7 questions on laser scanner design (DIY?)

[withlasers]

Hello fellow moths, I'm a Mechatronics student doing my internship and part of my system involves a laser vector projector, the application is actually industrial, but need not be super accurate, it's just a concept for guiding the staff around. So between the expensive precise industrial laser projectors, the expensive high power laser engravers, and the expensive proprietary laserist projectors, I'm not sure what's the best use of my money, so I need your help to answer some questions, if you don't mind:

1 do you think an off-the-shelf ILDA laser projector would be reliable enough to handle just a handful of ASCII characters (~20cm tall) and 1-3 rectangles at 4m away up to +-30deg? If so, what's the cheapest option you would recommend? I only need green (420-432nm), 1-5mw output (little, I know), no z needed. I know it's really not a lot to ask in terms of kkps, but I worry about drifting (one frame slowly drifting one way with the galvo's repetitive movement), shaking (image just not being stable, as it's very noticeable being projected on a flat surface instead of on the sky), and repeatability (if image one draws a square at top left limit, image two then also does that, they both should end up kinda in the same spot, I can tolerate +-2.5mm on this).

2 do you know any open source library that I can use to optimize the .ild file? From what I read it seems converting a vector image to .ild is the easy part, optimizing the .ild is the real hard part, which proprietary DAC software usually handles, is that right? My images would be automatically generated, so those VJ/mapping fancy programs are not useful for me, as I need it free and able to automate, so in essence I don't even need a UI tbh

3 Also any open source DAC's you would recommend? I can get a pcb done for cheap at uni, but I struggle with PCB design, so an open source diagram would be greatly appreciated. If not, a cheap one is welcome.

4 For single color, x-y movement, I only need 3 channels, right? One for X, one for Y, and another to modulate the laser? Does it get modulated by pulse width (turn off, turn on) or since it's analog it just varies the voltage?

5 About the laser, I must stay within the 1-5mw range, so I will probably source that separately regardless of what the full-build looks like, and I have a question about the modulation: can I plug a CW (continuous wave) laser and modulate it, or must I find a pulsed laser diode? If it's the first, is there a way to know how fast I can turn a specific diode on and off?

6 For anyone experienced with current galvo's market: The classics I read about online are either gone, or +400eur for used ones or more for modernized models, I also see enclosed galvo heads for powerful laser applications (co2, sla, engraving,etc ) going for 120-300 eur on aliexpress, and open galvo heads starting at 70eur that seems to replicate the classics dt-30's and 40's design, what can I expect from those in terms of performance?

7 Additionally, I should procure: an IR filter to ensure no invisible radiation, safety glasses appropriate to wavelength, f-theta lens to reduce distortion at higher angles, anything else I might need? no need for an UV filter? is the f-theta lens really necessary or compensating with software should do enough for this 'slow' application?

[withlasers]

Something like this?
DJ Laser Light with App Control, 3D Animation Laser Projector for Party, Supports Custom Drawings, Text Playback,

Although by now I'm already waist deep with a pair of chinese galvos (allegedly 30k), a custom made laser module, and an ildawaveX16 DAC.
But I'm struggling to set the scan angle, I have no clue where to do it, if it's by tuning the DAC firmware (force it to map the xy positions to a lower limit), if it's by tuning the .ild file directly (hope not), or if it's by tuning the potmeters on the galvo boards.

[mixedgas]

I'm the remaining guy on the forum who digs deep into galvo amps and interfacing.
I'm at work but since your working in an ILDA framework with what you can buy, I'll post some files.

I'll get back to you after a day of Co2 Galvo marker reverse engineering.

I suspect your fun task will be converting dxf to .ild (Cad file to ILDA)

There should be three pdf files here:

ILDA STANDARDS

Those are the old ones relevant to galvo interfacing including a few small schematics for positioning voltage to scan amp differential interface per the ILDA electrical convention.

Download a copy of LasershowGen (Free version) to talk to your existing styles of freeware hardware/play with

LaserShowGen – Bitlasers

see for the best:

ScannerMAX - High Performance galvanometer mirror scanners

Laser Galvanometers | Novanta Precision Manufacturing

If you have to go cheap, you Want Dragon Tiger DT30s with the new self protecting scanner amp.

