Sandaysoft

Hi
I've been a bit busy with real work so not much time to experiment.
Haven't had that much sun either which doesn't help.
Had a couple of instances of bright cloud light pretending to be sun but overall not too bad.
I think I probably still need a way of boosting the dawn/dusk sun so the setting can be a bit less
sensitive.
I wonder if anybody else has tried something similar with maybe better electronics ?
Regards Alan

Just had thought , what will happen with lying snow - will it need some sort of masking skirt ?

_________________

......................Imagine, what you will KNOW tomorrow !

Hi
Hadn't thought of snow - largely because here in West Cornwall it only happens every few years,
Haven't seen more than two days at a time in 20 years. I mentioned before that I wished I'd
put the 'hat' closer to the bulb which would probably cure the cloud light so I suppose a similar
shield below would sort snow and other random reflections. Up the mast again ! When I get
time.
Regards Alan

all comes down to the difference between daylight versus dawn/dusk/snow(!)

the normal light intensity changes need to be followed by the ldr sensitivity.

when i looked back to the at the circuit in your earlier post you see that the LM324 compares the value of the ldr with the value of pot R1,which basicly tells the LM324 at what value of the ldr it must trigger.

it would be a bit of a trouble to adjust R1 every morning/evening/snow,
so why not let the R1 change with the Daylight(NOT the sunlight).

maybe if you replace R1 with an ldr2 in the same range.or put one in series with it.
ldr2 pointed the opposite or another direction of ldr1(not seeing the sun,only "normal light".

some sort of self adjusting sunshine recorder
don't know if i am right,but worth to try

_________________
http://home.xmsnet.nl/hanmir/
http://home.xmsnet.nl/hanmir/radarbox.htm

Hmm .......... I see the principle, but what would happen eg. Moonlight

Perhaps two comparator and a gate so a minimum threshould can be set, or am I just babbling

_________________

......................Imagine, what you will KNOW tomorrow !

Hi
I did try a biassing general light ldr but to get exactly the right resistance variation proved
difficult. Moonlight (we did have a full moon during my tests) didn't seem bright enough to start the timer but if we can relate the switching to the time of day I guess some clever programming could would know night from day. I still think a modern version of Mr Mcleod's recorder will work and my rather primitive effort is pretty close. Thanks for all the comments/suggestions.
Regards Alan

the LM324 operates a transistor.which operates a relais.
you can also put a threshold (small resistor) before (on the base) the resistor.
it also needs a certain amount of voltage before switching(volt=type dependent)
so if it would switch on by moonlight,thus it can be eliminated.

something like this would be the circuit

_________________
http://home.xmsnet.nl/hanmir/
http://home.xmsnet.nl/hanmir/radarbox.htm

this might be use full for finding the average year elevation of the sun
so you can position the recorder without adjusting
http://www.jgiesen.de/SunView/index.htm

_________________
http://home.xmsnet.nl/hanmir/
http://home.xmsnet.nl/hanmir/radarbox.htm

Hi
Thanks Hans - I've had a quick look and that should give a good idea of how to get the axis and other angles correct.
The resistor on the output of the OPamp would not, I think, work as the threshold voltage/ sensor voltage only sends an on or off voltage to the relay controlling transistor.
I am starting to think of LDR's which can only see the dawn/dusk sectors - mounted to view only from north east to southeast and southwest through to north west.
For most of the time when there is the sort of weather when there is definite sun or not (as when there are drifting cumulus clouds) the system works perfectly it' s only when there is bright cirrus type clouds making the overall light level high that there is a problem.
Thanks for all the ideas - I'll be trying several things over the next few days.
One or two people have expressed interest in "buying" the finished product - it's more of an experiment than that but I'd be happy to produce more detailed descriptions - maybe for a small contribution - if anyone would like them. Let me know through Steve's email.
Regards Alan

ideal for the lm324
search for a window comparator.
references could be set by two ldr

_________________
http://home.xmsnet.nl/hanmir/
http://home.xmsnet.nl/hanmir/radarbox.htm

_________________

......................Imagine, what you will KNOW tomorrow !

yes,that made me thinking.(my electronic knowledge is a bit rusty,so i need those kind of hints,and i don't want to make it too complicated,all based on the original circuit)
problem with the ldr is the overlapping values of clouded during daytime and sun in evening.(lux values)
maybe for the evening a photodiode is more suitable since it is more sensitive for evening light.

_________________
http://home.xmsnet.nl/hanmir/
http://home.xmsnet.nl/hanmir/radarbox.htm

Hi
I haven't had time to do the latest modifications but the recorder has been ticking away for August and so far we seem to have had 137 hours of bright sunshine here in West Cornwall. That's an average of around 4.5 per day. I haven't found any definitive values to compare with but 200 would be a good guess and the impression is that we have had significantly less sun this month than normal. My feeling is that I'm recording about 10% less than the real value (still the dawn/dusk problem) so maybe I'm not too far away. The timer stops and starts with every small cloud which I am sure Campbell Stokes burn marks can't distinguish so maybe that system is over-reading. If there is anyone with a CS recorder or other and they have time to build my device it would make an interesting comparison.
Regards
Alan

Hi apenwith.

I have read the complete discussion for the sunshine recorder with great interest and have a few comments / proposals:

As the whole project one way or the other has to end up in some computer interface and software, I did some thinking:

It is fairly easy to implement a simple A/D converter using the LPT1 port on a PC with a fairly low resolution.
This could be used to measure the actual resistance in the LDR for further processing.

It would then be easier to experiment with sunshine decission thresholds etc.

Further the software could easely calculate actual sunrise and sunset times for the actual location for any date of the year.

Problem is still as I understand it the lower sun radiation at dawn and dusk - would another LDR measuring the ambient light toward north (assuming nothern hemisphere) maybe help?

Most of the time the sun radiation is dull at dawn and dusk but some times it is almost evenly bright untill under the horizon dependent on the humidity/polution etc.

I made an experiment with a halogen lamp, an old fashion buld and a scrap LDR - and it is easy to observe the change in resistance when the reflected spot is present and not.

Do you have an ideal geometry for distance between buld,and LDR?

Best regards

Ole

Hi
I've been running my system for some time and the results have been quite reasonable compared with the average sunshine results for my area (maybe the accuracy of the switching for quite small clouds compensates for the slight loss of dawn/dusk). I tried a north facing LDR as a compensator but with my very simple detection circuit it was difficult to get any sort of balance against the main LDR. In the end three LDR's seemed best - one central in the tube and one each east and west and wired in series. The geometry of the bulb and tube is that the central LDR can just see the edges of the bulb. As the sun is now getting lower the bulb itself largely compensates for the angle as the reflected spot is lower down the bulb and hence closer to the LDR's. I have to adjust the sensitivity slightly but only maybe every two or three weeks. I used an old fashioned game port to interface as the timing software I found uses a joystick switch as the on/off switch. I'm sure an LPT could be used but my programming skill for detecting that are non-existant. Glad someone else has tried the theory and the subject is not dead.