Hardware photographs

Photographs of the inside of a scanner, taken for this reference, so that claims about the hardware can be checked rather than taken on trust. They also give an owner something to compare against when they open their own machine.

Everything on this page is one F-135+, serial 16402, photographed from the component side of its main board on 24 August 2026. Getting to that view needs the bottom plate off, which is four security screws, with the scanner unplugged. Nothing here requires the scanner to be powered, and there is no reason to power it with the plate off.

What is missing

Most of the scanner. Nothing below has been photographed, and none of it can be inferred from what has:

• The CCD board. The most valuable gap. Its part number would give the sensor's total and masked pixel counts from a datasheet, which bears directly on the Offset question in calibration.
• The light board, which differs between the two models: the Plus has a thermoelectric cooler and temperature sensing the base F-135 lacks, which is why their LED current ceilings differ.
• The motor board.
• The reverse of the main board, so whether the two EEPROMs below are the only ones in the scanner is not known.
• Any base F-135. Its main board is PCB #125039A, a separate design rather than a revision of the one here, so no part or marking on this page can be assumed to apply to it.
• The DX sensor area, where an open question about which detectors read the barcode is still unresolved (see DX barcode).

The main board

PCB #125430 REV C, marked © 2005 PAKON, INC.

The board number

Silkscreened along the bottom edge. This is the marking to check before reflashing a controller, because the OEM images are tied to it: 125430A, 125430B and 125430C take hardware versions 03, 04 and 05 respectively, and the base F-135's 125039A takes 02. Putting a Plus image on a base board is the case the OEM readme warns about in capitals. See per-unit data and safety.

The USB bridge

U6, a Cypress CY7C68013A-128AXC: the EZ-USB FX2LP in the 128-pin package, date code 0801. It sits on the right of the board, below and right of the FPGA, with its crystal alongside. This is the bridge everything the host does passes through, described in USB identity and firmware.

The two I2C EEPROMs

U10 and U13, above the FPGA near the top right, both marked 24LC64I / SN 0614 / 1KS. A 24LC64 is 8192 bytes. These are very likely the chips answering at 0x51 and 0x52, though a photograph cannot say which designator carries which address, and the strapping that decides it is not readable here. The consequence for anyone archiving a scanner is on the per-unit data and safety page: if the capacity is right, only a small part of either chip has ever been read.

The round dot on each package marks pin 1, and the address is set by strapping pins 1, 2 and 3. Traces from each chip run to the decoupling capacitor beside it, C19 for U10 and C25 for U13, and the two are wired differently, which is what two addresses on one bus requires. Which is which has not been worked out; it would take a continuity check rather than a photograph, and nothing in the software depends on the answer.

Driving the light

Five Allegro serial-input constant-current LED drivers sit on the main board, not on the light board: A6277ELWT at U8, U25, U26 and U29, and an A6275ELWT at U27. So the illuminant is driven from here, through the connector, rather than by the light board's own controller.

Below them are four X9015U digitally controlled potentiometers at U20, U21, U22 and U23, each with a 1001 resistor alongside. U23 is the leftmost and sits rotated a quarter turn from the other three, so its marking reads horizontally where theirs read vertically. Four of them, on a scanner with four light channels, is a coincidence worth writing down: the OEM keeps Current_R, Current_G, Current_B and Current_Ir per resolution base and film mode, and a constant-current driver of this family sets its output from an external resistance. A digital pot in that position would be how a commanded current becomes an actual one.

That is a reading of the layout, not a traced circuit. Nothing here has been followed with a meter, the pots could as easily trim the A/D rather than the LEDs, and the count could be coincidence. [INFERRED, and weakly: from the part functions and the channel count only.]

The FPGA holds nothing when it is off

Spartan-IIE is SRAM-based, so U18 has to be given its bitstream every time the scanner powers up, exactly as the FX2 is given Pakon7.hex. There is no configuration PROM beside it, and the FX2 image is far too small to contain a 150K-gate bitstream, so the configuration arrives from somewhere else. The OEM package carries a "CCD FPGA file" alongside its controller images, which is the obvious candidate but is not documented here yet.

The consequence for the rest of the reference: the scan-window and gain registers described in calibration are registers in logic the host itself loads, not fixed silicon. Whatever those banks mean is a property of the bitstream in use.

Power

U15 is an LP3965ES-2.5 and U17 an LP3964ES-1.8, National low-dropout regulators supplying the 2.5 V and 1.8 V rails the FPGA and memory need.

Also identified

Read off the same board, without a photograph good enough to reproduce here:

Designator	Part	What it is
U39	LMD18200T	National 3A/55V H-bridge, the transport motor driver
U9	Micron MT46V16M16P-75 F	256 Mbit DDR SDRAM, date code 0708
U5	IDT 71V124 SA12YG	1 Mbit SRAM, date code H0805P
U18	Xilinx XC2S150E, FTG256, speed 7C/6I	Spartan-IIE FPGA, 150K system gates, 256-ball BGA. The scan-window and gain register banks live behind it
U34, U11	125506A, 125507A, both dated 2208	44-pin QFPs whose Pakon number is printed on a white label stuck to the package, not moulded or laser-marked into it, which is why neither carries a manufacturer logo. The original marking is probably underneath. U11 has a second line whose lower edge is just visible and could not be read
D13	B340LA	Schottky rectifier

Marks that have been read but not identified, in case someone recognises one:

Designator	Mark	Note
U14	LTBBW e3	Linear Tech logo. A top-mark code rather than a part number, so decodable from LTC's marking list by anyone who has it. Nothing else is printed on the package
U33	JM83AB over SOOO2VB	National logo, 16-pin. The second line is where a part number belongs but does not resemble one, so this may be house-marked like U34 and U11. Read as letter O rather than zero, though the two are hard to tell apart on a chip mark
U16	CJAB 2995M	National logo, 8-pin, beside the DDR
(bottom right)	X30 over UF400	a vertical part, logo like two overlapping Vs

JM83AF on the U39 motor driver and JM83AB on U33 share a prefix, so that field is a National lot code rather than part of either part number.

Contributing

Photographs of anything under what is missing would be welcome, the CCD board most of all.

Photographs here are by Ali Bosworth (alibosworth) and carry the same CC BY 4.0 licence as the rest of the reference. Contributed photographs need to be compatible with that.