Certified 5 mW safe green lasers come from Z-bolt, find their OEM module page. If I were you, I'd consider a unit with TTL modulation.

OEM Laser Modules – Z-BOLT® Electro-Optics

Next , Where are you located?



Steve

[buffo]

I'm a Mechatronics student doing my internship

Welcome to PhotonLexicon! Where are you located? My oldest son is currently working on an Electrical Engineering / Mechatronics degree with ECPI.

1 do you think an off-the-shelf ILDA laser projector would be reliable enough to handle just a handful of ASCII characters (~20cm tall) and 1-3 rectangles at 4m away up to +-30deg?

Yes, although you might see some flicker at that 30 degree scan angle if you have lots of characters. But just about ANY projector using standard 30Kpps scanners will be able to project ~10 letters with a box around the whole word, assuming you are using a single-line font. As you add more letters (or more boxes), the image will start to flicker more.

If so, what's the cheapest option you would recommend?

The Amazon listing that the Hermit posted above is probably your cheapest option. You could build something yourself, which would be a great learning experience, but it's going to cost you more. You won't need to worry about the image drifting or shaking; modern laser galvos (scanners) are closed-loop devices with accurate position detection and feedback. They are very repeatable, so long as they are tuned correctly and you don't over-drive them. You will need more than 1-5 milliwatts though, unless you're going to be using this in a dark environment. Well, 5 mw of green might *just* be visible, but it will be hard to see even under normal room lighting. (Unless this is just a proof of concept - in that case you can always add more power to the production unit.)

2 do you know any open source library that I can use to optimize the .ild file? From what I read it seems converting a vector image to .ild is the easy part, optimizing the .ild is the real hard part

You're solving the traveling salesman problem, which is an NP-complete problem. So, yeah, it's not trivial. Dunno if there are any libraries to help with that or not...

3 Also any open source DAC's you would recommend?

There are a few, but have you considered using a modified sound card as your DAC? There are a few open-source projects for this, and usb sound cards are really cheap. You still have to add the differential amp, and you need to bypass the DC-blocking caps, but it's a fairly straightforward mod... There are several threads here on the forum talking about this sort of project.

4 For single color, x-y movement, I only need 3 channels, right? One for X, one for Y, and another to modulate the laser? Does it get modulated by pulse width (turn off, turn on) or since it's analog it just varies the voltage?

Correct. Well, technically X and Y are sent as differential signals, so you have X+ and X-, then Y+ and Y-, but you can come out of the DAC with a single-ended signal and then feed that to a differential output amp to get the correct signal levels. And yes, your modulation is a single-ended signal (0-5volts). The ILDA standard uses the INTENSITY signal for monochrome projectors, but often you'll find that manufacturers just hook up the green color signal (if the output is green-only), or the red color signal if the output is red only. The type of modulation depends on the driver. The ILDA standard supports TTL color (0 = off, 5 Volts = on), but it also supports analog color modulation, were the voltage can be anywhere from 0-5 volts, giving you the ability to "dim" the laser. (Side note: Most DACs use 8-bit resolution for the color signals.) Finally, you can also use pulse-width modulation to control the color, assuming the laser driver can modulate fast enough. Bottom line though, analog modulation is pretty much standard these days.

5 About the laser, I must stay within the 1-5mw range, so I will probably source that separately regardless of what the full-build looks like, and I have a question about the modulation: can I plug a CW (continuous wave) laser and modulate it

Assuming we're talking about a laser diode, then yes, you can modulate it, provided you have an appropriate laser diode driver circuit. A laser diode driver is a constant current device. (See Sam's Laser FAQ for more info.) But assuming you have a good driver, you can do analog modulation, or TTL, or PWM. The laser diode doesn't care, so long as you don't exceed it's ratings. (And yes, the datasheet for the diode should include information about maximum frequency for PWM, but that's mostly to do with RELIABLY turning it on, rather than causing damage. Peak current is the real killer of laser diodes.)

