Fun with a K40 laser engraver!

Earlier this year I picked up a little desktop laser engraver and proceeded to play with it. It was interesting and fun, but it had limitations. Mostly power - the little diode laser it uses can only pew pew so hard, and because it's in the visible spectrum it doesn't like to mess with translucent or transparent things. Still, it was enough to whet my appetite. I learned the sort of thing

 A friend has a Glowforge which is a nicely integrated CO2 unit, but they're spendy. But there's a cheap Chinese engraver called a K40 that's available in a bunch of almost-identical variations. You can pick one up for under $400 shipped. A 40W invisible laser that's made to the lowest possible price point, what could go wrong?

I got a royalty check in the mail for some books I wrote nearly a decade ago, and I decided that if the publisher had sent me a laser cutter I would have been overjoyed. So I finally decided to pull the trigger. This is my story of learning this tool. I may talk like I know what I'm doing, but really I'm making this up as I go.

There are two ways I can proceed here. First option is to buy the cheapest option and learn a bunch of lessons. Second option is to buy one with the plan to upgrade it right from the start. I went for the second option because the weak points are pretty well known, and it minimizes the potential for fire/death/accidental holes/asphyxiation/etc.

First, the K40 itself. I went to Amazon because the parts ship from the US (shipping costs from China directly make Alibaba the expensive option) and because I could rely on a certain level of confidence thanks to reviews. I did choose to not go with "LATEST 2020 VERSION! NEW!" because of my planned upgrade path. 

I also picked up the following upgrade parts:

- new laser head
- new lens
- new mirrors
- 6" exhaust fan and some plumbing
- new controller board and camera
- a set of real goggles from lasersafety.com

I'll go over the reasoning for each of those as I go.

Wooo!

The first question everyone wants to know is "can you carry it in the trunk of a Miata?"

Well, not really.

Also, a 65 lb package that says FRAGILE GLASS HANDLE WITH CARE MADE IN CHINA is a little scary.

Luckily, the packaging was excellent. An outer box, a layer of foam, an inner box, more foam.

Packaging inside, well, it was good enough.

Here's what's inside.

Subbed for interest

once you install the "option" parts and like them can you put links to buy them ?

does the same base model seem to be sold be many sellers ?

Thanks

Now, you'd expect the first thing I'd do would be to plug it in and fire it up. But I want to make sure this is safe and reasonable and there's no point in not installing some of my upgrades. Besides, the pirated copy of CorelDraw that comes with it won't work on a Mac. So step 1 is the optics. Since part of the initial setup is mirror alignment, there's no point in waiting.

A new "air assist" head that allows for an 18mm lens (as opposed to 12mm) and three mirrors that are sitting face down. This should increase the power at the head nicely.

I don't feel bad about replacing this stock part. The machining of the replacement is much nicer.

How does it work? The laser tube is along the back of the case. A 45* mirror fires it down the side of the case (the X axis, basically), another mounted to the gantry ends it down the gantry (Y axis), then there's a last 45* at the head that shoots it down.

This was a pretty easy thing to do. One thing I did was measure exactly where the lens is because focus is important here, and I know the focal length is 50.8mm.

Then it was time to try the pew pew button and make sure everything is aligned. THis apparently can make people crazy, because the mirrors can be adjusted for angle in two axes but can also be moved up and down. You set the laser to a very weak power setting, put a piece of tape on the mirror in question, and tap the button. It's a matter of a fraction of a second, any longer and you vaporize the tape and then sublimate crap all over your mirror. Good lord this thing has some juice.

I used a lot of tape. It wasn't all that difficult and I got it dialed in fairly quickly once I got my head around the geometry. Mirror #2 needed a slight tweak to the bracket.

Impressed with the FDA sticker on it, meat engraving anyone?

I have a friend with a glowforge, it's a neat piece of kit.

I cannot explain the FDA sticker.  And my friend with the Glowforge is interested in what I'm doing here, as it's not perfect. But it has shown me what's possible.

About that control panel - it's the New And Improved digital version with a big percentage power indicator that seems very technical. But it turns out that it's actually showing percentage of the power supply's output, not what the laser can provide. Crank that thing to 100% and you will be buying a new tube. I drilled a 2" hole in the top of the case and wired in an analog ammeter. "Redline" on the laser is 18 mA on the gauge, which is 44.0% on the digital gauge. Eeek.

The little template is one I cut (later)  to set focus. 

Links added to the first post for the particular products. Three weeks from now, they'll probably all be dead links...

