EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Dannyx on December 07, 2024, 10:39:16 pm
-
Good day folks.
Tl;DR I had sort-of this brain-f@rt the other day, where I could hack into the deflection system of a CRT TV to create graphics using a laser and galvanometers.
Ok, ok, hold on :D I'm pretty sure this is impractical/impossible for several reasons, but there IS the tiniest bit of similarity going on between my idea and how a CRT operates......at least in my small simple mind :D
I'm also aware there exist purpose-built RGB lasers and I DO own a cheap Aliexpress one myself, which is what actually sparked up this whole idea, since it only comes with a bunch of pre-defined patterns and animations and I'd like to create my own, as well as, most importantly, get it to display the time in various ways, which is something I always dreamed of. More expensive lasers have this feature built-in, as well as the ability to design and load animations and patterns straight into them over ethernet, so it's not like I'm trying to reinvent the wheel and expect it to work as good as the real deal - it's more of a discussion topic to kill time, even if it's dumb :D
I spy PLENTY of issues, like I said: for one, moving a laser beam using mechanical means is going to be a lot slower than the beam normally sweeps inside the CRT. I suspect a "quick" enough galvo to even come close to "drawing" something useful doesn't exist or is impractically expensive. Reading up about CRTs I read some 15kHz is the horizontal "scan" rate, sooooo....does this mean getting a galvo that can match that SHOULD work ? Probably not....
Then there's the obvious problem of adapting the voltage levels of the TV to interface with whatever galvo I'd use.
Then I'd need to think about "blanking".....Plenty of stuff to talk about :D
-
I spy PLENTY of issues
Only one issue - it's physically impossible :-//
In CRT deflection system works by directly deflect electrons beam by external magnetic field, created by deflection coils. It works because elector has a charge, and moving charge (aka current) sensible to magnetic field.
Laser (any kind of) emit light (beam of photons) that has no charge, so CRT deflector will not affect it at all.
-
xvr, I’m pretty sure OP does not mean trying to deflect the laser beam itself magnetically. They repeatedly state “using galvanometers”. So I think they mean they just want to repurpose the deflection circuits from the TV to drive galvanometers. I mean… I guess theoretically possible, though entirely pointless since it’s probably easier to just build a compatible circuit from scratch rather than trying to interface the TV’s circuit.
The real challenge is getting galvanometers to move that fast and accurately. It’s telling that we don’t really do this. You’d probably have to make the mirror small to reduce mass, and maybe even have it in a vacuum to eliminate air resistance.
Consider that in practice, we only really use a few methods of deflecting (as opposed to filtering, like an LCD does) light to make images:
1. Galvanometers (like in laser light shows)
2. Spinning multifaceted mirrors (like in laser printers)
3. Digital micromirror devices (DMDs, like in DLP video projectors)
We have other methods used for deflecting light that don’t make images, like the servo-controlled mirrors in steerable dance club lights. But they’re much too slow to be relevant.
There were other ways of deflecting light to make images, like the Eidophor video projectors that used an electron beam to scan the surface of a rotating parabolic mirror coated in oil, creating ripples in the oil that deflected light away from the lens.
DMDs come the closest, in that they move insanely fast. But that’s because they are literally microscopic. So you need tons of them, at which point you don’t bother scanning the light source but just use each mirror as a pixel — and you’ve basically just reinvented the DLP projector.
-
Correct, that's what I meant - using the deflection system of the TV to "paint" images using a laser, since it comes the closest to doing so. Building a RGB laser show from scratch using some sort of microcontroller SHOULD be possible and WAS my initial plan, but goes way above my programming skills and I was unable to find any good resources on the topic - they're usually factory-made devices which are either at the expensive end and do what I need out of the box, or they're cheap and cheerful like the Aliexpress one, but have no customization options like the ability to load patterns and such. It does have DMX, but I don't know what that's good for - certainly not for sending actual animations and patterns - at most, it probably allows adjusting the scale of the patterns, brightness, colors, speed but that's probably it...
This is why I thought I could use a TV as the "brains" of my scanner....somehow...
-
It very well may be possible to tap off low-voltage control signals before they go to the deflection amplifiers. With that said, I don’t think it’s mechanically feasible. But that’s just a guess.
-
That was my idea too, but what's holding me back big time is the lack of an oscilloscope to see what the hell I'm doing, hence why this is stuck in the "thought" phase...
-
Is he trying to do a DIY DLP (digital light processing circa 1990's by TI)? On searching, I guess it is still being used.
-
No, a RGB laser show :D
-
I think the biggest obstacle to getting this working is, well, the CRT TV. I can see it plunging you into problems right along the line. You say that using a microcontroller is above your programming skills, but to re-create the raster scan that the TV might, if you were really lucky, do for you is really simple. It's just one loop repeatedly incrementing an analogue output from zero to a maximum value (the X axis) inside another loop repeatedly incrementing another analogue output (the Y axis). For each increment of the X and Y values (pixel), decide whether the laser should be on or off.
The difference is that you get to chose the scan rate. I don't think you have any hope of getting a laser galvo to sweep at the 15.xxx kHz sawtooth line frequency of a CRT TV in the first place, let alone without huge undershoots and overshoots. Changing the line scan frequency of a CRT TV by any significant amount is a huge pain.
Galvo Laser devices are used for vector graphics, not raster scan. Raster scan laser devices, ie. Laser printers, use a high speed motor driven multifaceted mirror for the line scan, not a galvo. I think you would be far better pulling the guts out of an old laser printer that a CRT TV.
