Night-Time Cloud Sensor - 1-Wire Project - Revisited

[TokKiwi]

I have had one working for over a year now and gives good relative sky conditions at nighttime like an Airport Metar would, if you knew how to read one...

The trouble was the quality of the Polyethylene kitchenware plastic container meant that there was no UV protection and after a few months showed signs of cracking up.  Inside it, I had a black painted paint tin lid on top side and white underneath, pushed to the top of the inside of the container and this had a 1-Wire temp sensor glued on the top black surface and wires fed back to a solar board to the side connectors and then listed in WD as an extra temp sensor then that Extra Temp sensor No was claimed as the cloud sensor under Setup, Miscellaneous settings.

Well this Tupperware like container cracked up and water got in, - which cancelled my Hobby-Board network - I pushed another plastic container over the top to temporarily maintain the weather protection.  Then we had a storm recently and the extra container blew off, most of the cracked body of the original container died further and went off into the distance with the winds, some polystyrene packing disappeared and then the paint tin lid was found on the back lawn and all I could see from the ground looking up was the lid screwed down still and the wires with the temp sensor partly attached - one leg was broken off the sensor when the paint tin lid left the scene rapidly so it is definitely not working.

I had a recent thought, and am now basing my latest project on this idea, using a car reversing lamp or front facing white light - caravans here have them as do long trailers etc forward facing.  Inside is a metal base plate and then a rubber base washer to seal against bodywork.  It looks ideal and I will follow through with some photos and design sketch of how to complete this simple project.  I will link back to a previous effort I posted on here as well.  I miss my accurate sky conditions at night

Graeme

previous images are here:

[TokKiwi]

Updating with a picture of the failed unit, and the proposed new equipment ready made for the job. 

You can see the damaged temp sensor and the broken off leg/lead

Graeme

[Windy Greg]

Graeme

I used a lab beaker bought off tradme, cut a thin aluminium disk the same dia, painted black and temp sensor affixed. Then cut a plug of polstyrene and glued the ali to the top. Inserted same into the beaker with the wire inset into the poly and glued a disc of galv steel to the bottom of the poly.
The bottom disc has a 6mm stud in the centre as a mounting device and then siliconed the bottom disk to the beaker. Been going now for nearly three years and no problems.
The lab beaker is made from a glass which has more clarity than standard glass and will withstand extreme heat. I have had readings of over 70 C during the summer. This works for both night and day sensor.

Cheers 

[Windy Greg]

Graeme

Further to last see
http://www.weather-watch.com/smf/index.php/topic,44315.msg376461.html#msg376461
For photos of my sensor.

Cheers

[TokKiwi]

I was getting a temp difference of up to -4.6ºC below ambient and up to 75ºC odd in the heat of summer.  I will try this new method as it is convenient to use.  Will paint the base plate tonight I hope, and get it ready for a weekend install.

[skyewright]

For the last 6 months or so I've been using a piece of 100mm drain pipe about 125mm deep. Inside are 2 "discs" of 50mm thick polystyrene. In the top disc I made a circular recess (rebate?) about 5mm deep. In the bottom of that rebate is a black painted alu foil disk with an OS "temp only" sensor (also blackened) just poking through the centre (its wire is fed through holes through the polystyrene). Over the top is a piece of clear plastic from a CD case siliconed in place (I was going to use glass, but it the bit of CD case fit so well, and it not shown any sign of degredation yet[1]). To help keep things dry I popped a couple of sachets of silica gel between the 2 layers of polystyrene, and then I siliconed between the edge of the bottom layer of the polystyrene and the pipe wall, and around the hole the sensor wire goes through. Around the outside of the pipe I wrapped alu foil tape.

This arrangement gives a clear night difference of -3.7ºC.

I don't really use it for daytime solar as I have a VP2 solar, good job too as the OS sensor can't handle temps over 80ºC and on a sunny day the insulated pipe sensor regularly exceeds that! When the senor is in range the speed of response is very good, with the graph giving a very similar shape to the VP2 solar sensor graph (except for the plateau caused by the 80ºC ceiling).

[1] We rarely see UV greater 6.

[Weather Display]

I might get one of these working again too....using a temperature sensor via my labjack...

[TokKiwi]

I made it Saturday afternoon, painted and glued and soldered, then placed in same position as the old one.  The result was not as good as the previous one put down to the different IR transmission properties of ABS vs Polyethylene - the soft kitchenware product I used last time.  

Perhaps the lamp bezel is polypropylene, I don't know but it was only a -2.7ºC differential max overnight with a clear sky.  There may yet be a "zero offset" difference at ambient which I have yet to define, but it did do what it was supposed to.

Check the comparison of the green (wx stn ambient) and the white lines - nighttime cloud sensor.  I just got it done right on sunset with son helping...

Graeme

NB you can see the red temp rise in the camera box when the heating circuit comes on to protect from condensation, inside and outside the roof mounted box in our frosts.

[niko]

I doubt that lens molding is polypropylene, more likely polystyrene, arylic, or polycarbonate IMHO. Is it thicker?

