Cumulus Support

Thanks Hans and Super-T
I'm not sure the hysteresis is going to work as the resistance of the LDR
only varies from about 600 ohms in sunlight to 800 ohms in cloud light.
Maybe if I 'sunglassed' the LDR the sun/cloud separation would increase
- I'm not sure.
The original circuit could cope with that but the dawn/dusk sun brings
the two even closer. I think the hysteresis circuit would need a higher cloud
resistance.
I've attached a diagram of the physical system (I have difficulty thinking of
spherical mirrors except as a lot of flat ones). Fig 1 shows midday sun
(viewed from the side) and would need a roughly 45 degree miiror to reflect
to the LDR which equates to position A on the sphere. Fig 2 is the dusk view
from above which gives the B reflection of the sun spot. So the sun is dimmer
and the reflection is 'thinner'. Compounding the problem. So fig 3 represents
the layout of two LDR's so the left hand one would see a fig 1 view (rotated
45 degrees) - giving a far better image. Thoughts ???
Regards
Alan

Don't forget there is also a spectrum shift eg. Sunset shifts towards the 'reds'.

about the hysteris,the problem is that when the clouded light is brighter than the dawn/dusk-light,it doesn't react to dawn/dusk.
setting it to those values,it would see clouds as sunshine.
so you would need this circuit three times,one reaction to daylight(as you have now) and two reacting to dawn/dusk,one ldr east,one ldr west,all outputs connected to your relais.
each with a different hysteris setting.
the lm 324 has 4 gates,so you have 1 to spare
so:
-sun comes up,ldr east reacts,ldr day and ldr west don't react.(vice versa ldr west)
-daytime,ldr day reacts,BUT ldr east and west are adjusted to lower light,so they would stay "on" the whole day,thus they have to be shielded for the sun during day period.( a piece of wood/plastic,next to the ldr,or a piece of tube over the ldr,width dependent on how long dawn/dusk must last).

for the redshift beetlejuice mentioned:
the ldr is best used for visible daylight,if it can''t see the reds good enough,you could try for dawn/dusk phototransistors instead of ldr's

just a suggestion:
how about just one ldr but with a suntracker.
http://www.josepino.com/other_projects/ ... racker.jpc

also can be made with a motor and a comparator and 2 H-bridge and 2 ldr's(if this does not get to difficult)

Hi
Many thanks to all
I don't think the red shift would have a great effect - it's rare where I live to get very long
sunsets so only a few minutes would be lost and as I've said before sunshine hours are a bit
of a rough guide - varying from place to place.
Giving a switch to the dawn/dusk by extra LDR's is a possibility but I was trying to use readily
available (commercial) circuits so non-electronic folk (like me) could build something with as
few needs to build circuits from scratch as possible.
The whole idea was to utilise the simplicity of the Mcleod recorder in a modern way - tracking
goes off in a different direction.
I wonder if a south only facing LDR couild bias the control voltage as the existing LDR resistance
drops to about 300 ohms in daytime sun giving a wider on/off range. The original circuit does
manage dawn/dusk (by manually adjusting the control voltage) but is then wrong for daytime.
Thanks again for all the ideas - I'm convinced it will work in the end - one way or another.
Regards
Alan

no matter how elctronic it would get,the credits for the (great!)idea are yours

please keep us updated how it get's along

ps;last question
have you tried replacing the lightbulb for one with a silver reflection on it?
then maybe the light of the sunset is a bit brighterand comes into daylight range.

Hi Hans
I tried a metalised sphere (Christmas bauble) but it reflected too much cloud light - virtually
a mirror. The white bulb almost seems to absord cloud light and appears quite grey unless
it is reflecting the sun spot. I though about using a clear glass sphere (specially blown) and
coating the inside black but black behind glass is just as mirror like.
And thanks for the compliment.
Regards Alan

Speaking of compliments, I'm just awestruck at how you came up with this idea in the first place. Amazing. And when it's 'finished', I'd like to buy one, please

.. Maybe if I 'sunglassed' the LDR the sun/cloud separation would increase

You also said one of your trials was too sensitive.

How about experimenting with the aperture in front of the sensor (camera / telescope style) ?

Thanks guys
I was sitting in my garden staring up at the device (as one does) and it struck me that at the
poles the axis of the recorder head should be vertical and horizontal at the equator. The sine of
the latitude came to mind and in my simple way I looked at what would happen if that related
to the axis and not the suns virtual axis. If I take the sine of 50 degrees where I am and multiply by 90 I get 68 degrees. As I said my spherical geometry is crap but in a worst case of the sun setting in the north (as if it did here) the reflected spot would be well down in the
acceptable area (not too thin an angle) and at noon still reasonable but the sun is stronger.
Up the pole yet again !! But not till I've tried the south facing biassing LDR. That is apertured
to increase the resistance.
Regards Alan

howabout not a bias on the ldr,but on the lightbulb,to evenout the difference in brightness morning/noon?

something like this?

I have something to add to this posting:

Here is a link to a Manufacter/Distributor of Sunshine Recorders,
http://www.thiesclima.com/Sunshine.html

I hope this may give you some ideas or inspiration for your project, I myself would be very interesting in buying the end product.

Regards

Jason

Hi
Because the suns noon elevation varies from about 15 degrees to 60 winter/summer I
where I am I don't think the bulb could be 'biassed'. And thanks for the link - apart from costing
loads the Campbell Stokes needs manual intervention to changed the cards and the others
measure total radiation which can be used - but I'm in the realms of hard up amateurs.
With all this interest I reckon I'm duty bound to make it work !!!
Regards Alan

Just a thought.
How about use what you have with the globe and LDR and make some electronics to compare readings every 10 seconds or so. You would only be interested in a reasonable amount of change to signify sun or not. This would also allow for the differences in sunlight dawn to dusk. Of course you would also need to know if it was day or night. Input to a computer, this should be quite straight forward for the programmers out there but could be done in hardware only.

Hi Terry
I'm starting to feel guilty using so much of Steve's forum !
I don't think that would work. The thing works on the basis of detecting the spot of light
reflected off the bulb. When we get biggish cumulus clouds the sky can go very dark and
then can be lit up by cloud reflected light so the 10 sec comparison would see a distinct
change but not neccessarily direct sunlight.
The bias circuit is running at the moment and picked up 20mins of sun early this morning
(I wasn't up so I don't know quite when but it was thick strato cumulus when I did). Now
at 10 am BST the cloud is breaking and bits of sun are still appearing and it's still working
just as shadows appear so maybe the bias is nearly right.
Rather than modify my existing setup I'll be making another to test my angle theory and
compare the two over a few days - it takes time 'cos it needs the right sort of sunshine !!
Thanks all - Regards Alan