-
Yeah, like I said - all this was just a thought, because in my mind a CRT TV seemed the closest thing to doing the job, while sparing me the headaches of trying to come up with a "software way", though I WAS fully aware from the get-go that it might not even come close to doing what I imagined.
ASSUMING it COULD work, while I do know my way around electronics way more than programming, it still doesn't mean it would be an easy job to interface a CRTs circuits with galvos, given my modest skills, so that's another nail in this coffin....
Laser printer mirrors did pop up briefly during one of my searches around the subject, but I imagine I'd still need two of them for X and Y, plus I have absolutely no idea how I'd interface with them....let alone if I'd go the extra mile to make it RGB. By that point, trying to align all this BS by hand and getting it to come remotely close even to my cheap laser show I already have would make this project impractical even further.....
-
Is he trying to do a DIY DLP (digital light processing circa 1990's by TI)? On searching, I guess it is still being used.
LOL you sound surprised, as if DLP were some obscure, fleeting, obsolescent technology. :P
On the contrary, it has all but displaced every other projection technology for most applications, with high end projectors (like cinema) being all DLP these days because DLP doesn’t degrade the way LCDs do in projection. Every DCI-certified digital cinema projector available today uses DLP.
(LCoS — liquid crystal on silicon — is now extinct, as far as I know.)
In fact, TI has been expanding DLP into other markets, like monochrome projectors used as advanced headlights, heads-up displays, and to project light onto the ground on car doors, for resin 3D printing, and apparently for near-IR spectroscopy and optical networking. (Each time I look at the DLP page every few years, they’ve added a few more applications.)
-
It's been done, i watched a video the other day, a guy used a laser and deflecting mirrors to draw patterns on a rudimentary screen made out of milky plastic plate. I do not have a link though.
-
Bitluni
https://www.youtube.com/watch?v=9qPc_I1V6go (https://www.youtube.com/watch?v=9qPc_I1V6go)
-
Consider that in practice, we only really use a few methods of deflecting (as opposed to filtering, like an LCD does) light to make images:
1. Galvanometers (like in laser light shows)
2. Spinning multifaceted mirrors (like in laser printers)
3. Digital micromirror devices (DMDs, like in DLP video projectors)
Distinctions get a little blurry with resonant mirrors:
https://eopc.com/sc-sys-xyg/
Which is basically what the OP is talking tangentially about. Keyword: raster
Building a RGB laser show from scratch using some sort of microcontroller SHOULD be possible and WAS my initial plan, but goes way above my programming skills and I was unable to find any good resources on the topic
Doing this DIY with microcontrollers is completely backwards. Computers will already output relatively arbitrary video timings with digital or analog signals depending on preference. Most of the work is the optomechanics and control theory/tuning.
-
With an angle dependent small signal bandwidth maxing out at 2.4 Khz /4 degree angle, SINE WAVE for a cheap scanner (tuned to the ILDA 30KPPS tuning), you cant even begin to follow the 15,570 Hz basic video rate.
Fast rasters are obtained using a Galvo for one axis, and a resonant scanner, polygon, or AO/EO deflection for the fast axis.
I have attached a convenient chart from the 1990s. Things have improved since then, close to the mechanical and DSP system theoretical limits but not much. Much depends on mirror size, scanner magnetics, waveform, rotor inertia, inductance, PID / PII loop tuning, I could go down the list. That chart is for a Sine wave, when you start doing discrete jumps the machine equation of state is very, very complex.
The chart is for sine response, small discrete jumps. ramps and retrace are another matter.
We do scan the odd raster in the laser show business, but 48x48 or 96 x 96 or similar and the flicker can be immense.
Using Galvos to produce video images can be / is done for biomedical and military applications, but the resonant scanner runs at F/2 or F/4 sine, and some highly customized electronics takes the video data and maps it to the mirror location. If you can find them, and if you can afford them, resonant scanners are tuned for 4, 8, or recently 12 Khz for video, and have a very small scan angle, on the order of 10 degrees.
The raster scan attached is from Pangolin Inc. marketing brochures. The flicker is something else, even with bidirectional scanning. The images is are stunning due to the coherent light.
Here is an award winner from like two decades ago, and it pushes the limits:
Enjoy Doug McCullough's Linea, and prepare to to fork over in excess of 7K for the Pro version of the software, another 2500 for a good set of galvos that can take raster day in/ day out. Sorry, 300$ sets don't last long at that level of performance, and the drive waveforms are very, very, carefully thought out. Youtube removes a lot of the inherent flicker. In the actual show flicker was used for artistic purposes. Your looking at a digital, not laser, rendering of the show, because, well, flicker.. Lots and lots of flicker. In actual laser there is always a dark line moving through the image like any other mechanical TV when filmed. Also, in Laser, the scan line ends are rounded, and drifting apart, in a noticeable manner.
The other trick is using multiple scan heads, the software for that has became available in the past four-five years, just add a 850$ output board for each additional projector used.
I'm a little miffed, having looked at quite a few Youtube videos in the past few minutes, many newer videos are from the control software video rendering output, not actually in scanned laser light.
https://www.youtube.com/watch?v=Q4a1hog3wwk (https://www.youtube.com/watch?v=Q4a1hog3wwk)
Thats about a year of artistic work, last time I talked to Doug about it.
Hooking up to a TV is likely a very, very, very bad idea. Especially when a shaft resonance is hit.
CRS prices are stunning in Q1.. Confocal microscopes are the main use. I have not seen a used set on Ebay in a long time.
chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://novantaphotonics.com/wp-content/uploads/2021/12/datasheet_resonant_scanners_CRS.pdf
Steve