Not long ago, [Marco] installed a powerful laser diode on a CNC machine, trying to etch the copper clad laminate and make some PCBs. The result is not very good, but the technology is very promising. In a new experiment, [Marco] bought a very cheap laser engraving machine kit from China, and now this technology looks like a winner.

[Marco] Purchasing his laser engraving machine from Banggood, which is almost exactly what you would expect from a CNC machine that costs less than $200. The frame is an aluminum extrusion, the motor is a ready-made stepper motor, the electronic device is just a Pololu-like drive, and the software is somewhere between bad and bad. Nevertheless, this machine can cut wood, leather, fabric, and can remove spray paint with a large blue laser diode.

To make his PCB, [Marco] first cleaned a copper clad laminate, painted it with spray paint, and then sandblasted it with a laser. The software of choice for this is LaserWeb, and for cheap machines, the results are very good.

After the circuit board is painted and laser sandblasted, there are some additional steps to create the PCB. This process still requires etching in ferric chloride or some other acid, but the result is good. Great, in fact, [Marco] is experimenting with copper foil and Kapton to make flexible circuit boards. You can view videos of these experiments below.

Thanks [ganzuul] for the hint.

I believe that laser cutting workshops do not like cutting copper, because their operators will have flash burns during processing. Anyway, I think it is copper.

He is not cutting copper, but cutting resist on copper.

He is not cutting copper, but cutting resist on copper. Nevertheless, once the resist is burned through, the laser beam will hit the metal copper surface and scatter.

I have been doing this for more than a year, and my result is that what you want to do is use the laser at a small angle. Otherwise, a large part of the beam may return to the diode and damage it. (This only happened when I tried to etch the circuit board with ferrous chloride paste. The paste melted and acted as a mirror, but it was still the case).

Ah, my bad. I misunderstood this process.

Can't see what's wrong with the tracing paper mask made with laser printer and Kinsten pre-sensitized PCB? Works great for me!

Compared to any material you might try laser cutting, copper has some of the most serious reflection problems, but no silver or gold. If you install a proper housing on the laser and do not try to pick up the workpiece with bare hands after cutting (it will be very hot), the operator will not be in any particularly dangerous situation. Laser optics is another matter; it is easy to blow the lens out of reflections.

I once used a 4 kW Amada Gemini to cut a 1mm piece of copper. The feed rate is about as slow as cutting 6 mm aluminum or 12 mm steel.

I do not believe. See what happens when you try to cut something and the height of the bed is the same level as you put the target on the laser focus. You mean the reflection of a piece of copper can cause skin burns a few feet away?

Maybe, just maybe it might damage the reflection of the operator's unprotected eyes. Choose a suitable pair of glasses to filter the light of the correct wavelength to prevent this from happening.

I want to know what type of laser is used...Specifications? ? ? ?

The only disadvantage of cutting copper is that it requires fine-tuning, and you have to run high power at low speed under high pressure, and the cost of cutting is very high. Therefore, if an operator is flashed, it is either because they are pushing the machine because of their stupidity, or because their boss is telling them to speed up because of their stupidity.

I also recently purchased one of the kits from Banggood, Eleks A3 2500mW, with a shipping cost of 170 Euros. The assembly is going well and the hardware is working fine, but the software provided by Eleks is very bad. SVG cannot be imported correctly for beginners. If your language/positioning setting is set to use colon (,) as the decimal separator, it will send the wrongly formatted GCODE to the machine and ignore the error it returns. It took me some time to find that. Other supported software benbox does not seem to be able to control the laser on my "Mana SE" controller. I haven't had time to watch LaserWeb yet. In addition, both the laser and the controller support PWM, but no PC software seems to support it. I haven't had a lot of time to play it yet, but I bet that I might eventually write the software myself.

Laser etching pcb is also very interesting to me because I want to get rid of messy FeCl3 etching, but I understand that cutting copper requires more energy.

I mean comma, not colon: ",". It obviously uses the locale settings to convert floating point numbers to strings, and GCODE always expects a period as the decimal separator.

The laser is almost the worst thing you can use to cut copper. It is reflective and conducts heat well. Really, it's better to use a micro mill.

