Reports until 15:04, Tuesday 01 December 2015
H1 TCS
nutsinee.kijbunchoo@LIGO.ORG - posted 15:04, Tuesday 01 December 2015 - last comment - 05:07, Wednesday 02 December 2015(23865)
IR sensor replacement work - NO REPLACEMENT DONE

John, Richard, Nutsinee

Today's plan was to go out to TCSX table, take out the old IR sensor, and replace it with the new one (+VP inspection). John wanted to make sure that the sensor actually does what it was designed to do before we unwrap the light pipe and put it in so we did some tests. The comparator box was thought to be calibrated to trip at ~45 degree C. However, I forgot to write down how I did it so the number was very shady... We waved a small soldering iron in front of it, starting at low temperature (~150 deg F). The box would trip when the soldering iron was waved at half a centimeter away from the sensor. Keep in mind that the sensor is mounted at ~2.5 inches away from the center of the viewport with a small angle (the sensor is looking straight down while the optical window is about 0.6 cm deep under the viewport cover. So, the sensor must be able to trip when seeing a heat source at 2.5" away with ~22 deg viewing angle. We did a quick test by waving the soldering iron in front of the sensor with several temperature settings at several distances (At this point the potentiometer was about a quarter turn away from the room temperature tripping point. It's a 10-turn pot!). 800 deg F setting tripped the box at ~3" away, 600 deg F and 500 deg F tripped the box when waved at ~2" away. But these temperatures are impractical settings. At 800 deg F, 3" away we were able to make the box trip when placed the soldering iron rounghly 1cm above the sensor's viewing axis. This gives as a viewing angle of ~8deg, about a factor of 3 less than what's required. We did the same test with the sensor that was mounted on the TCSY viewport. The results were similar (400F, 500F soldering iron didn't trip the sensor when places at the center of the viewport but tripped the sensor when moved to the edge of the viewport. 600F at the center of the viewport did trip the sensor.)

 

To confirm the temperature of the soldering iron we used to do this test, we used Richard's FLIR camera and a thermocouple. Turns out the actual temperature of a certain spot on the soldering iron we pointed the IR sensor to was lower than the display. The thermocouple measured 304 F when display reads 350F, 630 F when reads 700F, and 750F when reads 850F.

 

Conclusion: The sensor unit isn't doing what's it supposed to do and we should rethink about installing them without an improvement to the comparator box circuit. Note that we were only using half a turn out of 10 turns available on the pot to set up this test. There's no way to fine tune this thing to where we want it to trip!

Images attached to this report
Comments related to this report
alastair.heptonstall@LIGO.ORG - 19:54, Tuesday 01 December 2015 (23885)TCS
Link

Thanks for testing these.  It's maybe not so clear how they are meant to operate.  The IR sensor is seeing heat from a solid angle of roughly 60 degrees.  Hence the IR sensor is measuring something different depending on how far away you put a point heat source like a soldering iron.  If it is looking just at the room, then the room temperature acts differently because it is a constant temperature across the full 60deg view of the sensor.  It looks similar no matter how far away it is from a room temperature object (emissivity not withstanding) because the integrated IR power going through the sensor window is the same.
 
The viewport acts very different.  We choose ZnSe because it is transparent at wavelengths in the infrared.  As a result it also doesn't radiate at those wavelengths when it gets hot, so the emissivity is effectively zero.  This means that our IR sensor does not directly measure the temperature of the viewport.  The IR sensor might see something hot stuck on the viewport depending on size, distance and temperature.
 
However the IR sensors are incredibly sensitive to scattered CO2 laser light since this is 10um radiation.  Also, in this case the distance from the viewport isn't such a big effect because the light is probably only slightly diverging (depending on whether scatter is specular or not).   We have done a series of tests at Caltech and if you put something in front of a ZnSe window while the IR sensor is looking at it you get a massive signal from scatter, completely saturating the detector.  The sensor we use is listed as having an operating range up to 105C.
 
It's hard to calibrate the IR sensor potentiometer for this sort of signal.  We tested a prototype IR comparator at Caltech and were able to see that by setting the IR potentiometer to just away from room temp (but low enough that the temperature of my hand could set it off) it easily tripped with a small amount of scatter from the ZnSe.
 
So in other words, I don't believe that it makes sense to try to calibrate the IR sensor using a point heat source since this is not the effect used to protect the viewport.  The other IR sensors have always been calibrated such that they trip "slightly" away from room temperature, which I take to be roughly body temperature (which is around 37C).  Bearing in mind that enviromental temperatures could easily get close to 37C, it seemed unwise to set it any closer to room temperature.  This still gives a lot of headroom compared to the signal expected from the sensor if it experiences scattered laser light.
michael.zucker@LIGO.ORG - 05:07, Wednesday 02 December 2015 (23892)
Link 

That's very helpful, thank you. Can you please link here the design documents describing this system's design and operation? This should describe the proper way to test and adjust it.