After an absence of several months I have revisited the Cloud Sensor Project.
I have made some client software for the Cloud Sensor. Please can you try the software. Versions are below for Linux, Windows and OSX :-
LINUX: http://www.amateur-astronomy.org/Filez/CloudClientLinux.zip
WINDOWS: http://www.amateur-astronomy.org/Filez/CloudClientWindows.zip
OSX: http://www.amateur-astronomy.org/Filez/CloudClientMacOSX.zip
If it works, this should connect to the Cloud Sensor Server which is uploading data from the Cloud Sensor in the garden and you should see a graph showing the data from the two sensors.
Please let me know if it works or not for you.
Well its 12:17, and according to your software
Ground sensor is -0.06
Sky sensor is -0.37
And its saying its cloudy.
This is on Windows XP Pro, SP3.
Looks like its working fine.
Mac :D
Update. 12:33
Still running fine, auto scale on graph works ok,
ground -0.25
sky -0.43
Only problem is the dark blue is a tad hard to read.
I would say that Mike's cloud sensor is being a bit economic with the current situation saying "cloudy"
Yea well it was obviously only going to say cloudy on a day like today. I really wanted to make sure the client/server side of things worked. Thanks Mac it appears to be doing it's job.
I'll work on a version where the user can choose what colour they want the lines to be.
Mike - I tried windows vers at 2.15pm & didn't seem to do anything - have you taken it off line? John
It was offline for only about 30 seconds. Did you close it down and restart it?
If necessary, open Port 27539 on your router.
All seems fine now.....
(http://homepage.ntlworld.com/john.punnett/graphics/oas/mike1.jpg)
Great. Thanks John.
The Server may be on and off for the next few hours as I do some code updates. If your client doesn't connect then try again later.
Just out of interest how many connections can it support?
It should be unlimited. The server program simply sends out a simple text string through a specified port. That string contains the data values for 4 sensors (2 of which are not graphed. Internal Temp plus Light Sensor).
The client connects to the server IP address and port and listens for a send. When it hears one it reads in the text string. From then on it is processed into the graph.
Why do you ask?
Just curious.
Wow. At 05:07 this morning the temperature went down to -6.75 degrees C.
Yes, it was very cold.
I came to work by train instead of bike today because the roads were covered in ice. Yesterday, in spite of the snow, at least they weren't icy.
Mark
Right, last night i spotted a bug in the algorithm that was preventing it from detecting the clear/cloudy state.
I have also tidied things up a bit and made the blue a lot lighter. I am currently working on a version with a colour picker that will probably be up by tonight. The program will now keep a log text file in the same directory that logs each data reading as it comes in. This is in the format - DATE, TIME, SKY SENSOR, GROUND SENSOR, SKY STATE (Between 0 and 1 : 0=cloudy 1=clear). You can open the log file in a text editor and view all of the temperatures and sky states logged over time.
The updated version 1.1 is available here :-
WINDOWS: http://www.amateur-astronomy.org/Filez/CloudClientWindows.zip
LINUX: http://www.amateur-astronomy.org/Filez/CloudClientv1_1.linux.zip
MAC OSX: http://www.amateur-astronomy.org/Filez/CloudClientv1_1.macosx.zip
Note that the sky state, i.e. Clear or Cloudy or anything inbetween, is an average of the whole sky and not just the sky above the sensor. i.e. if the sensor is showing 50/50 clear/cloud and you look up and it is clear it means there must be cloud lower down nearer the horizon. Also remember the further away from my location you are the less accurate it will be.
Please let me know if you have any problems or spot a bug. Thanks.
Mike, you are very clever inventing a cloud sensor, perhaps you will be more reliable than all the other forcasts!
Well it doesn't forecast clouds, just tells you the current sky state. It is a way to see if it is cloudy or not without having to go outside. Will be useful for imagers who fancy a nap, etc. or who are watching TV and want to know when it clears. It will eventually have alarms, etc.
Oops!! :oops:
Sorry Windows users I posted a link to the old version :oops:
Here is the latest version 1.1 ......
http://www.amateur-astronomy.org/Filez/CloudClientv1_1.windows.zip (http://www.amateur-astronomy.org/Filez/CloudClientv1_1.windows.zip)
Yup the blue is alot easier to read, and the data logging bit is working ok,
and it seems fairly accruate, its saying 50% between cloud and clear,
Just checked and its about right.