This only applies to very thick copper or CO2 or very low power diode and fiber lasers. All over the world, laser cutting of copper materials with KW-class lasers has become a daily job. Cutting copper with blue wavelengths should be no problem at all, because the absorption characteristics of copper in this wavelength range are similar to steel and CO2. Those who use Blue for KW-class high-brightness laser sources will revolutionize industrial metal cutting.

These little blue engravers are a big step in bringing this wavelength into the CNC world, and what I have seen so far has left a deep impression on me.

PCB laser direct etching can be done with triple Nd:YAG laser. LPKF sells prototype manufacturing machines. But as mentioned earlier: holes, especially plated vias, are the real problem.

Isn't copper the worst thing you can cut?

As mentioned above, YAG or visible lasers in the kilowatt range will not be too troublesome for copper. If you assume it is a CO2 laser, then the worst material I have encountered is ordinary window glass, which is a quarter of an inch thick. The problem is that it breaks too easily. I can laser cut 1/8 inch thick window glass, but thicker things are very difficult.

Gold is said to be almost impossible to cut with a CO2 laser, although I have no first-hand experience in this area. There are also some materials that fall into the category of "No. It just doesn't work." For example, if you use nitrogen assist and a large enough nozzle, you may be able to cut magnesium-but if something goes wrong, I don't want to be in the same room. Beryllium copper alloy is another on this list. The smoke it releases can make heavy metal poisoning look like a common cold.

You need a larger CO2 laser with air assist to cut copper, even then it will suck because copper has very good thermal conductivity (the laser melts the metal and compressed air blows it out of the cut). Small semiconductor lasers like diodes (probably initially used in Blu-ray recorders) are completely hopeless for this kind of work.

CO2 is a bad choice for copper... Q-switch DPSS YAG is a better choice, but it is obviously more expensive and has more complex optical components.

He was just burning paint and acting as a resist. This is of course much easier than laser processing directly on copper.

This is basically how Texas A&M manufactures PCBs. It can very easily "print" to the SOIC pitch. However, there are many shortcomings in via plating. When a 330uF capacitor is charged to a 270V arc on a blow-through hole, it becomes very exciting...

Definitely look at LaserWeb! I noticed one thing about the 2500mw I purchased (https://www.aliexpress.com/item/DIY-mini-2000mw-2w-large-engrave-carve-machine-350-500-laser-engraving-carving- machine-laser printer/32611833032.html) is that they will not provide updated GRBL source (only pre-compiled and old). Found that they made their own board and swapped the output, so I just remapped it in GRBL and uploaded it. This is necessary for it to be used with LaserWeb.

So I did take a look. The standard firmware of Arduino Nano provided by Eleks seems to be grbl 0.9, but laserweb needs to be >1.1d. I verified that the step and dir pins are the same as the default configuration, and the laser PWM is on PB3 (D11), all of which seem to be consistent with the default grbl configuration. So I tried the stock 1.1f grbl firmware and the machine moved but did not report the location. So I changed $10 to 0 or 2. Also set $32=1 to enable laser mode (so the machine will not stop when changing S). It can now report the position correctly, but the laser is not lit with any S value of M3 or M4. It also enters alert mode after startup, because homing is disabled by default (no reference switch anyway). So I must issue $X after startup to clear the alarm state. I also need to find the correct step/mm setting again because the EEPROM is overwritten with an incorrect new value. I tried to compile the firmware on raspbian (to disable the alert state at startup), but the generated hex file only gave me an "error 7" (eeprom read failure) and then stopped responding. I haven't verified this toolchain yet, so I might try again using arduino IDE on windows. So I still need to make the laser work.

Ok, so these "Eleks Lasers" ($100, $101) have a speed of 80 steps/mm, and after setting the test power in Laserweb4 to> 0%, the laser is now working properly. The stock firmware grbl v1.1f runs well on the Eleks Mana SE board. This seems to be the way to go, and the only possible convenience are some tools for quickly creating lines and text. Oh, and there is a running time estimate.

I think this is what I came up with, 80 steps/mm. It's close enough.

What is the accuracy of the laser? Can it be like a 10/10 wear scar?

It depends on how much time you spend focusing on it, but usually 10/10 is realistic, if you spend a lot of time playing with it and have a good mill to start, 10/6 is feasible. For smaller marks, the etchant will get under the paint and you may get cut marks, so don't do this.

