Integrated fire suppression?

I’ve been thinking about integrated fire suppression. Not for leaving the machine unattended! Not for replacing a dedicated CO2 fire extinguisher! But I have a 24oz CO2 canister sitting around unused, and see some potential benefits. I probably won’t ever get around to it, but it’s fun to think about.

  • Could be two-stage. If I monitor flare-ups, I could dump CO2 into the air assist alone to put out a fire before it spreads; could also have a second stage that floods the whole enclosure with CO2.

  • Can keep a lid on it: Don’t need to open the unit and provide more O2 before setting off a CO2 extinguisher.

  • If I can find a reasonable price on 1000 psi solenoids, could be triggered automatically; first stage could be triggered by IR camera seeing more pixels light up than expected; second stage could be triggered by temperature sensor at exhaust, which if it triggered would also turn off the exhaust fan (and close a blast gate, if configured) to stop feeding the fire.

I see that paintball CO2 regulators take 4500psi canisters down to 800psi; probably an inexpensive small orifice downstream of that would be fine for the first stage for flare-ups, and a much larger orifice through a manifold and out through a network of copper pipes to flood the box.

This seems to me obvious enough that it must have been done before, or at least talked about before, but search terms I can think of aren’t helping me much. A full-sized CO2 extinguisher intended for open fires is often 5 lbs. I’d think that in an enclosed space, 1.5lbs of CO2 would be sufficient in most cases, and might avoid needing to recharge a full-size extinguisher some day.



What about a shielding gas from welding?

That’s an interesting thought!

I think CO2 is a cheaper fill? My shielding gas tank needs a new hydrostatic test before I can get it filled, so I have no idea on current shielding gas prices around here, though. (Need to do that at some point, though; my last weld was with flux core wire and I hated it, so I’ll find out.)

I think the key problem is sensing a fire vs normal operation especially when cutting.

1 Like

That’s definitely the problem for automating it. I could have manual controls for the two stages and expect to run them while monitoring it. I could even try to automate alarms but still have manual controls.

I would expect that CV could recognize flames, but honestly an IR sensor and total IR energy might be enough to differentiate flames from laser; while the IR laser beam is powerful, it’s also directional; even a PIR sensor looking from the side at focus level should register way more energy from flame than from operation, I think. And if there is a larger fire, I’d expect a definite change in temperature measured at the exhaust port.

If the alarms prove reliable (including inducing tests I guess :grimacing:) then automating setting off the suppression of one or both stages might be possible. That’s where I’d have to find out how much solenoids would cost. Also what regulators are available and how much they would leak; I’d probably have to also have a valve that I turn on upstream of the solenoids because I wouldn’t expect them to be as tight as a high pressure valve.

When operating the second stage (manually or automatically), I would also want to pause the job and rapid the head to a front corner to try to move the lens away from flames. If the first stage is effective, I could imaging it being effective without necessarily pausing a cutting job. No experience here, just guessing.

Not planning any of this for phase one. :slight_smile:

1 Like

I have put some thought into this for a while now. I think it could be cheaply done using a small 5-pound bottle of an inert gas such as argon. A temp sensor could be set to 125 - 150°F to control a solenoid located on the bottle, and a hose from the bottle goes inside the cabinet. If you put the temperature sensor wire on the cabinet lid, you will pick up the heat.

My guess is that by the time a thermocouple temp sensor heats up enought the box will be totaled.
I still doubt that an infrared sensor will be able to differentiate normal laser operation from a fire.

Perhaps a camera with image learning algorithm that can differentiate the laser beam from a fire :)…

Most cameras have an IR filter. You have to go out of your way to get one without, generally.

But really, I think that starting with manual controls for things like “dump a bolus of CO2 through the air assist” for snuffing out flare-ups makes most sense; automating things is down the road if I’m having fun. I’m planning not to leave this running unattended!

1 Like

First step of automating would be automating an alarm (audible and visible) for the operator to react to. Only if it were very reliable in both senses (not false positives or negatives) would it make sense to invest in automated deployment based on either alarm.

Noodling this problem again …

My view is that automating the extinguishment of a fire in a laser cutter is a “diminishing returns” exercise.

Overshadowing this interesting engineering problem is the practical rule that you should never leave your machine unattended and that means you are present at the front of the machine while it is operating. This is true with or without automation. Right next to the machine should be an extinguisher.

