New project: Fire sensing laser cut out system - and maybe more later.

[Graham Facer]

Graham, CO2 cylinder pressure is as I recall around 900 PSI. You can get small R size tanks which are 20 CF filled at any welding supply house. The companies like Pepsi or other local fountain distributers can also sell and fill the larger tanks.

I know, I make my own seltzer (first with a DIY setup and now with a hacked sodastream). I get them filled at the fire extinguisher place for $20. I have a 10# that I can test with though ideally if a paintball sized bottle would work they are cheap to refill. My guess is a 5-10# is needed. I think paintball ones are less.

Anyhow, checking the mail - no sensors today.

[Graham Facer]

Hi Matt,


Anyhow, until I get the sensors I don't have a lot more to add but I'd still be interested to get details on Scott's attempt. I ordered up a bunch more sensors on DX last night and a 1/4" air solinoid so now I wait for parts and maybe work on the program a bit.

Graham

Sorry Scott - I had it in non linear view of the thread - I see the details now.

[Bill George]

Graham I added to my post info on RADIANT ENERGY FIRE DETECTORS which need to be considered.
And after detecting a fire by any means you would automatically shut off the fan (and machine) and discharge the CO2 inside the laser. So all you need to do is fill the machine plus a little extra so 10 or 20 cf should be enough. Then of course you need to ventilate the room and bring in fresh air.

[Graham Facer]

Graham I added to my post info on RADIANT ENERGY FIRE DETECTORS which need to be considered.
And after detecting a fire by any means you would automatically shut off the fan (and machine) and discharge the CO2 inside the laser. So all you need to do is fill the machine plus a little extra so 10 or 20 cf should be enough. Then of course you need to ventilate the room and bring in fresh air.

Yikes - am I right in seeing those are $2k and up each? If so, I might stick to the cheap route. If there was an in-between sensor in the $100 range I could try that but I think I'll still see how far I can get on the $5-10 sensors first. But great information if I want to beef it up.

Thanks!

[Bill George]

Yikes - am I right in seeing those are $2k and up each? If so, I might stick to the cheap route. If there was an in-between sensor in the $100 range I could try that but I think I'll still see how far I can get on the $5-10 sensors first. But great information if I want to beef it up.
Thanks!

I will check with my commercial Honeywell supplier here in town tomorrow. But they are under new ownership and I don't know the owners anymore.... so discounts are more than likely gone. I am sure Johnson Controls also has more or less the same thing. But when your talking $30,000 for a new laser machine and lots more for a new building, what's a few dollars

Amazing the amount and different types of flame detectors out there available.> http://www.globalspec.com/industrial...flame_detector

[Scott Shepherd]

I can't find any reference to the sensors we were looking at. I thought they were on Mouser or DigiKey, but my searches for them hasn't turned any up and I've looked through old emails with the guys and I can't find where they emailed me any links to it. I do know they weren't cheap. The things were like $150 each and they detected fire, but not like fire on a pilot light, they visually saw the flame no matter where it was, not like a pilot light where it's just detecting a flame in a certain spot.

[Graham Facer]

OK well, The project is started - and I'll be honest I'm not as optimistic tonight. I got my first 2 sensors in on Friday and started messing with them. Well you get what you pay for!

That said, I'm not prepared to completely write them off yet but they do have me baffled a bit.

Long story short is that they are very sensitive to incandescent lights. So in my test lab/kids playroom where the lighting is pretty high (and all incandescent), they trip all the time. Not only that but they supposedly have LED's on the tiny boards that tell you when they trip and they are supposed to have adjustable potentiometers that would vary the trigger. What they really have is a potentiometer that varys when the LED trips up but does nothing to the actual analog output - which only has a value but never trips per say. As for the digital output, it basically always sees something to register as high or is defective (as it was still showing that when clasped in my hand.

On the plus side, they do register changes in light/IR and I just did test them with a compact fluorescent vs an incandescent lamp and it basically does not see the fluorescent.

I think I will test them out in the cabinet this week to see if the relatively dim light (and fluorescent at that) of the cabinet helps (though then I'd have to also program them to only look for flame when the door is closed - but easy bridge if any of the rest works - could likely tap into the existing lid sensors).

I think the board versions of the flame sensors might have been a waste of money but at least they are easy to wire up. The biggest challenge is that the field of vision for the sensors is very limited. Even with the fan shaped, 5 sensor, version it sort of needs to be dead on, on the right plane.