Also, I'm assuming you are worried about staying under 5 mw due to regulatory concerns? If so, there could be ways around that limit, particularly if this just a demonstration build and not a completed product for sale. Is this project for the school, or for your internship company?

6 For anyone experienced with current galvo's market: The classics I read about online are either gone, or +400eur for used ones or more for modernized models

OptLaser in China sells DT-40 scanners for around $225 or so, and they're very good scanners for the price. I can't speak to the stuff you find on AliExpress.

7 Additionally, I should procure: an IR filter to ensure no invisible radiation

Unless you are using a DPSS laser, there won't be any IR anywhere in the device. Even if you did use a DPSS green laser, it should have an IR filter built-in.

Adam

[withlasers]

They are precisely tuned

: is there a guide on how to safely tune them, then? I just got everything working again, and was about to tune them as per Zenodilodon's video from last week.
I am asking but will be searching it for it too, as when displaying a circle I hear knocking and 2 sides of the circle are clipped flat, meaning its hitting a limit, so I guess I am looking for some centering/offset setting.

common with most ESPxx MCUs

: I just checked the amplitude of the wave on the scope and it's xxx, from what you wrote I understand 5v will work, but since the ILDA standard expects +-10v I will be limited to half of the angle for this galvo+amp kit anyway, right? I've seen conflicting info regarding the amplitude of the signal (dac->driver board) saying both 10v and 20v for amplitude...

But, if the ildawavex16 DAC's firmware is open source, that might be your best option. Perhaps the firmware is intentionally shipped with reduced gains to avoid hobbyists over driving their galvos. Stanley Projects should be able to advise.

: yep, as a matter of fact he just released a new firmware last week that's much better than the previous barebones one, perks of living in an AI I guess, btw he updated it as he released a new version of the ildaWaveX16, now with a raspberry Pi MCU for all the network/file reading/whatnot leaving the esp32 dedicated for buffering and timing of the output signal, it's looking really cool from what I have seen, and yea he's pretty helpful on his discord channel. You guys should check it out he made a youtube video on the the new release.

Most laser imaging apps provide setup options for kpps, RGB balance, and image optimization. Otherwise, those functions need to be coded within the MCU's firmware.

: I see, well maybe I pull a free one from the interwebs to test some images, but my usage will be all automated so I don't think I will be using imaging apps unless they got a free API.

What are you using to generate your imagery?

: Right now: nothing. I am still building the software, but I have my eyes on a library that allegedly can do dxf to ild, not sure how good it is with all the timings, cornering and blanking... if there are any free suggestions I'd love to see it

[withlasers]

I suspect your fun task will be converting dxf to .ild (Cad file to ILDA)

: ngl I'm banking on openlase or any of the forks on github

Next , Where are you located?

: Eindhoven, the Netherlands.

[withlasers]

Welcome to PhotonLexicon! Where are you located? My oldest son is currently working on an Electrical Engineering / Mechatronics degree with ECPI.

: I'm in Eindhoven, south of the Netherlands

Well, 5 mw of green might *just* be visible, but it will be hard to see even under normal room lighting. (Unless this is just a proof of concept - in that case you can always add more power to the production unit.)

: I ended up getting a 20mW 520nm, hoping that I can drive it at 5mw and still be readable, don't want to go higher because it's meant to be used in manufacturing and it would be a hassle as per IEC 60825

have you considered using a modified sound card as your DAC?

: yes but since it's an internship project I've been trying to keep things simple and documented... maybe for a personal project after this!

The type of modulation depends on the driver.

well guess who made the driver? I won't expect much from it haha I spent almost a whole month struggling to get the board to work, using an lm317 for current ceiling, a series of transistors and a 2 channel opamp for the photodiode feedback (I wanted APC because I want to drive the laser at max power permissible by law hehe and without APC based on the pd feedback it darkens as temp rises and then it might be too dim to view, since it's moving fast on the surface) but that failed and now I switched to a barebone board with an lm317 hard set at 60mA, a mosfet and some protection diodes/cap, not proud of it but at least it works, but it seems the DAC is the bottleneck right now, the "always on" is blinking I assume it's an optimization issue while cycling through the internal routines of the MCU

Also, I'm assuming you are worried about staying under 5 mw due to regulatory concerns? If so, there could be ways around that limit, particularly if this just a demonstration build and not a completed product for sale. Is this project for the school, or for your internship company?