BTW, k40laser.se is my go-to reference here. Loads of good info in bite-size chunks.

Smoke control. Smoke is bad because it diffuses your beam and deposts itself on the lens. My second generation (of three?) K40 came with an upgraded exhaust fan, but it's still pretty wussy. So I took a two-part strategy.

First, pull out the little 4" fan and replace it with a 6" inline fan. Also, put that fan at the end of the duct instead of the beginning because that's better in case of leaks and fans would rather pull than push. I'll poke a hole in the side of the shop later. I then took the original 4" fan, cut a hole in the case and turned it into an inlet fan. If you've been watching my Facebook Live videos, almost everything I do turns into relative pressures. So here I was helping the airflow by increasing the pressure in the case.

I also took the chunkily designed exhaust vent that covered half the bed and chopped it down. This had the effect of almost doubling my bed size to 8.5" x 9".

No pictures of any of this, sorry.

Now, I also want to control smoke right at the point of burning. The "air assist" nozzle is supposed to do that but there are some questionable aerodynamics involved here that may actually suck up smoke. Instead, I added a regulator to one of my air compressor drops and simply bent up a piece of hose to aim right at the point of contact. It's not pretty - especially with my attempt to crimp it down to make a "nozzle" - but it's solid and very effective. I can dial up the pressure and airflow as desired, and it runs off my 60 gallon tank. It's mounted with a hole and grommet through the lens mounting bracket. The thing on the bed is my first serious cut, it's a guard for the new 4" inlet fan.

The laser is actually water cooled and the K40 comes with a submersible pump. One 5 gallon Home Depot bucket with lid, 5 gallons of distilled water (not as conductive as tapwater) and a hole with grommet in the lid with an instant read thermometer. I'm not really worried about the coolant overheating, more making sure it's not too hot or cold to start with. THe happy temp is about 15-20C.

Enough for now, the ribs are done (NOT in the FDA laser...yet) so I shall continue later.

Okay, they're not ready. So, next step...

Control board. These things ship with a fairly rudimentary controller. I've been using LightBurn software for my baby cutter and quite like it. There's an open source controller called a Smoothieboard that would let me use that, and there's a commercial variant that comes with support and actual instructions and is complete. It's a big chunk of change - about half of what the K40 cost - but I like the idea of support and it gives me the potential for future upgrades like a Z axis control. Think of it as buying a Megasquirt for a $700 Miata from DIYAutotune instead of soldering up your own :)

The install was stupid easy, the new board is literally a plug-in. Attached it to my Mac and they immediately made friends and boom, I was up and running.

So naturally, it was time to test. I grabbed some cardboard, doodled out a shape in LightBurn and bzzaped it. No focus or anything, just "let's see if I can make it move".

I then calculated my focal point (44.44mm from the bottom of the head or 17mm off the bed), stacked up a few shims and took a shot at some wood.

Enormously faster than my little one, and it slices through the wood in a single 25% power pass at 25mm/s. The little laser needs three slow passes at full power and has much more trouble with scorching and never gets a clean cut. I also stuck a piece of acrylic in there and made my fan cover. The pieces fell out with a perfect edge. Love it. I could never get the little laser to cut through acrylic. This is basically the maximum size of my bed, by the way.

This was kind of a neat effect - it's a laser warning symbol that started as a raster engrave, then had a bunch of vector lines converge on it. It made a little serrated cone.

So after a bit of careful prep work, this thing is up and running far quicker than I expected. Some of that is thanks to learning on the little one, but really this thing is pretty solid. I did find that doing tiny (1/8") text in vector at a high speed (100mm/s or so, which means the head is moving at over 220 mph) led to some calibration drift, but that's asking a lot of the poor stepper motors. Engrave as raster (think dot matrix printer instead of plotter) and it has no trouble because it's missing the violent changes of direction.

I guess now we mostly work on what I can build with it. I have some projects planned!

The FDA sticker is because they regulate lasers and the importer of the lasers needs to be registered with FDA 

Incredibly cool!

Had some fun tonight and learned some things. Those two are often related.

First, my bed is bigger than I thought. Dunno why I had 9x8.5" in my head, but it's over a foot in one direction. Right after I picked up a bunch of 9x8.5 stock, too. I'm still learning just where the limits are - I actually bumped into them a couple of times tonight.

First I did some playing with 1/8" hardboard. This stuff is under $9 at Home Depot for a full 4x8 sheet and has a decent finish on one side, so it's great for experimentation. Turns out it cuts more like acrylic than thin plywood, which is probably because it's basically just sawdust in glue. These are test cuts to find the speed and power needed.

