High Performance, Low Cost…

Stacks Image 108

SITe Imaging System

Here we describe an area array detection system constructed using a high performance scientific grade CCD image sensor. The sensor, a SITe STA-001A, was removed from instrumentation that had been previously used elsewhere in the human genome project and had subsequently appeared on the surplus market. This project provided an opportunity to learn about, and use a field programmable gate array (FPGA) in a scientific instrument application.

This imaging system is designed to be placed at the focal plane of an imaging spectrograph (see below) and connects to a PC or laptop via a high-speed USB interface. The graphical interface for controlling the instrument is implemented using the LabVIEW™ software package from National Instruments.

The photograph above shows the area array detection system with its cover removed and with the CCD chip (format is 1132 x 330 pixels) clearly visible atop a heatsink “island” in the upper part of the picture. The dark blue ezFPGA circuit board (Dallas Logic) and its associated Cyclone® FPGA can be seen in the lower compartment. The board in the upper compartment mates to that in the lower by a 96-way edge connector that brings signals from the FPGA to a number of clock driver chips located around the CCD.

The surface mounted chip on the carrier board near the bottom of the upper PCB is an analog-to-digital converter that converts image data into digital format. The digital lines from this chip in turn feed back to the FPGA on the lower PCB. Not visible in this photo are two static RAM chips that are located beneath the FPGA module on the lower PCB. These devices store the image prior to an upload to the host PC via the USB module that can be seen at the lower right hand side of the lower PCB.

Stacks Image 105
An FPGA (Field-Programmable Gate Array) is capable of operation at extremely high clock speeds. An FPGA is an ideal choice to provide the core of a complex instrument by virtue of its large number of fully programmable I/O pins. The function of an FPGA is software re-configurable and thus it differs from a normal integrated circuit whose functions are hard-wired during manufacture and cannot be altered later.

The FPGA lies at the heart of the area array system described here. Other sub-systems incorporated in the design include clock generator circuitry to clock images from the CCD, a 16 bit analog-digital converter, a 512k x 16 static RAM (for image storage), and a USB module.

A block diagram of the system is shown at left and indicates the relationship between the EP1C3-8 FGPA and the other components of this imaging system.

CCD-Based Optical Spectrometer

In the photo at left, the SITe imaging system is seen mounted at the focal plane of a SOL Instruments imaging lens spectrograph.

Light enters the spectrometer from the left hand side, with the monochromator slit width being controlled by the micrometer adjustment.

A second micrometer (not visible) moves the spectrum laterally, thereby defining the spectral range that is registered by the CCD imager.

The imager and its associated electronics can be seen in the lower right of the photo.

2D Imaging LabVIEWTM Front Panel

The LabVIEWTM vi developed to run an XMOS Startkit-based controller for the STA-001A imager is shown below. A full description of that system can be found here; this serves as an excellent example to illustrate just how much can be achieved using a low cost XMOS Startkit development board.

Recall that the imager format is 330 rows each containing 1132 pixels. The user specifies a “binning parameter” that enables multiple rows to be combined prior to readout. In the example shown here, the binning parameter has been set to 2 so that pairs of rows are first combined and a total of 330/2=165 rows are read out.

On the right hand side of the screen is the resulting image; here we see the spectral output from a Hg-Ar calibration lamp. In this experiment the imager is placed at the focal plane of a SOL Instruments lens spectrograph.

The trace in the middle of the screen is the wavelength-calibrated spectrum that is obtained by summing up the intensities recorded in all pixels having the same index position within a row.

The vi loads in three wavelength calibration parameters that convert pixel numbers into wavelengths using a quadratic polynomial. The wavelength coverage of the spectrometer can be changed by adjusting a micrometer setting on the spectrograph – in this run the wavelength covers from ~ 520 nm to 980 nm.

Additional user parameters control the exposure time, the gain and the offset of the image acquisitions. A “frame reads” parameter is also provided; this is used to rapidly read out an image prior to an exposure - which has the effect of clearing any residual charges prior to an actual exposure.

In the spectrum shown below there is a small amount of stray light leakage that is more visible at one end of the detector; this gives rise to the sloping baseline that is seen as one moves down towards 500 nm. This could easily be removed by paying more attention to covering the detector head. For serious work with any CCD one should ideally record a dark “frame” using identical imager settings (exposure time, gain etc) and subtract this from the “light” frame to yield a dark-corrected spectrum - but this has not been done here.