: yes from what I recall of the IEC 60825 the limits are based on power (energy/time) in such a way that PWM does nothing to safety, in fact below the blink response time it the regulations get tighter, as we can't blink when it happens, so at the end of the day it's the power reaching the surface divided by the area of the dot that's key for the regulation. Our internships are 'professional' here, so I'm working for a company with an intern contract, and then school will grade me based on it, then 6 months specialization, then a final internship for graduation and I'm out

Unless you are using a DPSS laser, there won't be any IR anywhere in the device. Even if you did use a DPSS green laser, it should have an IR filter built-in.

: bet, I got a direct diode after reading up on dpss ones, thanks for the advice!

[withlasers]

Zenodilodon just posted this great video covering galvo troubleshooting & tuning.

excellent recommendation, although as I said in my previous comment I'm not sure if I should give it a go at the pots just yet, atm I'm trying to figure out where in the pipeline I should tune what.
Like per example, where to store and correct for the surface angle?
What exactly should be tuned at the hardware level (trimpots)?
What should be tuned at the (DAC) firmware level?
What is normally configured at the .ild file level?

[withlasers]

What does xxx mean?

: means I got carried away before I could scope the exact peak to peak amplitude of my signal, but I just did and it's 9.92v

Yes, your galvos will complain with a 'clicking' sound, while clipping and you should see it on your scope..

wait, what do you mean I should see it on the scope? shouldn't the signal DAC ---> DriverBoard remain a perfect sine wave (for a circle) regardless of what's going on downstream in the control loop between DriverBoard <===> Galvo?

For clarification, most ESP's GPIOs output 0-5v. So there would be a 2.5V offset.

: Indeed, the ildawavex16 takes care of stepping it up "The analog output stage consists of a DAC80508, which is an 8-channel, 16-bit SPI DAC with a precision internal reference, used in gain-2 mode, producing a 0-5V output, which is then processed through a 2-stage operational amplifier circuitry to generate ILDA-compliant voltages."
What I meant that makes me confused is how some people say +-10Vdc to mean a 10v differential peak-to-peak, while others say +-10Vdc to mean a 20v differential peak-to-peak, but I'm fairly certain at the moment that the ILDA standard for the signal out of the DAC towards the driver board should range from -5Vdc to +5Vdc, regardless of how it might be written.

Cloudlase is free, but don't have any experience with it and like most 'free' software it needs an ILDA DAC (EtherDream) that will set you back a few hundred $.

: thanks for the suggestion, I downloaded lasershowgen as instructed and will see if I can make it dance with the ildawave, and then check Cloudlase.

Don't you have a deadline looming on your science project?

: yes sir, got till 9th of June for the uni's report, and then 10th of July for the work contract to expire, just yesterday I finally got the galvos-laser-laserdriver-DAC all working (together) for the first time, now I need to fine-tune without breaking anything, and then implement this operation into my application.

So after having scoped the signal from DAC to Driver Board, I see the signal isn't that pretty, but it's solid and symmetric across the zero, which means it's asking it go to the right by the same amount as it's asking it to go to the left, at this point I think this is happening as a combination of 2 things:
1. the circle being full size (it's a test circle built in to the firmware)
2. the galvo being just a little off to the right and down
So I'm thinking I will change the firmware to include a scaling operation by any values 0<v<1, set it to 0.9, and enjoy a clean canvas without having to worry about the little offset at the center, because if the signal is clean, and the galvo is hitting the corner, it must mean that I should find the right potmeter and adjust it, correct? Unless I'm mistaken about the system, which very well could be as the firmware also offers a cross and a square, both are full size, and neither is getting clipped, which is strange, almost as if only the circle is "out of bounds".
Guess I will load up a different circle, see how the wave looks and how it behaves when within bounds, before downscaling in the firmware.

[withlasers]

It appears that it is indeed a hardware induced bias, because even after flipping X and Y axes (which for center-symmetric shapes shouldn't do anything signal-wise), the bias remains over limits on the right and bottom