Then time to build a prototype.

Not bad. I think there's some slight variation in thickness of the material because it's almost but not quite a perfect cut. Either that or the laser power is dropping off on long cuts. Still, it'll do the job nicely. 

I also played with transluscent acrylic. You want to use cast acrylic for this, it cuts better than extruded. I picked this up locally - I can order from Tap Plastic, but it's faster and less expensive to get it here in town. Assuming they have the color you want. This is a blue. It takes a little more to cut than the black I was playing with earlier.

What are you looking at? Quilting rulers. Quilters take these patterns and layer them and do things with them and make amazing shapes. And the rulers are goofy expensive. I don't know why, but this is half of a set that costs about $60-70. So here's part of my justification for owning this thing. It's cutting another one while I take the picture.

Of course, I could cut the fabric directly...

Things I learned - my air assist hose is out of the way almost all the time, but it did catch and drag the material once. There's a solution for that, so I'll add it later. I'm also still finding the edges of my bed. Ideally, I should have some sort of fence to mark the edge of it so I can hit the origin every time. Right now, I have a piece of tape with a little pinhole marking 0,0 and then I align the edges with the gantry. But that's a bit prone to error. The bed's also quite low - I have to lift stuff up about 17mm to get it into focus. I could either change the lens to one with a longer focal point or lift the bed. I can't see myself trying to cut something 34mm thick, although I suppose I could be trying to engrave the top of something 17mm thick.

Also, I really need to put the exhaust fan outside. After cutting two copies of that hardboard control panel, the shop was getting pretty smokey.

Lasers are cool. Very informative. I dont know why I kept saying Pew Pew thru the whole thread every few sentences and pictures

Glad to see this one's been set up to run proper G code instead of Correl Draw.

Does the air assist head have any provisions for centering the nozzle over the beam? A correctly centered assist gas has a huge effect on cutting materials that melt; it may be less effective on wood or acrylic. Speaking of alignment, using a piece of flat acrylic or cardboard instead of tape combined with a cross-hair in place of the mirror makes it a lot easier. Use a hair drier to blow off the smoke.

Is there any way to enclose the beam path in a bellows? That goes a long way to keeping deposits off the mirrors.

I have amassed a good collection of things that run G code of late. Embroidery machines, vinyl cutters, laser engravers...there's a dead 3D printer sitting on the bench, maybe it'll contribute a Z axis bed! I come from the home of Corel, but I didn't realize the company was still in business.

The air assist head fires the air directly across the surface of the lens. There's concern that the venturi effect will actually make the nozzle into an inlet. I could easily give it a try. My current setup may not be the ideal but I can compensate by moving a ridiculous amount of high pressure air if necessary. 

Beam path in a bellows. That's an interesting idea. The tape, combined with a set of calipers, was a pretty easy setup for the mirrors. Just had to figure out the geometry.

Keith Tanner said:

The air assist head fires the air directly across the surface of the lens. There's concern that the venturi effect will actually make the nozzle into an inlet. I could easily give it a try. My current setup may not be the ideal but I can compensate by moving a ridiculous amount of high pressure air if necessary. 

Does this one let air escape around the lens? I was expecting it to have a sealed lens assembly and the only way out for the air to be through the nozzle.

Beam path in a bellows. That's an interesting idea. The tape, combined with a set of calipers, was a pretty easy setup for the mirrors. Just had to figure out the geometry.

The bellows is what you'd typically see on a large commercial laser cutter. The bellows would also typically be pressurized with filtered shop air or even pure nitrogen as well, but venting the bellows outside the case would be preferable to letting smoke in there.

It should fire it straight out the nozzle. I'll try it.

Spent some time playing with the thing today. It's down on power. I suspect I have some dirty mirrors. Yes, the smoke from cutting the hardboard was pretty epic, but it seems a little soon.

Cleaned the mirrors, rerouted the exhaust outside, changed to the "real" air assist head...and the laptop battery died so I couldn't do any testing. Soon.

I am really interested in this cutter Keith. Thanks for all of the information! Lightburn is the program that you run? I would be transferring files from illustrator.

Did you ever run it right out of the box? I wonder how it does in stock form?

And yes one can laser etch food. I  did this at art school to baloney (BalonaLisa)

Also did this 300sl sculpture 

Out test out of masonite then 1/4" Acrylic 36" long - My wife dropped it and broke it. 