It seems that PCB can be quickly turned over.

Now I know what to ask Santa Claus;)

The problem is not accuracy, but spot size.

The very expensive ($5,000) laser in our hacker space has a spot size of about 0.02 inches, which is about 1/2 mm.

The positioning accuracy is very good, about 0.02mm, but when I count the trace width and the minimum line width (one line is etched with a 1/2mm line on each side, plus the trace width) I found that you can’t run two lines Between the pads.

I would love to use it as a technology, but it is only suitable for large components.

If you want to try, you don't need an anti-etchant-just spray some black paint on the PCB and dry it in the oven for about an hour. Etch twice-once to remove paint and once to remove splashes from the first run.

Then gently wipe with alcohol to remove any remaining splashes, etch and wipe with acetone to remove paint.

The beam from the laser may be too wide, but can a mask with a narrower diameter be mounted on the laser assembly?

In an unrelated note, first use some salt water to coat the copper surface with an oxide coating so that it does not reflect. This probably avoids the problem of beam reflections back to the laser assembly.

I sent Hackaday a video showing how to do this earlier this year, but received no response at all. What gives?

This video is at https://www.youtube.com/watch?v=t8zkrFdetjk, you can see how to use a peristaltic pump to speed up the etching time.

(My video also shows how to use ferrous chloride to accelerate etching).

There are few guesses: you did not show any actual circuit boards, or even close-ups of the etching boards, so that we can evaluate the quality. Also, maybe they can't bear that everything in your video is so messy :) No offense.

Oops, holy personal protective equipment, Batman! Why not wear gloves? Acetone, ferric chloride, and all by-products (such as copper dichloride)...if you don’t need it, why should you let yourself be exposed to these?

Maybe any machine like this should have a big red toggle switch on the laser head. Therefore, when it is closed, you can be sure that it is closed.

On the bearing/belt guide combination. There may be some accidental wear, because unless the bearing contact and the belt contact radius are the same, there must be some slip somewhere.

But it looks good value for money.

The one I sell has a tilt switch, which only fires when it is aimed down and moves relatively slowly. Security issues do not have much effect on repeatability. However, I found that a 3D printer with a Kevlar belt works well.

The biggest problem remains the same: you can't make plated vias.

Yes. We have a micromill in the university, so it is not used much. Electroplating machines are still very expensive.

People did this many years ago and managed the double-sided solder mask with a second engraving after etching. Electroplating requires drilling before etching, a simple low-resistance chemical coating (usually toxic), and a CV CC DC power supply.

But for less than $60, I can get 4 layers @10 pieces with higher accuracy. After all, most people's time is precious.

Or a simple non-isolated wire, welded on both sides, suitable for hackers with limited budgets. It puts "no" in "rapid prototyping", but hey, at least you won't wait for the PCB to arrive.

so? Just don't design plated vias.

For decades, people have achieved success by designing electronic devices without plated vias. As an amateur, must everything you do must match the cutting edge of the industry? Think of it as a design challenge. Show that you can create a good board even if there are fewer functions available than the board has.

I once made a board like this: an 8051 circuit from the 90s and 100 vias with hand-soldered wires. This is a fairly large PITA, and I broke some 0.5 mm drill bits while drilling the board. I only noticed after finishing the layout that the default through holes are very small :-( I am really happy that electronic products have made progress at the same time. But now my focus is to use modern SMD components to optimize circuits and functions, as well as technology, instead of using Outdated technology. Fortunately, there are affordable real PCBs (with plated vias, solder mask and thin wires) from places like OSHpark or DirtyPCB. I also can’t understand this "retro computing" hype or "make We make some things with old-style TTL, 6502, Z80 or digital tubes or other tubes.” I have a C64, which is very suitable for its time, but the times are changing. When I got the 486-33 PC, I was very The C64 stuff was sold soon.

totally agree. For the "home hacking" proof of concept, although this is a good choice. It can be frustrating to wait a week or so to prove that the circuit is closed. I would like to do things like this in my home laboratory

Agree to avoid vias (and double-sided) in home manufacturing. I prefer to use 0 ohm 1206/1210 resistors instead of cross rails.

I don't know of any rivets that can achieve 0.3mm through holes. If you want the durability of the circuit board and not just a bunch of intermittent connections, you have to solder them. Rivets are not compatible with modern PCB circuit/space design rules.

