Aerospace and Mechanical Engineering
Adjust Font Size: Normal Large X-Large

Instrumentation Laboratory 2008 Rebuild

The original set of circuit boards, which buffer all the i/o to and from the microcontroller, were built by hand by Jim Frame in 1998-9. This was an extremely tedious process. Due to the fact that the BL1600 microcontroller is no longer available and the solderless contacts are wearing out, the lab stations needed to be rebuilt. The basic organization, concept and capabilities of the lab stations were kept the same. The processor was upgraded to the Rabbit 2000 and because they are now available on a module, the RCM 2000, there was no need to purchase a complete controller. The PC's were also upgraded, but the rest of the lab station (oscilloscopes, DVMs, function generator and power supply) are unchanged. The circuit is directly derived from the original one designed by Jim Frame.

New Bread Boards

The image below shows the new bread board assembly. It is made up of three printed circuit boards mounted on a plexiglass support. Clicking on this image gives a larger labelled version.

Bread Board overview
Click for a larger, Labeled image.

Features

  • Fully buffered digital inputs (12)
  • Fully buffered and output current limited digital outputs (12) (outputs also have pull downs to force outputs when in tristate conditions on RCM 2000)
  • Buffered serial ports: Port B on RCM 2000 and COM2 on PC.
  • 8 buffered LED indicators (that accept +/- 12 V inputs)
  • 5, +/- 12 and variable power supply connections
  • Two isolated BNC cable connections
  • Direct connection headers for hard wired experiments
  • Jumper selectable TTL levels on PC COM2.
  • Jumper selectable run program from flash (stand-alone mode, no PC needed)
  • Double sided PCB's with minimum exposed traces on top surfaces.
  • 15 bypass capacitors

A fuller description of the Bread Board is also available.

Design Documents

The following are copyrighted by T. W. Shield and released under the GPL.

Document PDF Native Format
Parts List single_station_parts_list.pdf single_station_parts_list.xls
Main Board Schematic all-main-board-2008-11-18.pdf all-2008-11-18.sch
LED Board Schematic all-led_board-2008-11-18.pdf (in all-2008-11-18.sch)
Power Board Schematic all-power_board-2008-11-18.pdf (in all-2008-11-18.sch)
Main Board PCB main_board-2008-11-18.pdf main-board-ss-pad-2008-11-18.pcb
LED Board PCB led_board-2008-11-18.pdf led-board-ss-pads-2008-11-18.pcb
Power Board PCB power_board-2008-11-6.pdf power-board-ss-pads-2008-11-6.pcb

The SCH and PCB files were generated with free Express PCB Cad software. The schematic is designed to be linked to the printed circuit files.

Vendors

The parts list. lists the needed parts that were available from the following vendors (as of late 2008).

Changes and Usage Notes

Assembly problems

The major problem in the assembly process was the close tolerances on the holes for the metal tabs on the solderless connectors. If these holes are tight the metal tabs bend and do not go through the holes. If a new set of boards were ordered, all these holes should be enlarged. The holes for the plastic tabs are also tight (particularly on the LED blocks), but can be cleaned out with a drill and because they need to be tight, should be left as they are.

Power Problems

The only major problem discovered with the above design is due to using the same power supply for the controller and the breadboard area. One of the most common student mistakes is to short out a power supply. This puts large transients on the power rails which destroy the RMC 2000 modules. The previous BL1600 boards had their own voltage regulator which avoided this problem. The solution to this problem has been to use a separate power supply for the controller. Surplus brick style power supplies (5V, +12V and -12V) where purchased for this purpose and connected to the power terminals on the main board. The IDC cable from the power board to the LED board was reduced to just the three ground wires. This has solved the problem with destroyed RCM 2000's.

A 120V power switch and an outlet housed in a pair of electrical boxes connected back to back have been added to the station. This provides a single switch that turns on and off the power to the Elenco Power supply and the brick supply used for the controller. This allows the power to the scope and meter, controlled by the lab station main power switch, to be left on.

To also aid the students in identifying shorted power supplies, two additional LED's were added to the power board on the +12 and -12 supplies with 2.2 k-ohm current limiting resistors.

Replacing Buffer chips

The 74LS244 (now 74ALS244) chips are used to buffer the digital inputs and outputs and are the most common part that needs to be replaced on the lab stations. I check them once a week and the most common problem is input bits failing. The 1k-ohm resistors on the output bits seem to be doing a good job at current limiting and protecting the outputs. I have developed a board with eight LED's and connection wires to all 24 i/o bits. This allows a program running on the controller that prompts the user to observe the LEDs to quickly test the i/o bits. The next paragraph has a suggestion for a design improvement that would allow this process to be completely automated.

These chips used to buffer the digital inputs and outputs are currently arranged so that a single chip does all 8 inputs on Port E and another all 8 outputs on Port A. A third chip does the 4 inputs on port B and 4 outputs on Port D. This latter arrangement makes testing the ports very easy. A board has been constructed that cross connects the input and output ports (A to E and B to D). Because the 8-bit output port is on a separate chip, 8 LEDs are required to check that the output bits of Port A are working. Then the 8 input bits of Port E can be checked by software. The low nibbles on ports B and D are much easier. If the software detects that the inputs don't match the outputs, then this buffer chip needs to be replaced, it does not matter if the output or input is the problem as they are all on the same chip. Thus if the chips for ports A and E where similarly split to do both output and input for each nibble, the testing could be completely automated and no LEDs would be required on the test board.

A nice upgrade to the main board would be an IDC header that a cross connect cable could be connected to. This would avoid the rather tedious use of the solderless connectors to do the cross connections.


Last Modified: 2011-07-01 at 10:53:46 -- this is in International Standard Date and Time Notation