It comes with a pirated copy of Corel Draw. So, yes, it comes with software? Won't work with my Mac and by all accounts it's pretty awful. There's an open source piece of software called K40 Whisperer that seems pretty well regarded.

Now that I've upgraded the control board, I can use my licensed copy of LightBurn with it. 

Great thanks, edited my post after I saw the LightBurn image. I use a Mac as well. 

I'm clearly down on power. I have notes for settings on various materials I've cut so far, and it's barely tickling them now. I realigned all the optics and cleaned everything up...but no love. I was careful not to run the laser too hard and while I'll admit my coolant did get to the upper end of the range - it should be 15-20C, and I reached 23C - this seems an awfully short life of maybe 60 minutes of run time. I have reached out to the seller to see if they have anything to say. At worst, I may be out the cost of a tube.

Gosh darn it. I had a new control panel for the synth all ready to cut out.

I have heard back from the seller of the K40, they asked for a video to explain what I meant about it being down on power. No response to that yet.

I ordered a new tube anyhow. They are a consumable over the long run and they have an indefinite shelf life, so if the vendor sends me a new one I'll put it on the shelf. I can't really return this particular unit as someone has modified it!

The new tube arrived almost immediately but I have held off on installing it while I wait to see if the vendor wants any sort of proof of life other than the video. 

Anyhow, I carried on with setting up the engraver to be the best option. First was some drag chain to keep the air assist hose from drooping down and catching on things.

I also installed a camera. The LightBurn software can use it to overlay an image of the bed so I can align my patterns perfectly. This is the Official LightBurn Camera installed in a freshly printed Official LightBurn Camera Mount. I did not spring for the "cable management kit" because I already have stick-on ziptie anchors and zipties.

The camera is a fairly wide angle unit, and there's a correction process you can go through (built in to LightBurn) that will turn the image into a nice flat one. It involves putting a calibrated image (that's it in the first photo) at various points in the camera's field of view so the image can be properly processed. There's a 20 minute video for an instruction and I have very little patience for YouTube video tutorials that are 5x as long as they should be. I got it corrected, but I'm not happy with the image quality. It's not as sharp as it should be. I may try moving the camera closer to the bed and (sigh) sitting through that full video to see if there's some little 5 second tip I missed somewhere 12:31 minutes in.

Once it's corrected, you calibrate it. The LightBurn software burns some targets, then you overlay crosshairs. My tired tube is still able to scorch cardboard, so at least I have that capability :) All calibrated, I did a quick test and nailed the location of the pattern perfectly even with the blurry picture. So that's good.

In order for the camera to consistently sit at the same place, I installed some little hydraulic rams to hold the lid up. Previously, it went over center and leaned back against a stop but then the camera couldn't see the bed. The arm installation proved to be remarkably difficult to get right and it still wants to foul the #2 mirror on the end of the gantry. In retrospect, a simple stick prop on each side would have been easier to implement. But hydraulic assist feels so cool!

Running the air assist through the actual air assist nozzle instead of my homebrew piece of pipe seems to work pretty well. When engraving cardboard you can clearly see it blowing the small bits of incandescent debris away.

The seller did not respond to my email with the video and nobody has watched the video, so I contacted them again via the Amazon "contact seller" interface. Again, a very quick response with...a request for a video. Sheesh. So I responded by email AND via the Amazon "contact seller" interface once again. Let's see if it works this time. If I don't hear back in a few days, I'll swap in my new tube and consider options.

I'm glad I didn't buy this thing through Alibaba. 

The seller is still playing the "we need to talk to our engineer" game now that they finally got the video - I had to make it public on YouTube so they could find it via search.. Anyone with a small amount of troubleshooting ability and understanding of the general concepts can tell it's a bum tube. It sounds like they'll send me one eventually but good lord they're dragging their feet. 

I'm no longer waiting. I stuck my new tube in and the thing fired up with the vigor it had when new. So yes, clearly the tube. Swapping it out was easy enough - I crimped some connectors into the wires so it's an easy swap later if need be. The only downside is that you have to realign the optics but that's a satisfying job.

Used it to do some fast prototyping with cardboard. It's actually a bit difficult to deal with because it cuts so easily - trying to just engrave is difficult. I did some vector engraving on this at something like 300 mm/s and 5% power, and it has trouble keeping alignment when doing a small radius circle. You can see how the text is a bit offset because of it.  I'm going to call that user error, it's asking a lot of the mechanism. 

Sure is nice to have it back, though.