Air Assist Boost - Possibly Daft Idea 🤔

I am looking at it to improve cut quality.

One of the main benefits I get from other lasers such as the Trotec and thunder is their air assist power means I get cleaner cuts without masking, plus I think deeper cuts would be possible on the Glowforge but I would need to test this to actually see if there is a measurable improvement (I just know when the air assist is clogged I get worse cuts now)

1 Like

are you cutting on 40w trotec and thunders? higher power usually means cleaner cuts at the same thickness. i get a cleaner cut on the universal at work on 1/8", but it’s a 75w laser, so i would expect it to because it’s going significantly faster through the wood.

and yes, i would expect worse cuts when the air assist isn’t working as well. that’s because it isn’t clearing the smoke/debris out of the way fast enough for the laser to cut clearly through.

but there’s a bell curve of how much that’s going to improve with more focused airflow. i don’t think that a clean air assist fan on the GF isn’t giving you a clear cut.

not trying to discourage you from trying it, i’m curious to see if it would have any effect. i’m just trying to temper expectations. i don’t think it will be as significant as you seem to.


I know it made a huge difference when I upgraded my K40, with substandard air assist it couldn’t cut 1/8" plywood in one pass, keeping all the same optics and just upgrading the air assist allowed it to cut basically anything the Glowforge could. In 99% of cases an air compressor worked better than a fan …

One of the interesting experiments in that community was using a bent metal tube instead of the lens nozzle to get even better air (better focus rather than directing straight down) which might be another thing to try.

1 Like

Go ahead and try it.

It would be cool if a small 3D printed part that snapped onto the existing duct could focus the air onto the cutting area and produced cleaner cuts without masking.

I thought something like that had been shared here before but can’t find it.


Because I was not aware of the bad effects of magnets, I had a lot of problems with the air assist. When the exhaust fan died, I got a Vivosun 190 Cfm external exhaust which is not sufficient when cutting smokey stuff. When the replacement machine was set up I was no longer using magnets and the Vivosun was backed up by the regular exhaust. The first thing I noticed was that the smoke caused a lot less crud because it did not hang around long enough to do so, and that included the air assist.

What I have done is to smooth the path of the air to make it work best. If it is trying to blow past a hold-down pin or even the edge of the material it can cause turbulence that can interfere with the efficiency of the fan’s wind.

With the magnet problem, I found myself analyzing everything else, not expecting the real cause. The real direct cause I found was when the smoke ignited, the laser beam lost a lot of power that even large amounts of smoke did not cause it. I realized the fan was sucking up goop even to the point of jamming it from working at all.

What I did to clean the fan was to fix a vacuum hose to the outflow of the exhaust (with the machine off of course) but the moving air would spin it up, and I would pump hand sanitizer through that would dissolve everything, and running the vacuum a while longer would dry it completely with nothing staying in the machine.

Now two machines later the fan is very clean, even without any special cleaning. The higher volume caused by the external exhaust has been a primary cause of that in my estimation.

1 Like

Previously… Laminar flow nozzles yield cleaner cuts:


that’s really interesting. it has the same structure as the universal (and, from @jbmanning5’s comments, i’m guessing the trotec, too). i’m totally willing to be wrong here. i haven’t seen evidence that the cuts they’re doing in our model shop are particularly dirty compared to the ones we’re doing in our office with the cone. but they’re also running 150w compared to our 75w, so that may be part of it, too, with the speed increases.

one of the limitations that you would probably have doing that on the GF is not being able to direct the flow along the same path of the laser like nervous did. you can’t push it straight down, there isn’t space. and i’m guessing part of the reason GF didn’t is the need to make that head easy to remove for cleaning.


So the laminar flow would be an additional design aspect beyond the cone both the Universal and Trotec are using. Blowing through the nose cone (like those two) is an axial air assist - just because it’s on the same axis. Nervous designed a nozzle that smoothes the flow of air to fit their Trotec. I never picked up a 3D printer to make one, but I was curious about how well it would work.

On the Trotec, I have a built-in air compressor as well as hookups for an external compressor. I generally run the external air compressor because it does keep things a little cleaner. But, I can switch between which one I use in the software - since things sometimes need different air settings. Rather than adjusting the pressure, I keep the external set at one value and the internal set at another and just switch the source.

The Universal air system is fantastic because of how it creates an air curtain around all of the optics. Incredible how clean it keeps them! I’d love to see a beefed up air assist on a Glowforge, but don’t know the best way to go about it.


IMO there’s plenty of space for a cleveryly-crafted “nozzle” to direct a jet of air directly onto the cut area using the existing fan and duct, and I have a 3D printer (two actually) - I just don’t have the skills to turn my ideas into a physical object.

I thought a lot about it since I first saw the concept (before the link above, years earlier) but gave up given I’ve never really had any problem with the current arrangement, other than surface “smoking” when not masked.


i get that. but would it generate the same benefits blowing from a 60 degree (random SWAG) angle vs an axial path straight down through a nose cone? i really don’t know.

we have a lot of limitations on this machine that were built in by the shipping packaging dimensions that were meant to keep it box shippable by UPS/FEDEX. this could very easily be one of them, since it’s at least partially tied to the height limitation which doesn’t leave space to do axial air assist. plus the whole head being on the horizontal gantry.

i don’t disagree with this. what i’m unsure of is whether blowing at that angle will get you the same results as blowing straight down would. it’s a relatively extreme angle. and without modifying the head unit, i think you’re unlikely to be able to change the angle much. and any change you do make to that angle will probably lower the height you can cut to be less than 1/2" on the crumb tray (and 2" overall).

again, i’m not saying it can’t be done or that it has no purpose, just pointing out limitations you would need to work around.