Mac.
Might leave this running all night and see what the temperature gets down to.
Yeah that's how I knew it was -6.75 last night.
I'm pretty pleased how well it is working. Whenever the trend changes between cloudy and clear I take a look outside and so far it has been spot on. It's interesting for it to be 50/50 and you look out, see it appears to be completely cloud, or completely clear, and yet when you pop outside and have a proper look another part of the sky is the exact opposite.
Mike,
It's very "iffy" on my version of windoze XP.
I got it to run once (clicking on CloudClient.exe directly from the zip) but every time I try it now I get an error dialog "Could not find the main class. Program will exit" from the Java Virtual Machine Launcher - whether running direct from the zip or copying the files locally.
The one time it ran, it all worked perfectly.
Maybe I have a dodgy version of JVM.
Mark
Sounds like it to me. Try reinstalling Java. I downloaded it onto my laptop and it works fine. Mac has his running too. Let me know if you have any further problems.
So Mike,
Out of interest (you probably said all this before but I can't remember) does your cloud sensor literally just decide how cloudy it is by the difference between the 2 temperature measurements? Also if I knew how to could I access your computer via the s/w you have provided & could you do the same to mine....? The cloud sensor seems to update realtime so the cloud level is always changing - do you have plans to plot a trend line or something... i.e. so I can see for the last 2hrs it's been 90% clear even though at the moment it is cloudy...
John
John,
Yes the two sensors are housed differently to ensure that one reads the ambient air temperature and the otehr is shielded and housed in such a way it onyl reads the 'black body radiation' from space. This will create a temperature difference when it is clear and will have an almost identical temperature when it is cloudy (as radiation is bounced back from the clouds to the top sensor).
Access would be possible I guess but no more or less thatn usual. There are no holes opened up if that is what you mean.
Yes I intend on creating a 24 hour cloud state graph so you can see the last 24 hours clear/cloudy state. From the change in trends the user can then establish their own alarms, etc. from that.
Anyone wanting their own cloud sensor made for them can then use it to set of alarms, control observatory roofs, interrupt imaging scripts, etc.
Once I get it to work this software will be uploaded to my website so people don't need to download the software at all they can just visit the page. The software is available now as a workaround as I can't yet get the webpage script to work properly. Anyone with their own cloud sensor can obviously use the software locally.
Very impressive - Have you got plans to market it..? Seems like you have put an awful lot of work in & something that could be very useful to a remote imager.. i.e. one of those that sits indoors & controls everything... I could see this more as something people wanted to buy rather than looking at what the clouds are like in Orpington...
QuoteAccess would be possible I guess but no more or less thatn usual. There are no holes opened up if that is what you mean.
No I was just wondering how secure it and why loads of porn has suddenly appeared on my laptop.... :lol:
Very good work anyway - John
I do plan on marketing it as long as I can manufacture it in a small package, around the size of a box of Cooks Matches would be good. I would like to make a wireless version (and possibly also solar powered) too.
I think a price of around £100 is reasonable and a lot cheaper than the Boltwood unit at around £700.
Problem with Cloud sensor :(
Having left the sensor program running since 22:09 last night,
i just checked and found the temperatures at 10.06 & 6.31.
and that the display has frozen at 12:11!
The server is still ticking away. Have you restarted and if so is it the same?
Must be your end Mac. I've just fired up the client on my laptop and it is displaying the same temperature as the server is outputting.
I've just restarted it and its showing the correct time and temp,
I was just puzzled as to why it froze.
I might run a few of them together, just to see what happens
Could be a bug. I'll leave it running on my laptop and see what happens.
Right, I woke up this morning to find the server had froze. Weirdly, when you told me your client had froze yesterday the server was still ticking away as normal.
I'll do some digging and try and find out what is going on.
Just to let you know it's crashed again. :(
this time it was running from 10:51 through to 12:18,
Where its frozen.
The java sub system is showing as running at about 15-20% cpu, with 25Meg of memory allocated.
Im going to reboot and load a system debugger, and run it again, if it crashes again i'll try and get you some more information.
Just to see if its your coding or the main Java subsystem.
Have tried to get it going, Mike, but get "could not find the main class" do you know what this is?
