MLX90614 & Sensiron Temp/Humidity - Outdoor Enclosure

everest

All,

I'm building a unit that I need to sense the temperature of the sky directly overhead, and provide ambient temperature and humidity. Neither the MLX90614 nor the Sensiron temperature/humidity sensor looks terribly weatherproof. . .and I need them outside all the time.

Any suggestions on creating a waterproof housing for these units? Particularly the MLX90614, which must be able to accurate sense sky temperatures. . .so covering it up isn't really a viable option.

Any suggestions would be great. Thanks!

-Jeff

Mike2545

I have also been looking for a suitable enclosure for outdoor air temperature and humidity sensor placement. I have not been able to find a commercially available unit so I was thinking of making one myself out of six- 4" saucers for potted plants.

Construction would involve drilling a 2 1/2" hole through 5 of the saucers and stacking them with 1/2 to 3/4" spacers. the top one would be left intact with only 3 holes for the screws to penetrate for the top level of spacers. These screw heads would be sealed with a high quality silicone sealant. 3 spacers per level would suffice for each level.

The bottom could be mounted to a pipe flange and secured with a section of galvanized pipe. I would only be worried about a driving sideways rain getting in but a deep enough saucer selection, ones with a higher rim, might be suitable to keep out all but the most sever storms.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Mike2545

This message sent to you on 100% recycled electrons.

ruperthart

Jeff, can't you just have the sensor under a roof and point slightly sideways? The sky should be big enough, no? Try the 10 degree field of view version rather than the 90 degree field of view.

Franklin

http://www.ambientweather.com/amwesrpatera.html·or
http://www.wxqa.com/shields.html
both work to keep the weather off but I think you need something else for the sky temp.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

everest

Thanks! Anyone have any ideas on the IR sensor? That's the really tricky one. . .

-Jeff

Tracy Allen

Jeff,

May I ask what purpose you have in mind for measuring the sky temperature?

The MLX90614 is sensitive to long wavelength far IR. I am not sure what the cutoff wavelength of the MLX is, but far IR starts at about 3 micrometers. It is not directly sensitive to visible and near IR wavelengths, although although those would heat the window of the sensor and it would "see" the resulting longer wavelengths.

There are materials such as some plastics that are transparent to long wavelength IR, such as the plastics that are used in PIR sensors that detect body heat in primarily in the 8 to 14 µm wavelength band. For atmospheric monitoring, the pyrgeometer is an instrument that monitors net radiation at long wavelength (> 4 µm). The fancy ones usually use a silcon quartz dome with a visible cutoff filter.

Microclimate effects like dew deposition are highly dependent on radiation balance. Objects drop in temperature due to exposure to a cold sky, which basically means that the net radiation is negative, outward from the earth. Some objects cool and collect dew faster than others due to higher emissivity and other thermal effects.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

everest

Hi Tracy,

I'm creating a cloud sensor. . .basically a simple routine to get the ambient temperature, and compare it against a measured sky temperature. If the difference is great than about 35 degrees C then it's clear skies. . .if it's below that, then it's progressively more and more cloudy. I also want to measure the current dew point. So I really need both my Sensioron and my IR sensor to be outside, looking up, but protected from inclement weather.

I'll experiment with some various plastics. . .I already pried the cover off my PIR and tried that, but that didn't work at all. . .it measures the temp of the plastic.

-Jeff

FearTurtles

I don't think that sensor will do as you need. I believe it sends out a IR beam which needs to bounce off an object back to the sensor. You are trying to have the beam bounce off a vapor thousands of feet away. I don't think the beam will be strong enough to achieve a reading.

everest

Hmmmmm. . .really??? That's a pretty weird way to do non-contact IR sensing. I think I can sense the temps of the undersides of clouds, I just need to have NO clouds now so I can sense some blue sky. It should be around 40 degrees C cooler than ambient. Clouds are coming in around 15 degrees cooler than ambient, exactly what I'd expect.

I tried the cover from a PIR sensor I have, and that seems to add about 3 degrees C to the readings of the unit, but it does still sort of work. The question will be whether the ambient temps will dramatically affect the differential, and how complex the math will need to be to compensate for that.

Any other suggestions would be great. I'm going to try milk jug material today since I've heard that's very transmissive of IR.

-Jeff

Timmoore

You could try these lens http://glolab.com/pirparts/pirparts.html and see if they are for the same frequency that the mlx90614 uses.

FearTurtles

Its not weird it uses a laser. but I do see we are talking about two different devices.

Have you tried measuring the sun yet? I just gave it a try and found it to be really cool right now. Only about 23degF. The clear sky was about -14deg f and the ambient air temp was around 50 deg. My non contact thermometer is a cheap one and is not a laser thermometer. I know distance causes a bigger degree of error in the reading but I would think that the sun would be hot enough to cause the device to max out at 400 degs.