[TokKiwi]

Yes more likely acrylic or polycarbonate, agreed.  Overall, I guess it is thicker   and has "ring" profiles inside to help distribute the light so if lower performance for IR transmission and thicker, then there will be less temperature differential, ie less sensitive as a device.  But the main criteria this time were a) Long Life  and b) Easier to install - more purpose built and less work...   And hopefully more folks can follow this simpler but more effective long term design.

WD does a great job of doing the Math and setting the weather condition ranges based on the temp differences, so that is what makes it work, despite a lower range in performance.

Project photos to come later today. Lunchtime now...

Graeme

[TokKiwi]

Well I have posted my project on Flikr where it may be easier to view with descriptions for each photo, but I will link the photos into here so they are embedded in the text.  

Note - this whole project just took a Sunday afternoon.  

All photos are in a Set here on Flikr website also there are larger views of each with more descriptive text as well.  And some more photos not shared here...

http://www.flickr.com/photos/tokweather/sets/72157627194727261/

The old nighttime cloud sensor died of UV damage on the soft PE Kitchenware plastic, and then a storm blew it to pieces, with parts distributed onto the ground, and the temp sensor destroyed with a solder leg broken off when the metal plate the sensor was glued to, blew away.  So I needed something that was UV resistant and designed to be out in the weather.  Hence the selection of an auto spares NARVA reversing lamp.



Link = http://www.flickr.com/photos/tokweather/5992068933/in/photostream


THEORY:
I took it apart and sanded the galvanised steel base plate - the sensor uses "black body radiation" theory with the -273ºC sky sucking heat radiated from this black body.  Ie just as a laser beam transmits energy from one "source" to another body through space, the sky, like a freezer, allows heat to be transmitted through space and to cool down the object.  If there are clouds in the path, then transmission is reduced and the ambient air and other objects will heat the plate up again as it absorbs energy from its surroundings.  Note - try putting your wet hand into a freezer for a few seconds - does it feel colder??

Stripped and sanded own and ready for painting:



Link = http://www.flickr.com/photos/tokweather/5992626616/



Here I have spread the center leg - the signal one - away from the outside pair which are joined together for this operation.  Then I have added a small pair cable for the 400mm length I needed to reach the sensor board I was connecting too.  see three posts above in the earlier photos.  I have the heat shrink sleeving on there just loosely showing the pairing of the legs.

DS1820S Temperature Sensor - Hobby-Boards 1-Wire networks device, it has its own ROM ID and is polled by the weather station PC and brought into WD for data manipulation and storage.



Link = http://www.flickr.com/photos/tokweather/5992069091/


Now we see the heat shrinking in place and some elephant glue setting with a weight...



Link: = http://www.flickr.com/photos/tokweather/5992626718/in/photostream/


Up close you can see the fitted sensor and wiring etc...



Link: =  http://www.flickr.com/photos/tokweather/5992626872/in/photostream/


We now see it almost together ready for the clear bezel to be screwed into place and then fitted up on the roof...   excuse the slightly out of focus image...



Link = http://www.flickr.com/photos/tokweather/5992108859/


Now the finished installation - excuse the poor quality cabling - I should use some cable ties up there...  NB there is 12vdc going inthe box for Hobby-Boards 1-Wire Network, 7.5vdc for the ex PS fan, 15VAC for the three heater resistors - demister circuit, the small usb cable for the camera, CAT5 for 1-Wire network, nighttime cloud sensor...



Link: = http://www.flickr.com/photos/tokweather/5992626758/in/photostream/


Inside the Webcam box is the Solar Sensor board connected to the modified and relocated sensor in the clear ABS bubble on top of the box.  But it shows how you can connect an external sensor through the tiny block connectors center left - the overdrawn grey wires...



Link: = http://www.flickr.com/photos/tokweather/5992825010/in/photostream/


You can now see the webcam box and the anemometer on the small 2m mast:



Link = http://www.flickr.com/photos/tokweather/5992627010/in/photostream/


As you can see, if you have some Hobby Boards stuff already, it is an easy add-on.  Otherwise, this might encourage others to use HB equipment for solar sensor and a nighttime cloud sensor.  These are shown in the graph below:



http://www.flickr.com/photos/tokweather/5992203243/in/photostream/


Good fun adding to your weather station using Weather-Display!!!

Graeme

[Weather Display]

I like the idea of using a tail light like that , well done 

[niko]

How is the "black body" isolated from ambient temperature?

[TokKiwi]

How is the "black body" isolated from ambient temperature?

Because the plastic is a reasonable insulator with low conductivity, and there is no air to air connection - the plate inside is not connected to ambient air via conduction or convection.  I noted that during daytime with heavy cloud the temperatures did come close together within 0.1ºC so the IR transmission from the local clouds did bring it down to ambient and that the "ambient" values may have no differential or offset which is good.

[niko]

Oh, I didn't realize there was a plate, it looks like the temp sensor is just glued to the metal base.