theoretical sparc length should be 5m. Beside some troublesome calculations and simulations with my model generator Tesla the device will get a computer driven measurement system. The Tesla transformer is biased with a 20kV/15kVA 3-phase transformer. I'm curious how the transformer and the computer will behave together... Besides that we'll build a Marx-generator, we target 1.2MV/2.4kJ. A Marx generator produces a short DC-puls. It is used for generation of artificial "norm-lightnings". More infos and pictures can be found on