Has anyone tried this process on a Co2 laser cutting machine? I have been thinking about buying a cheap Chinese 40w Co2 laser cutting machine from that famous website. You can bid there, but the main thing is to click to buy it now.

I want to know if Co2 Laser can remove spray paint "resist" as effectively as this machine.

I use a 60W CO2 laser to burn off the paint. good results. The problem in this case is the copper below. Copper (as far as I know) is highly reflective in the infrared range of typical CO2 laser output, and reflections can damage/burn the lens and release dangerous chemicals in the process. Working at a slight angle may help to deal with reflections, but I'm afraid to try it yet. The machine is too expensive.

Using 40W, it will definitely evaporate the spray paint. However, you need to turn off the power because the bare copper under the paint will reflect light back to your light source. The CO2 tube may not care, but the lens may. In my experience, China’s carbon dioxide pipes are very good, as long as those of Western famous brands.

I have one of these cheap Chinese laser cutting machines. However, I have been considering installing a low-power UV laser on it and simply using it to expose photoresist. Because the driver board I installed has multiple laser outputs. So there is no reason not to try. The laser is not expensive. In my case, building a small driver board for it is just an operational amplifier and a few resistors. As a 5mW 405nm laser should be sufficient to expose the photoresist. In the worst case, I need two or three times. But it is better to wait a week or two for the prototype to be shipped from other parts of the world.

However, it is very possible that a 40W laser tube can also be used at a relatively low power, and only spray paint or the like can be used as an etching barrier.

I do not recommend this. I tried it 2 years ago and describe it here: http://www.eevblog.com/forum/beginners/practical-proscon-of-laser-vs-milling-of-pcb's/msg766269/#msg766269

The speed is too slow, the quality is poor, and some etching resistant film is still left after evaporation. I have used a 1W laser, maybe 2-3W or 40W CO2 would be better, but it will open other worm cans, either unsafe (I think the 445nm power laser is the most dangerous thing in my laboratory) or damage the ZnSe lens. Copper not only reflects laser light well, but also absorbs heat, which makes the task of paint evaporation more difficult.

I want to try the UV exposure method on a 3D printer (see http://www.diyouware.com/node/161), the laser power is much lower (so faster than CNC). It should be better, but it adds more wet processes that I don't like.

Has anyone tried to expose the photosensitive plate with a laser?

Yes, it is not the correct wavelength. But you can bring a UV lamp and focus it.

What are the benefits of using "any" paint and burning it off?

Fiber laser technology can cut brass, copper and bronze up to 10 mm thick without any problems. There is a laser that can cut it, haha.

If it were me, I would use a Banggood laser engraving machine and a small Dremel type rotary tool (instead of a laser), and then simply remove the copper mechanically. Of course, the cheap function of this unit does not really support the milling method, but I think it should be within its ability to remove the soft thin copper layer from the board. I believe someone has already done this somewhere.

If there is no suitable Z axis, it will run around. I tried:)

In fact, you need to use 3D leveling because even small changes in depth can weaken or overcut the PCB. It only needs one alligator clip to the board and the other for the drill. Let it have problems in the grid, and then adjust GRBL to a depth map. Look!

I use a 40w fiber laser to directly laser ablate the copper on the copper clad laminate. The effect is very good. I will get a picture tonight.

It’s unbelievable how people try to use lasers (buzzwords!) to do something instead of just using a simple desktop milling machine. If there is a complicated method, why should it be simple, right?

I saw people with a Swiss accent use one of the 3040 milling machines to do this. The process seems very complicated.

I have used both and found that the laser is easier and can provide finer traces and spacing.

I am surprised that no one mentions the T2 laser (t2laser.org) as an alternative to benbox or laserweb. This is a very good software, updated regularly, and developers will respond to support requests personally. The price is not expensive, about 40 dollars if I remember correctly. I have been using it on my BG laser engraving machine for about a year and I am very satisfied.

Hello, can anyone tell me what machine he uses in this process? Or customized? Thank you in advance.

Please be kind and respectful to help make the comment section great. (Comment Policy)

This website uses Akismet to reduce spam. Learn how to handle your comment data.

By using our website and services, you explicitly agree to the placement of our performance, functionality and advertising cookies. learn more