Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Steve Paddock wrote:

 > 2.      How difficult would this be to implement on an existing
 > microscope, Zeiss, BioRad, Leica etc....???

Steve,

Implementing software to control a confocal is a major effort. The
confocal microscopes I've seen (all laser scanning) use a similar
architecture:

1. A scanning system composed of mirrors steered by galvanometers.
Usually two galvanometers are used, one for the x axis and one for the y
axis. The z axis is controlled by a separate motor. The galvanometers
are driven by scanning waveforms, produced by D/A converters from
digital waveforms supplied by the controlling computer. The waveforms
are computed to cause the galvanometers to scan linearly across the
field of view, or, as an alternative, the scanning system measures the
position of the mirrors as each sample is acquired, and the software
supplies the necessary corrections to produce an image with uniform
linear dimensions.

2. Photomultiplier tubes, usually each connected to an A/D converter.
The software controls gain and offset on the photomultiplier amplifiers.
The software also reads the A/D converters, and must keep up with the
data stream, which can be fast, especially if many channels are involved.

3. A collection of miscellaneous electronics that control the light
sources, the filters, monitor the power supplies, and the like.

The confocal control software handles this hardware through computer
interfaces. The interfaces are usually proprietary PCI bus boards that
require their own device drivers, although I've seen people use
off-the-shelf hardware, on the order of boards from National Instruments.

To write software to control the confocal, you need the specifications
of the microscope hardware and any proprietary computer interfaces.
These specifications are generally not supplied by the manufacturer,
because the manufacturer can't afford to document the hardware in the
detail required so that someone can write software without the help of a
design engineer. The manufacturer can't afford to have the design
engineers talking to customers about how to control the hardware,
because then the engineers can't work on the next generation instruments.

After you have the specifications, you get to write:

a. The kernel-level device drivers for the interface hardware. Most
manufacturers use one of several commercial interface chips on their
proprietary PCI bus board. Since much of the device driver is related to
the interface chip, at least you start out with example drivers from the
chip vendor (perhaps for the wrong operating system) and a detailed chip
manual from the vendor that is reasonably accurate.

b. The low-level software that controls the confocal, handling the
control electronics, the scanning system, and the photomultiplier tubes.

c. The software that operates the confocal, acquiring images, processing
them, and storing results.

d. The user interface. You can judge the size of the user interface
yourself from the programs supplied by the manufacturer.

Creating this software is a major effort.

-- Dan

--
Dan Brown
Bruxton Corporation
http://www.bruxton.com
206 782-8862 x103