X-ray Image Scanner Teardown. Philips / Fujifilm PCR Eleva S (3 part series)

[Mads Barnkob]

Teardown of a Philips / Fujifilm PCR Eleva S x-ray image scanner / digitizer. Also known as Computerized Radiography (CR), it fulfills the job of scanning a x-ray exposed image plate with a laser beam and extract each pixel illumination with a photo multiplier tube setup. It can scan 78 plate per hour with a general resolution of 3520 x 4280 pixels in 12 bits per pixel gray scale.

/>
The system is built up on Altera Flex (EPF6016QC208-2N), Altera Cyclone (EP1C12Q240C8N) and Altera Max (EPM7128SQC160-10N) CPLDs. CPLD is a Complex Programmable Logic Devices and is a sort of device that is one step closer to the hardware layer, than a FPGA.



The stepper motor drivers are Sanyo STK672-050 packages. There is 4 identical stepper motors from SHINANO KENSHI, model STP-59D3002, with properties 1.1Ω 1.8°/step is certainly reuseable for other projects. The small diaphragm vacuum pump and DC gear motors can also come in handy for fun stuff.



The harddisk was destroyed with a hammer, while the PC was running. To my big surprise it actually took a while for Windows to realize that the drive was bend 45 degrees. I eventually had to command the machine to reboot in order to get it to trigger on a faulty harddisk.

[Mads Barnkob]

Reverse engineering of laser scanner module from a Philips / Fujifilm PCR Eleva S x-ray image scanner / digitizer.

Working with lasers require proper safety equipment to avoid damages to eyes or other parts.

/>
A laser scanner consists of 3 main parts, laser, moving mirrors and optical lenses. The laser diode is marked with J9788H2, which is an unknown laser diode part number. The laser diode driver LDD23A is quite complex with multiple power supplies and diagnostics built into it.



The rotating hexagonal mirror is made by Sanyo, marked LS8820. It uses a LB11872H driver IC for three-phase brushless motors. The timing signals for IN1 and IN2 signals on the LB11872H is supplied by a 74HC4020 14-stage binary ripple counter.





The laser diode is driven with a LM317 constant current power supply. Configured with 80 Ohm and a 1000 Ohm potentiometer, it supplies between 1 to 15 mA. The original laser diode is most likely a 633 nm red diode rated for around 5-10 mW.

I used a green laser pen for demonstrating the optical parts is a standard office pen, to keep the demonstration as safe as possible, driving a unknown laser diode is not safe.

[Mads Barnkob]

Reverse engineering of the Fujifilm PMT26A PMT module that contains the photomultiplier tube, high voltage power supply and analog amplifier

/>
The Fujifilm PMT26A board has two connectors, both goes to the SCN23A card in the main computer of the Philips / Fujifilm PCR Eleva S scanner. CN1 is a 34-pin flatband pin header and CN2 is a coaxial output signal from the output amplifier.



The output amplifier is shielded, the differential amplifiers on the board is using +12/-12VDC from a 88M12T and 79M12T regulators. The +15/-15VDC input goes directly to the HV module. The HV module is a YE DHFF-001 from 2008



The Anode output signal from the Hamamatsu R6233 photomultiplier tube goes to a Analog Devices AD8610. The AD8610 is a precision, very low noise, wide bandwidth, JFET operational amplifier.

The next amplifier IC in the chain, which seems to handle Offset and Gain potentiometers, is the Analog Devices AD744. The AD744 is a precision BiFET operational amplifier.

A sampling IC from Analog Devices ADG417 is placed to the side of the signal chain, indicating that it might have to do with sample and hold, test, calibration or blue LED referencing. The ADG417 is a LC2MOS precision mini-DIP analog switch.

The output amplifier next to the CN2 output coaxial connector is a Analog Devices AD817, which is a high speed, low power, wide supply range amplifier. It can supply a minimum of 50 mA at up to 50 MHz.

Test setup with BC408 scintillation material



The following 3 waveforms have a clear muon signature with a steep pulse and a subsequent event of 2/3 energy about 2 to 10 microseconds later. The mean life span for a muon is 2.2 microseconds.



Something that is properly not a muon, but interesting signature