Engram Enterprises


FlashCut CNC IB462 Driver Repair

In the course of developing the BURPS stepper controller interface, I fried two of the IB462 driver modules in a FlashCut CNC Model 4020 drive system by inadvertently connecting the Step inputs to +5V without current regulation. With nothing to lose, I disassembled the drivers and discovered they used mostly common parts and were easily serviced. This process is detailed below to aid anyone in similar circumstances.

The driver inputs can be quickly tested without disassembling the driver. Look for a 1.2-1.75V diode drop between each of the four signal inputs and Logic Ground. If the inputs read open or short, the optocoupler LED on that input has failed. (The Current Adjust input is not isolated; it is referenced to motor power ground.)

flashcutdriverbox.jpg IB462.jpg IB462_pinout.jpg

The FlashCut-branded end caps are self-adhesive aluminum plates, easily pried off. The plastic lid is more difficult to remove. It's possible to pry up on the plastic until it bows up and pops free. Compressing the heatsink fins in a vise also works, but risks damaging the heatsink.

The power semiconductors are fixed to the heatsink with cap head screws which take a 3/32" Allen key. Note that a plastic insulating washer is present on the LM317HVT voltage regulator.

IB462_lid.jpg IB462_opened.jpg

The topology of the driver is very typical. The Enable, Half/Full Step, Step Clock, and CW/CCW inputs are isolated by Avago Technologies HCPL-2630 dual optocouplers, connected common-cathode to Logic Ground and without series resistors; a potentially fatal design flaw. The outputs are routed directly to an STMicroelectronics L297 stepper motor controller IC, operating an L298N dual full bridge driver. Logic power is supplied by a LM317HVT adjustable linear regulator. Current adjustment involves a potted circuit in SIP form, marked IB46X MTI 9945; possibly custom. The main PCB is easily traced, although it has at least one buried layer.

IB462_pcb_top1.jpg IB462_pcb_top2.jpg IB462_pcb_top3.jpg IB462_pcb_bottom.jpg

Below is a schematic for the input portion of the driver, and a complete bill of materials for the board:

IB462-FL1_input_schematic.png IB462-FL1 input schematic diagram in PNG format
IB462-FL1_bom.csv IB462-FL1 bill of materials in CSV spreadsheet format

The Step signal uses the first optocoupler in part U4. This was easily replaced by clipping the leads close to the case; desoldering the loose leads; and clearing the holes of solder. This 3-step process is recommended because it places minimal mechanical stress on the plated through holes. If the plating pulls out with the part, connection to buried layers is broken and you then have a bigger problem.

The Avago HCPL-2630 was replaced with Fairchild's version, Mouser P/N 512-HCPL-2630. Of minor interest was that while both versions are rated for a typical forward voltage of 1.4V, the Avago part consistently measured at 1.37V while the Fairchild replacement measured at 1.27V.

IB462_opto_removed.jpg IB462_opto_replaced.jpg

The drivers were reassembled and returned to fully functional service.

Contact: reboots at g-cipher.net
XHTML and CSS compliant
:wq