1 Like

There’s about 1/4" to work with, and a “nozzle” that fits snug against the head then slopes down, even if it didn’t extend down below the current “vent”, might be effective.

It would restrict airflow but if more “focused” it might work better than the current setup.

I wish I had the skills to model it. Imagine a small funnel that perfectly fits the vent, but is twisted to point down at the laser focal point.

I have no idea if it would be effective. That’s why I wish I could knock something up to test.

It’s tough to say. I know other systems use something similar. Actually, there are several different styles.

Epilog Zing uses what they call an air curtain, which I believe just blows high pressure air across the work surface. No idea how well it works.

Others use a hose set off to the side, but pointed at where the beam is striking. Universal has one of these, they call it a backsweep air assist / but primarily for engraving as it blows whatever comes up from the engraved material back towards the exhaust.

I think one of the things that makes the GF air assist less effective (from a cleaner cut surface at least) is you are fighting two forces — the air assist is blowing a fairly large swath of air forward, but the exhaust is also trying to suck it backwards.

With the small cabinet area and a strong exhaust, I could see just putting a rail into the front passthrough slot (or magnetic attachment to the front of crumbtray?) and blowing everything back - kind of like the Epilog set up. But, you’d have a problem with the air assist not reading correctly (if disabled, and then you’re just kind of fighting two forces again if it’s enabled).


i think you’d want to upgrade the fan if you’re restricting the flow too much. my guess would be focusing it more generally with the current fan would be more effective than funneling it, which may lower the actual flow too much. especially when you’re changing directions.

It’s already changing direction by nearly 90 degrees.

This is a volume vs pressure thing. Further constraining the air would lead to higher pressure but lower flow (volume) and that’s generally ok for air assist. It’s really about the pressure rather than the flow/volume. But the fan also has to be able to flow/push through through a constrained environment well enough to generate the pressure, so not sure how that would work.

yes, and i wish we had one of those for the GF. IIRC, the universal can be redirected in different directions. but also, the universal can be run front to back or back to front for pathing, too, right? (i seem to remember our rep explaining that, but it was 4 years ago and i haven’t tried to use it).

which lowers the flow already. adding another bend will lower it even more. just like the vent exhaust. every direction change will lower the speed.

As far as I remember, just for engraves (top to bottom or bottom to top). I always run bottom to top. But I could be wrong. I thought beyond that, it was just the direction of the path in the code.

I think it’s more a velocity thing. And also a turbulence thing. Though these are both essentially also volume and pressure.

The air comes off the fan in a turbulent vortex. As it moves away from the fan it spreads out, so both pressure and velocity drop. The turbulence accelerates this. So to ensure you’ve got enough airflow at the cut with an un-ducted fan, you need one that moves a lot more air. Most of what the fan is blowing is wasted, passing way off to either side of the cut where it does no good.

Jet engines and ducted fan motors for aircraft have stators to “straighten out” the vortex coming off the fan, and these significantly improve the efficiency of the fan.

Taking the air off the fan and smoothing it out and ducting it through a narrow orifice to the laser spot will essentially take all the work being done by the fan and direct it at the cut, none of what comes off the fan will be wasted. It’s got to be a big boost in the effectiveness of the fan.

I wouldn’t be hard to design and 3D print a plenum for the GF fan. I just lack the copious free time to do it… been too busy working on this project. :slight_smile:


Static pressure for the standard air assist fan is wretched. Airflow drops like a stone if you impinge the fan.

It’s a little weird to read this, you sort of read left to right to see how much airflow drops as static pressure increases. The key thing to notice is how steep that curve is on the lower right and upper left, your airflow won’t sustain much restriction or redirection. (I am assuming the Glowforge puts ~12v to the fan so this graph is fairly indicative of our situation)

Notice that we’re talking about very small pressures here: the maximum static pressure for this fan at 12v is about a twentieth of a PSI (1.4 inches H2O is about 0.05 PSI). It’s really hard to find a fan in the same form factor that can push more pressure, check out my vacuum table thread for a deep dive into the available fans.

Add to this mix that any fan you choose to replace here needs to be about the same overall mass, or you might throw the clean corners algorithm off. If you want to use a more massive fan you’ll need to figure out how to mount it off the head assembly and use a hose to direct things (you know, like a more serious laser does)


Absolutely!! I tried to address that in the second portion but probably didn’t make it overly clear or important enough. Enough resistance and you’ll surpass what the fan will be able to handle.

I was probably making an assumption (and we know how that goes) that one of the approaches would be to essentially duct the air from existing components/fan, down to a smaller point, rather than “broadcasting” like it is now. As far as the volume/pressure comment - pumps in general go HVLP (high volume/low pressure) or the opposite. Boyle’s Law.

That’s one of the things about the in-line exhaust fans, but also applies here since the fan type is the same (axial fan). They are generally suitable for high volume/low pressure but like your chart shows, they can stall pretty quick under pressure.

I think personally, if I were to try this, I would look at something like the air curtain set up at the front of the enclosure and block off or reduce the shroud from the existing fan (IIRC, you experimented with that, right?). I have no idea how effective it is on the little Epilog, but they continue using it, so it must not be not horrible?