VIPCO LED display

From Hackerspace ACKspace
Revision as of 09:37, 8 September 2015 by Xopr (talk | contribs) (added more information, linked datasheets locally and a first test-run video)
Jump to: navigation, search
Project: VIPCO LED display
Featured:
State Active
Members Xopr, Da Syntax
GitHub No GitHub project defined. Add your project here.
Description
Picture
No project picture! Fill in form Picture or Upload a jpeg here


synopsis

VIPCO LED display, euro coin as size reference, complete with self-projection

Got a hold of two VIPCO displays (120x32 red pixels).

used ICs

  • Driven by HD64180RP6 CPU (datasheet)
  • ULN2003A (High-Voltage, High-Current Darlington Transistor Arrays, datasheet), both used as horizontal and vertical driver
  • CD4514BE (CMOS 4-Bit Latch/4-16 Line Decoder,s datasheet), row select
  • HCF4094A (8-stage shift and store bus register with 3-stage outputs, datasheet), 6 per panel, 5 panels per display

pinout and protocol

While the PCB segments are designed to first shift in data for the top 16 leds (24 wide) and then the bottom 16, at least the display without serial connector is modified to use 2 bits per clock (top and bottom half).

back side (top right)
----------+
    VCC oo|data2
    VCC oo|VCC
  clock oo|D
 strobe oo|C
  data1 oo|B
inhibit oo|A
    GND oo|GND
    GND oo|GND
          |

Where:

  • VCC is roughly 7.8v without load
  • data1 is top 16 LEDs shift register
  • data2 is bottom 16 LEDs shift register
  • A+B+C+D is row select
  • inhibit is the row 'strobe'

notes

  • the main controller board uses a HD64180RP6 and is controlled by RS232 and probably RS485
  • 16 TIP107 (8A darlington) transistors drive the 32 lines per two lines (1 and 17, 2 and 18, etc..)
  • one TIP107 seems worn out/broken (number 16)
  • a couple of leds have worn out and might need replacement
  • strobe seems to be latch, inhibit seems to act as strobe
  • the transistors seem to be on a separable voltage rail (tested on 5V@1A and 7.5V@2.2A)
  • the shift register shifts from the right, but the pinout per HCF4094A is reversed per 8 columns (as seen in the video) like this:
8 9 10 11 12 13 14 15 0 7 1 2 3 4 5 6 7


plan

Since we have three LED displays which all work differently (this one, the PAL0123 and DE-DP14116), we might want to write a display base class and build on top of that.

todo

I (xopr) have cheated on the Arduino controller because you need a clock above 250kHz for the display to draw smoothly; the demo writes two bits every 128 cycles and the columns are reused every row, so you'll see bars pushed in. The noise generator shifts in between 0 and 2 columns (of 2 random bits)