Ambient temp will be a problem for you unless you are able to measure the temp layers between you and the clouds. Not sure a milk jug will work. At one time it seems that they all went to white jugs because light was bad for the milk. Caused it to loose it's vitamins. Maybe you can pick up the plastic protector from an old TV remote or VCR?

Edited

Thinking about the FIR Ir here. Most glass and plastics emitt thier own FIR Ir and quit often will reflect FIR Ir from nearby objects. So even a remote cover might not work although if you have one around it would be worth a try.

Post Edited (FearTurtles) : 5/31/2009 5:41:27 PM GMT

everest

Thanks, that's a great idea, I'll give it a try. On the sun, the reading will be dependent on the cone of the sensor. . .10 degrees for example is quite large, so when you are measuring the sun, you're also measuring a LOT of sky around the sun. If you had a very narrow cone that quite literally only saw the sun, you would indeed get very high readings. For cloud sensor, ambient at the IR sensor is all you need. . .the basics of a cloud sensor are pretty simple. . .take ambient temp, and subtract sky temp. . .then look at the resulting value. If the difference is greater than about 35 degrees you've got clear skies. Less than that and it's getting progressively more cloudy. -14F is about right for clear skies. . .I'm guessing ambient was between 50-60F? The key is a big differential.

I'm going to go tear up and old remote now [noparse]:)[/noparse]

-Jeff

everest

Another approach I'm considering is mounting the unit with a 12" PVC "shield" and then mounting it at a bit of an angle, so most rain would tend to hit the sides of the shield and run down the sides vs. striking the sensor. Other sensors seem to be fairly water-tight, they just obviously will read the water temp if they are wet. . .so it might work actually, exact in the heaviest of rains.

-Jeff

Tracy Allen

A ceilometer is a device that measures the height of clouds by means of the bounce of a laser beam.

The MLX90614 is a differential thermopile. A sensor on the bottom senses ambient temperature, while a sensor on the top is exposed to far IR. The circuit measures the temperature difference. Then it calculates the radiation flux from theory (which is outlined in the MLX90614 data sheet), along with a calibration table that is stored in its internal eeprom.

That is similar to a net radiometer, which too is usually a differential thermopile, one side exposed to radiation coming in from above and another exposed to radiation from below. A wideband pyrannometer is a device to measure incoming radiation only, using a differential thermopile connected to alternating black and white plates all facing up.

Pyranometers that measure only short wave + near IR are often based on photodiodes and are much cheaper and can be DIY. For example Forrest Mims work with the Globe science program uses LEDs as photometers to measure atmospheric water content.

A shadow band pyranometer is a device that measures the energy coming from the whole sky minus the direct beam from the sun. Sometimes thin cloud cover actually increases the total energy due to increased scattering, but thick clouds obviously decrease the energy. Cloud cover sensors sometimes use a fisheye lens to image the sky onto a sensor array.

Check out the Kipp and Zonen web site for a run-down of professional ($$$!) solar and atmospheric instruments.

I wonder if there is an easier way for you to measure the "cloud cover", without devoting the MLX90614 to the task.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Tracy Allen

Another material you might try as a protective window is clear polyethylene. Something like a 1 mil storage bag. It is relatively transparent to Far IR.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

everest

Tracy,

I'll give that a try. . .I tried saran wrap and that added several degrees onto my readings. That's something I could compensate for but it would require some math against the ambient, because the temp increase is directly tied to the temp of the covering material.

At this point I'm thinking about a simple flip top arrangement where the unit can flip open the top and take a reading every so often, but only open if the humidity is below 90%, otherwise it's irrelevant for my application because that tells me the conditions are "bad".

Thanks!

-Jeff

Tracy Allen

I think Saran wrap too is polyethylene. All have different additives, so results may vary. But I think there is no getting away from the fact that any window is going to have its own radiation that the sensor is going to see. The flip top sounds like a good idea. The sensor itself has a window, but I don't see anything offhand in the data sheet to indicate whether that window has a hermetic seal to the package, or not.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

everest

The QA forum on the manufacturers web site indicates that their other sensors are indeed sealed, but that obviously water on the window will result in a reading of only the water. They don't specifically indicate if the 10 degree instrument is sealed, but I suspect it is. I'm guessing there is just a small lens at the bottom and the little black metal cone is to restrict the view of the sensor.

I may just give it a go with a trial unit and see how it performs outside. I can even simulate some rain. I guess worst case I kill a $60 sensor. That wouldn't be ideal but I suppose not the end of the world.

-Jeff