This rule is true for any automated (CNC) piece of machinery,

Having the plumbing built in so you can just push a button to kill a flame is a nice feature but is the cost and space usage worth it vs lifting the cover and killing the fire with an extinguisher?

Maybe a FIRE ESTOP button is a alternative consideration.

  • Push-button fires extinguisher
  • Main power removed (same as Estop)
    • Fan shuts down
    • LPS shuts off
    • Motion is stopped
      The above is the same as an ESTOP but with the addition of fireing the extinguisher.

Automation considerations;
While cutting some materials a flame may be present. It’s hard to differentiate a necessary flame from one that will start of a fire. In fact, this is how a fire starts. This situation will likely create lots of false positives.

The uncontrolled kindling of the material you are cutting occurs much faster than any sensor that I have found. I tinkered (not extensive tests) with thermocouple and IR arrays.

  • Sensing fire by looking at the IR component (thermal arrays) is complicated by the laser’s frequency.
  • Sensing fire with a thermocouple takes to long. By the time it gives a reliable signal the fire is in full force.
    • Where do you put the thermocouple, on the head? A fire may start at one location but the head has since moved to a new location.



How about making the fire suppression system a proactive system and flood the cutting area with CO2 while the laser is cutting as a normal operating procedure? I think by keeping the potential to flame down it might be doable without using a ton of gas. Just thinking out loud too.

Some commercial systems use inert gas for assist, but that would be spendy. It would make some sense for cutting; it would make it sensible to use the integrated assist in the nozzle. Already bought a 105W pump for normal air assist; not starting with full-time inert gas assist. This is my intro to laser cutting; I’ve never done any laser cutting in my life. Hopefully in a few months this build will be complete and I can start…

I hear ya. I’m thinking of using the gas (argon or CO2) like a shielding gas is used from the welding world. Makes sense to integrate a solenoid valve connecting the gas to the air assist port (since it’s right there near the cut) that could be actuated on any/all laser fires. I think it would at least give the user more time to react to a potential fire situation.

Having an inert gas for air assist might reducing cutting efficiency for things like wood. You are burning the wood with the laser after all during cutting. You can actually increase cutting efficiency by having a high air flow placed close to the cut.

Does the extra oxygen help it burn, or is it just blowing away char and ash that gets in the way of the laser? I’ve been curious about this. :slight_smile:

1 Like

Could be some of both, but burning wood under reduce oxygen is likely to increase char formation and decrease cutting efficiency. :thinking:


@jkwilborn discovered that someone has already successfully implemented this idea.

Here is the description from the referenced video with links updated and annotated in case they break again:

1 Like

I ordered a 5 lb’r CO2 extinguisher, but I like the idea of being able to flood the whole enchilada with gas. If it’s running, I’m going to be there and try to put it out with a hand held if possible. I’m not going to get burnt to save myself a couple of buck… Maybe just a ‘fire’ button to activate it. I would think a short activation would be sufficient to stop a fire. Assuming there’s not enough heat for it to start again. I’ve always ‘over engineered’ things…

I am waiting on the 50 watt Omtech model that I purchased from them. Looking at cameras Lightburn has a list of them and I was wanting to put one in my machine. To follow their Lightburns instructions, they specify the distance to the bed. I have no idea how far it is from the open top to the laser bed.
I thought maybe you had a similar machine and give me a value?



The #monocle is my own, fairly unusual design. It is a monoframe design. The lid is 1.5 meters across and 870 mm deep. It has adjustable Z so the height will depend on what is being cut and what lens I’m using (I have four different focal lengths).

Also, it’s still incomplete. I still have to wire it, enclose it, attach cooling, make and install enclosure panels, and set up exhaust. And buy a fire extinguisher. :grin:


I was told by Omtech that mine was delayed in shipping (customs) and should be here by end of May. She also stated that it could show up sooner. Trying to gear up for it, along with trying to move about 100 miles north. Need to build a table for it. Bought the wheels for it.

Haven’t heard anything from my High Voltage post. If I don’t hear anything by early next week I’ll try and re-post it on redits electronics forum. I would have thought I’d get some feedback, even if it’s don’t kill yourself

Take care…
Jack :slight_smile: