4 ms·

for what it's worth, i wouldn't trust any sensor in the form of the sparkfun CO sensor (link 0 in your comment) -- there was a fellow in denmark who did a ton o

by durkie 6y ago

for what it's worth, i wouldn't trust any sensor in the form of the sparkfun CO sensor (link 0 in your comment) -- there was a fellow in denmark who did a ton of research with these sensors to create an electronic nose that exploited their huge lack of sensitivity. most of them are sensitive to a huge range of gases beyond what they are ostensibly labeled for, so he would drive them at different voltages and could basically get fingerprints of various substances: the nose could smell the difference between different types of olive oils or detect different spices. they are also very finicky to changes in temperature / humidity.

rckoepke 6y ago
Oh, I completely agree. Particularly the ones for various flammable gases. I've tested some of them in a lab and found them lacking for scientific analytic data collection, although I still found use for them as binary "presence" sensors to provide additional leak-point monitoring. I wouldn't recommend using them to meet safety requirements; I only used them to go above and beyond the minimum safety measures. I found they worked well enough to alert humans "hey this fitting might be coming loose" well before a whole-room LEL monitor on the wall noticed anything. I'd love to see that Danish fellow's research to get a sense for how terrible the non-flammable gas sensors are as well. I might have suggested a more expensive/reputable sensor but as I remarked in another reply -- it's my opinion that any solid-state gas sensor should probably be verified and calibrated against reference gas compositions. Let me know if I'm mistaken though! I just find it difficult to trust any manufacturer's datasheet claims. I also tried to state in my original post that the CO sensor shouldn't be used in a safety-critical application. But I'll make another note to make sure people read these down-thread discussions so they're not mislead.

by durkie 6y ago

rckoepke 6y ago

Oh, I completely agree. Particularly the ones for various flammable gases. I've tested some of them in a lab and found them lacking for scientific analytic data collection, although I still found use for them as binary "presence" sensors to provide additional leak-point monitoring. I wouldn't recommend using them to meet safety requirements; I only used them to go above and beyond the minimum safety measures. I found they worked well enough to alert humans "hey this fitting might be coming loose" well before a whole-room LEL monitor on the wall noticed anything. I'd love to see that Danish fellow's research to get a sense for how terrible the non-flammable gas sensors are as well. I might have suggested a more expensive/reputable sensor but as I remarked in another reply -- it's my opinion that any solid-state gas sensor should probably be verified and calibrated against reference gas compositions. Let me know if I'm mistaken though! I just find it difficult to trust any manufacturer's datasheet claims. I also tried to state in my original post that the CO sensor shouldn't be used in a safety-critical application. But I'll make another note to make sure people read these down-thread discussions so they're not mislead.