LED/"LASER" Printers - How are +5000 LEDs addressed/controlled?

  • Thread starter Thread starter NTL2009
  • Start date Start date
  • Tags Tags
    Leds
Click For Summary

Discussion Overview

The discussion revolves around the technical aspects of addressing and controlling over 5000 LEDs in LED printers, particularly in the context of printing resolution and speed. Participants explore various methods and considerations for efficiently managing the LED array in a printing system.

Discussion Character

  • Technical explanation
  • Debate/contested

Main Points Raised

  • One participant proposes using shift registers to control the LEDs, suggesting that a long register could operate at approximately 11.377 MHz, but questions the practicality of this speed across a large area.
  • Another participant mentions that using multiple shift registers in parallel could reduce the required clock speed to around 711 KHz, which might be more fault tolerant and cost-effective.
  • A different participant contrasts the LED printer's requirements with those of a 4k OLED screen, implying that the LED printer may be simpler in terms of data update speed.
  • Discussion includes a comparison to laser printers, noting that they utilize a single laser diode and a rotating mirror for scanning, with some historical insights into common failure modes in early models.
  • One participant highlights that writing data to hardware in microcontrollers can be similar to writing to RAM, mentioning memory-mapped I/O as a common method in larger systems.

Areas of Agreement / Disagreement

Participants express various viewpoints on the methods for controlling the LEDs, with no consensus reached on the best approach. Different technical solutions and comparisons to other printing technologies are presented without resolution.

Contextual Notes

Participants reference specific technical details and historical context, but the discussion does not resolve the assumptions or limitations of the proposed methods for LED control.

NTL2009
Messages
637
Reaction score
393
I recently re-watched this video:



And I wondered, how are +5000 LEDs addressed/controlled? At 600 x600 DPI and an 8.5"x11" printing area, you need ~ 5120 LEDs across and 6600 scans. At 20 ppm, the scans would need to be less than 3 seconds total ( need to allow time for fusing, etc), so ~ 0.45 ms to "load" 5120 LEDs.

I'm thinking shift registers? One long register would be ~ 87 ns per LED, or ~ 11.377 MHz clock? I'm not sure that speed would be practical across that large an area or not, but it's far less than motherboard speeds. Maybe several (16?) in parallel. With 16 (a convenient digital hardware and software number), the shift registers would only need to be 320 deep, so 1.4 us per LED, or ~ 711 KHz clock? That speed may be more fault tolerant and/or cheaper to implement?

Or a different approach? It just seems like shift registers would be the simplest layout for such a small space (1:1 output with the LEDs, no matrix wiring).

-TIA
 
Engineering news on Phys.org
A LED printer would be easy compared to the significantly faster update of a 4k OLED screen.
 
Reply
  • Like
Likes   Reactions: Rive, DaveE, russ_watters and 2 others
Laser printers use a single Laser diode. The horizontal scanning is done with a rotating octagonal (or maybe it hexagonal, don't remember for sure) mirror.

The vertical (length-wise) scanning was just the paper movement thru the printer.

One of the common failure modes in some early, major brand, printers was the IC (chip) that drove the scanner motor. I don't remember the details but the electrical design was rather borderline. I have repaired a few by replaceing the IC in the scanner assembly and adding a few resistors.

The other, less common failure mode, was the bearings wearing out, requiring a motor replacement; but it was more practical to replace the whole scanner assembly.

The service manual for an early (2004) Laser printer is available here:
https://www.printertec.com.br/arquivos_download/1160.pdf
Lots of technical info there!

Have Fun!

Cheers,
Tom
 
There are a lot of potential methods for writing data to HW in a uC.
In general, I don't really see that this is any different than writing to RAM as viewed from the uP buss. Memory mapped I/O is one common method in larger uC systems. It doesn't take long for a special purpose uP or FPGA to fill up 5K of RAM.
 

Similar threads

How Do You Program a Sequential LED Light Pattern in MIPS Assembly?
  • · Replies 3 ·
Replies
3
Views
5K
Multi-Cycle Processor Implementation in Verilog?
  • · Replies 1 ·
Replies
1
Views
13K
Satellite internet constellations
  • · Replies 6 ·
Replies
6
Views
4K
Requesting suggestions for languages, libraries, and architectures for parallel (and sometimes non parallel) numerical and scientific computations
  • · Replies 8 ·
Replies
8
Views
4K
Can Evolved State: Dichotomy Capture Your Imagination?
  • · Replies 1 ·
Replies
1
Views
4K
News Patriot Act 2: Summarized by Alex Jones
  • · Replies 10 ·
Replies
10
Views
4K
News Bush Honest & Trustworthy Until How Many Lies?
  • · Replies 65 ·
3
Replies
65
Views
11K