If anyone knows of some other flame sensors that might work, let me know.

I'll post again once I see how (badly) they work in the cabinet.

Oh and just for kicks, one sensor is high value when sensing and the other is the reverse!

Graham

[Graham Facer]

Bill would something like this be suitable do you think? SharpEye 20/20 UV/IR

[Keith Colson]

On Ebay "10 pcs NTC Thermistor Temperature Sensor" 10 of them for $2.80. Sorry I am still trying to talk you into thermistors haha. No reaction to light!

I am very interested to know if there is a reason they won't work with the right thresholding software. I have contemplated building one with an array of thermistors. The price is right for testing.

Cheers
Keith

[Graham Facer]

Keith - I might just add them to the list of things one the way to try. At that price, its worth playing with isn't it.

Graham

[Bill George]

I really don't know why a fast reacting temperature sensor set at say 165 Deg F mounted inside the exhaust air intake with a lock out circuit would not work? Yes a flame detector would work, but I think the "real" ones are pretty expensive. I really have not had time to research.

[Scott Shepherd]

Sounds like you got flame sensors. Flame sensors won't work (as you are figuring out). They are designed to tell whether or not a pilot light is lit to know whether or not to shut the gas off or allow it to come on type situations. They aren't designed to catch a flare up across the table, they are made to stick into a fixed place where flame is either present or it's not.

Measuring average temperature in the cabinet isn't going to work either. It simply doesn't work that way. I did a lot of testing a while back on air flow through the cabinet and the airflow varies throughout the cabinet a substantial amount, and the placement of your cutting in those different areas would give quite different results. If you have great air flow and a fire in the center of the cabinet, you might be getting the same reading as you'd get on the edge without a fire. It's just not that simple. If it was, we'd all have temperature sensors or $3.00 XYZ sensors on our machines doing this. I don't think you are going to get this done low tech (basic sensors). Going high tech and getting sensors that can see the flame across the table is the only option I know of and those aren't cheap.

[Dave Sheldrake]

Sounds like you got flame sensors. Flame sensors won't work (as you are figuring out). They are designed to tell whether or not a pilot light is lit to know whether or not to shut the gas off or allow it to come on type situations. They aren't designed to catch a flare up across the table, they are made to stick into a fixed place where flame is either present or it's not.

Measuring average temperature in the cabinet isn't going to work either. It simply doesn't work that way. I did a lot of testing a while back on air flow through the cabinet and the airflow varies throughout the cabinet a substantial amount, and the placement of your cutting in those different areas would give quite different results. If you have great air flow and a fire in the center of the cabinet, you might be getting the same reading as you'd get on the edge without a fire. It's just not that simple. If it was, we'd all have temperature sensors or $3.00 XYZ sensors on our machines doing this. I don't think you are going to get this done low tech (basic sensors). Going high tech and getting sensors that can see the flame across the table is the only option I know of and those aren't cheap.

Scotty's pretty much bang on, my Mitsu LPE 160F extraction system has fire suppression provided by CO2 gas, start thinking around the same sort of money as a high spec Speedy 300 and you won't be far off as a start point. It relies on FLIR technology to "see" a flame and it's core temperature then decides how long it will allow that hot spot to remain before purge.

If it could be done at reasonable cost then China would be on that train many years ago as would the big Western names.

[Bill George]

FYI we were using UV flame detectors 40 - 50 years ago, to sense flame in large industrial boilers. When the boilers had a flame out for some reason you did not want to have that 2 and 3 inch gas main dumping gas into a hot chamber. The UV detection ignored the light from the red/white hot refractory walls.
The technology today on flame detection is light (pun intended) years ahead of those days.

[Graham Facer]

Ok I've done my first in cabinet testing. I took the single sensor model and pointed it straight down the channel in the knife table and then set some acrylic alight and then started calibrating. After I while I moved on to a laser test.

Then good news is that it basically does not see the laser itself while cutting acrylic. I then worked on creating a flare up condition - luckily my table was pretty cruddy so it wasn't too hard - and dialed it in some more. I've uploaded a video to youtube here. The test is a little more sensitive than I would like for a shut down but it would be OK for a 2 stage system (alarm - then shut down). Careful - the alarm is loud if you have the sound up.

Anyhow, the downside is of course that on a 1200x900 table like I have I might need to wire up and monitor 50 of them and at a few bucks each it still adds up. But its a start and much better than I thought after last night.

Anyhow - time for paying work now.

Graham