How is FLIR One limited to only -4 to +248 DegF?

[Ben321]

Is this limiting done in the FLIR One app itself (so that one could easily write their own app to get around this limit)? Or is it something in the hardware? If it is in the hardware, is it a limitation of the Lepton chip itself, or is it something programmed in the firmware (and thus possibly hackable)?

[uncle_bob]

Hi

If it's like the camera's we have, it's a limitation of the imaging chip. Noise limits you on the cold end, overload limits you on the hot end. (no that's not the whole story, but it's close enough).

Bob

[Ben321]

Hi

If it's like the camera's we have, it's a limitation of the imaging chip. Noise limits you on the cold end, overload limits you on the hot end. (no that's not the whole story, but it's close enough).

Bob

Typical thermal imager chips can go up to at least 600 DegF and down to something like -25 DegF. I was thinking they intentionally "crippled" the device, by implementing (either in the firmware, or in the official app) a value clipping/clamping function, which artificially causes it to top out at 248 DegF and bottom out at -4 DegF in the FLIR One, in order to make it intentionally not suitable for industrial/commercial use. And by doing so, FLIR forces people who need it for industrial/commercial usage to buy a more expensive product. If this is the case, then a mod (either as a firmware hack, or using the SDK to write your own non-crippled app) should be able to extend the device's apparent range to include all of the device's actual range.

[efahrenholz]

The reason a thermal range exists at all is due to calibration. The sensor might be able to detect higher temperatures, but because it can not be calibrated accurate past a certain point, FLIR clamps the range. The sensor is uncooled, which means that it heats up. The pixel size is getting smaller, and means it saturates quickly. Also, many newer FLIR sensors don't use a shutter. To overcome this, the sensor takes movement of the scene and calculates the noise to subtract out.

So ultimately, what does all this mean? Little nuances here and there tend to throw off the accuracy of the sensor beyond certain guaranteed limits. At 60C, the error could be as little as 0.2C. However, at 300C, the error could be as large as 5-10C. When a specific application requires knowing when a temperature is hit, being off by 10C could have consequences. This is a tool designed to give somewhat accurate measurements of temperature. Imagine a calculator that can't give an accurate sum after 1000. What if it reported 1000 + 1000 = 2050. That's terribly inaccurate. If you added 2050 to say 1000, the error compounds further.

I hope this helps a little. Just because the sensor is capable of reporting a higher and wildly inaccurate range is why FLIR clamps it.