Cloud Sensor
After looking around on the Internet I found this and this which I found very intriguing. The Boltwood Cloud Sensors look great but seem very expensive for what they are (out of my budget anyway!) so I decided to have a go myself.
I used a Peltier thermoelectric cooler, purchased for a few £ on ebay, sandwiched between two small pieces of aluminium sheet. I used a small amount of thermal grease to make sure heat transfer would be as good as possible between the sheets and the cooler, which here works in reverse, i.e. the temp difference between the sky and the ground changes the potential difference generated by the Peltier device. The whole lot was held together with nylon bolts, mounted to the outside of the observatory and connected to a USB multimeter, purchased from Maplin for £20!
I used the provided software over several nights to calibrate the sensor - it was amazing to see the differences in the voltage generated by it changing as the cloud conditions changed.
I soon found though that rain did affect it - rain drops had the effect of cooling the top plate, but this usually changed the voltage much more than a normal clear night did, so is quite easy to spot. I did have a go with covering the sensor with cling film, but it didn't work for me unlike as mentioned in the second link above.
I soon found though that rain did affect it - rain drops had the effect of cooling the top plate, but this usually changed the voltage much more than a normal clear night did, so is quite easy to spot. I did have a go with covering the sensor with cling film, but it didn't work for me unlike as mentioned in the second link above.
It was a fairly simple matter to talk to the multimeter in my program using serial comms.
Rain sensor
So my program now had a good idea about the cloud condtions, and could react to them, but I felt like I needed a rain sensor as a sort of failsafe.
I tried a KEMO rain sensor (purchased from Maplins) but I kept having problems with it giving false alarms. I have read others having issues with these so I decided to make my own, with a view to possibly buying something else later.
I purchased a small piece of gold stripboard from Maplin and used that to make a simple rain sensor. Each alternate strip is soldered together, with a power resistor mounted on the underside to help combat dew and dry off the sensor after the rain has stopped. The sensor was linked up to a simple transistor circuit controlling a relay, which I wired in to be read as an input by the K8055.
This sensor works well, though I do have to wipe the stripboard with wire wool about once a month to get rid of oxidisation. For this reason I will eventually upgrade to another commercial unit.
Sky Brightness Sensor
In addition to the Peltier-derived cloud sensor and the rain sensor, I wanted the observatory to be able to check that the sky was actually dark.
I used an old LDR that I've had since er.. school (! - that's >20 years!), made a mount for it using bolts and bit of waste water pipe. I also put a small piece of thin perspex over the top to protect it from the worst of the weather.
Rain sensor
So my program now had a good idea about the cloud condtions, and could react to them, but I felt like I needed a rain sensor as a sort of failsafe.
I tried a KEMO rain sensor (purchased from Maplins) but I kept having problems with it giving false alarms. I have read others having issues with these so I decided to make my own, with a view to possibly buying something else later.
I purchased a small piece of gold stripboard from Maplin and used that to make a simple rain sensor. Each alternate strip is soldered together, with a power resistor mounted on the underside to help combat dew and dry off the sensor after the rain has stopped. The sensor was linked up to a simple transistor circuit controlling a relay, which I wired in to be read as an input by the K8055.
This sensor works well, though I do have to wipe the stripboard with wire wool about once a month to get rid of oxidisation. For this reason I will eventually upgrade to another commercial unit.
Sky Brightness Sensor
In addition to the Peltier-derived cloud sensor and the rain sensor, I wanted the observatory to be able to check that the sky was actually dark.
I used an old LDR that I've had since er.. school (! - that's >20 years!), made a mount for it using bolts and bit of waste water pipe. I also put a small piece of thin perspex over the top to protect it from the worst of the weather.
This was then mounted to the same pole as my electronic weather station.
I have found that the LDR is great for not just detecting day/night, but also clouds! As I live in a light-polluted area, the sky's light level at night changes depending on whether it's a clear dark night as opposed to cloudy. So I now factor this in to the decision-making as well as the Peltier cloud sensor.
Weather Station
I use a Maplin (that shop again!) USB "Professional" weather station along with Cumulus software, which re-generates a text file every minute which my program uses to check if there's been any rain in the last hour (which I class as "bad", in case of showers and to allow the cloud sensor to dry out a bit) and for high winds. I have to check the file is both present and current to ensure the data is up to date, so no old data is used.
Still lots to do - mostly programming, working on the logic for coordinating imaging while it's clear and dark. But I have a few updates pending still on this blog. Keep an eye on my Twitter account @cho_web too for updates on this project.
I have found that the LDR is great for not just detecting day/night, but also clouds! As I live in a light-polluted area, the sky's light level at night changes depending on whether it's a clear dark night as opposed to cloudy. So I now factor this in to the decision-making as well as the Peltier cloud sensor.
Weather Station
I use a Maplin (that shop again!) USB "Professional" weather station along with Cumulus software, which re-generates a text file every minute which my program uses to check if there's been any rain in the last hour (which I class as "bad", in case of showers and to allow the cloud sensor to dry out a bit) and for high winds. I have to check the file is both present and current to ensure the data is up to date, so no old data is used.
Still lots to do - mostly programming, working on the logic for coordinating imaging while it's clear and dark. But I have a few updates pending still on this blog. Keep an eye on my Twitter account @cho_web too for updates on this project.