Fay. That means you are using an old version of Java. Download the latest version - http://java.com/en/download/index.jsp - and it should be fine after that. Let me know if you have any further problems.
I'd suggest anyone else that has problems to also install the latest version as it was developed under the latest Java JVM.
I've had a quick look but it won't start at all. I suspect it's a Java version which I'll try and fix if I can...
It should run if you have the latest java JVM.
OK, the last log file gives a clue as to what may have gone wrong...
Quote2009/2/4 21:33:19 0.12 -0.18 0.0
2009/2/4 21:33:25 0.12 -0.18 0.0
2009/2/4 21:33:31 0.12 -0.18 0.0
2009/2/4 21:33:36 0.12 -0.18 0.0
2009/2/4 21:33:42 0.06 -0.18 0.0
2009/2/4 21:55:2 NaN 23.93 0.0
For some reason it sat and did nothing for 20 minutes before outputting a spurious reading from the device. There is, admittedly, no checking that the data from the device is valid, which is probably it's downfall. I will code in some data validation at the server end before it spits it back out again.
The whole thing works on 3 layers of code.
1st, the Atmel Atmega168 microcontroller is reading data from the 1 wire temperature sensors, converting the data to a form readable by humans and then outputting that data to a serial port.
2nd, code on the server simply takes that data and spits it out to the net (until I fit an ethernet module to the microcontroller).
2rd, the client software listens to an IP address until data is received, parses it and then graphs it (along with a few other things).
The microcontroller end is pretty reliable and unlikely to cause any problems and as I have not reset it inbetween resetting the server the problem is probably at the server end. I guess the client gets locked up when it receives data it cannot parse successfully.
I'll code in some error checking over the weekend as well as the colour picker.
Hi Mike,
Just thought I'd give you a 'heads up...'
Whilst at AstroFest (DrinkFest... boy did I have a sore head this morning..) Ian & myself were talking to Terry platt from Starlight Xpress. He was showing us a cloud sensor he has just developed ( should be released in a month or two) - It uses a single sensor similar I think to the IR ear thermometers you can get which I think is a single waveband radiation thermometer. It look very neat & simple to use & funnily the control box was about the size of a large box of matches. It used wireless technology to connect between sensor & control box & could give visual or audible warnings of cloud cover. According to Terry in trials it has been very accurate & works very well (but then I guess he would say that). We asked about price & he said it was likely to be £150 which seemed reasonably - it looked like a good quality product.
I've checked their webpage & their are no details on it yet but would expect them to have it up there soon...
Cheers, John
Thanks for the info. John. I thought about going down the IR route but it pushed the costs up considerably.
The Server will be on and off over the net few hours whilst I run some tests.
Fay did you update your Java and if so did it work after that?
works for me in Ubuntu 8.0.4 mate...
Thanks Ian. I was hoping someone would give the Linux version a go.
OK I have migrated the Cloud Sensor over to a new Server adn it is now up and running again. Hopefully I can now leave it in peace to run non-stop now.
I will be updated the client software over the next few weeks time permitting.
Please let me now if you try it and cannot get it to work.
You can download the Client Software here :-
WINDOWS: http://www.amateur-astronomy.org/Filez/CloudClientv1_1.windows.zip
LINUX: http://www.amateur-astronomy.org/Filez/CloudClientv1_1.linux.zip
MAC OSX: http://www.amateur-astronomy.org/Filez/CloudClientv1_1.macosx.zip
I cannot open it Mike. I see in an earlier post you asked if I had updated JAVA, perhaps I should do that?
Yep!
http://java.com/en/
Mike, did this project get any further? Think i might get me an arduino. Can you recommend a supplier???
Try this one.
http://www.EarthshineElectronics.com (http://www.earthshineelectronics.com)
;)
Is there commission for recommendations :cheesy: ;)
Mac
Quote from: Ian on Mar 07, 2011, 14:25:23
Mike, did this project get any further? Think i might get me an arduino. Can you recommend a supplier???
No but I have the parts for V2. It is on my ever growing 'to do' list. Should take that up again soon.
Mac has suggested the best supplier. If you talk to the owner I hear there are special discounts for OAS members too. ;)
Time to go shopping. I shall pm you:)
Ian,
This is a great idea - building something to sense clouds. Can you keep the project going as long as possible because it has cleared the sky wonderfully, leaving you no clouds at all to sense!
Mark