Awesome piece of kit and the engineering involved. As I expected, the optoelectronics separates the printer from being a huge pile of cogs and boards into a sophisticated machine. Really interesting stuff.
A separate look on the optical units would be cool, but I (personally) wouldn't try to operate the laser section. Not with the lid off at least or with the right laser goggles. I read around on laser printers and it's probably 5mW - 10mW IR (780-800nm), so that kinda sucks because it's invisible and more than enough to burn stuff.
I can imagine what SeanB said, the beam has to be extremely collimated. 300dots/inch makes it 0.1mm per 'pixel'.
I believe laser diodes usually aren't that well collimated compared to their gas 'alternatives'. They are lasers in that they multiply light inside, (that's the term laser) , but their viewing angle are at worst 10 degrees. Well that's bare diode, maybe this laser diode already has a tiny lens inside it, so the viewing angle is much smaller.
Other than that it are almost like LED's that are extremely sensitive to ESD and other nasty things (reverse polarity).
I don't get what Optical filtering you would need, as lasers have a very small light spectrum (like <2nm wide compared to several tenths of a color LED). I only think the lens is so to focus the light very carefully on the rotating mirror.
As this is an 'extreme' teardown anyway, disassembling and plowing through all boards would be interesting to see what's involved.
Especially the optics.. Lasers usually contain TEC's for temperature stabilization, so it would be interesting to see whether it just got a big heatsink or a TEC (with H-bridge?) as well. Ofcourse, it's only a tiny diode, I may be ehm, blinded , having researched controllers mainly for 3W lasers in my last internship (and for 3W output you do need a TEC).
The scanner could well use a
linear CCD sensor indeed, the sensor looked very thin compared to it's width. I know they are also used in spectrometers with grating mirrors; the grating material reflects the light where the angle is proportional to the wavelength, so you basically get a rainbow inside. Measure it with a 1D CCD sensor, and whoila, you can scan the optical spectrum.
I see
Toshiba makes sensor with resolutions up to 10k (wow!), so you can scan up to 2400dpi. Probably not in this model, but really quite cool technology you don't get to see everyday
![Smiley :)](https://www.eevblog.com/forum/Smileys/default/xsmiley.gif.pagespeed.ic.R8GFI-pF6f.png)
Didn't even know such high resolutions existed, the spectrometer I used had 2048 pixels and was branded as 'top class'. Ah well, probably something to do with spectral sensitivity and noise, because the device used a 16-bits converter to read the data.