I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

Failing something that one might have kicking about, I wonder if the
local glazing shop would be able to cut a bit of picture frame glass
and if that is likely to be transparent for this role?

Cheers, T i m

[1] How likely is it that the fairly old DG has some sort of 'Solar
glass' in it and is already impacting the readings somewhat?

On 27/05/2021 15:47, T i m wrote:
I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

I have not read the spec on the sensor, but normal glass is opaque to
some wavelengths of IR, so that may account for the reduction in level.

(for thermal imaging cameras they often have to use lenses machined from
germanium).


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On 27/05/2021 15:47, T i m wrote:
I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

Failing something that one might have kicking about, I wonder if the
local glazing shop would be able to cut a bit of picture frame glass
and if that is likely to be transparent for this role?


In addition to John's point about IR bear in mind that reflection from
the 2 surfaces mean the light transmitted will be less than the light
incident - about 8 per cent less for normal incidence and glass with RI=1.5


--
Robin
reply-to address is (intended to be) valid

On 27/05/2021 15:47, T i m wrote:
I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

Failing something that one might have kicking about, I wonder if the
local glazing shop would be able to cut a bit of picture frame glass
and if that is likely to be transparent for this role?

Cheers, T i m

[1] How likely is it that the fairly old DG has some sort of 'Solar
glass' in it and is already impacting the readings somewhat?


Clear cast acrylic sheet is what I would try - inherently UV resistant
and with up to 92% visible light transmission - highest when the light
strikes perpendicular to the surface. I can't find values for its
transmission in the IR range though.

--
Colin Bignell

On 27/05/2021 19:39, nightjar wrote:
On 27/05/2021 15:47, T i m wrote:
I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadbandÂ* photodiode (visibleÂ* plus
infrared) and one infrared-responding photodiode on a single CMOS
integratedÂ* circuitÂ* capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

Failing something that one might have kicking about, I wonder if the
local glazing shop would be able to cut a bit of picture frame glass
and if that is likely to be transparent for this role?

Cheers, T i m

[1] How likely is it that the fairly old DG has some sort of 'Solar
glass' in it and is already impacting the readings somewhat?


Clear cast acrylic sheet is what I would try - inherently UV resistant
and with up to 92% visible light transmission - highest when the light
strikes perpendicular to the surface. I can't find values for its
transmission in the IR range though.

I think, go with whatever is to hand and re-calibrate the thresholds
expected in the software to allow for the chosen window material.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On 27/05/2021 15:47, T i m wrote:
I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

Failing something that one might have kicking about, I wonder if the
local glazing shop would be able to cut a bit of picture frame glass
and if that is likely to be transparent for this role?

Cheers, T i m

[1] How likely is it that the fairly old DG has some sort of 'Solar
glass' in it and is already impacting the readings somewhat?

No idea.

Have you considered an old camera lens glass for the cover? If you don't
have one, a camera shop selling secondhand film cameras might have some
old partially-broken lenses which could serve as a source, or even an
old UV filter. However, depending on what spectral range you are trying
to measure the light level of, you might want to have a look at this first:
https://petapixel.com/2020/06/04/why-uv-filters-are-basically-useless-on-modern-cameras/

--

Jeff

On Thu, 27 May 2021 20:26:49 +0100, Jeff Layman
wrote:

snip

Thanks for all the suggestions so far guys. A few things to try there.

Have you considered an old camera lens glass for the cover?

No, I hadn't (so thanks). I had thought of similar in the form of a
microscope slide but I can remember buying 'daylight filters' for my
SLR cameras to (sacrificially) protect the lens from physical damage.

If you don't
have one, a camera shop selling secondhand film cameras might have some
old partially-broken lenses which could serve as a source, or even an
old UV filter.

Our last camera shop went a few years ago but we do have a 'Cash
Converters' that may have a bits n bobs bin.

However, depending on what spectral range you are trying
to measure the light level of,

According to the spec I believe it covers the range of 300-1000 nm
(even though I understand we can only see around 380 to 750 nm etc)
but what I would describe as 'daylight' is all I'm really interested
in re managing general light levels indoors. Given this sensor seems
to be designed to 'see' the sort of thing we see, I think that as long
as whatever light it gets isn't filtered at all, it should be ok.

you might want to have a look at this first:
https://petapixel.com/2020/06/04/why-uv-filters-are-basically-useless-on-modern-cameras/

Interesting.

Whilst I had a few film cameras along the way I never really used them
much because of the time delay between taking the / 'a' picture and
getting the result. Ignoring my time with Kodak (microfilm / fiche)
and the Polaroid and a scanner on my PC, my first digital (A Fuji
DX-5) revolutionised it all for me and the fact that it was only
640x480 didn't matter for my needs, 'instant reference photographs'
(engineering stuff).

I hadn't thought of a round 'window' but one makes sense and one I
could easily incorporate in a 3D printed box.

I'm not trying to actually 'measure' anything (although the sorts of
levels are interesting if they are correct and when compared with the
stated light output levels of domestic lighting) but to just get the
comparable levels that could be useful to automate what might equal
'it's a bit dark outside, better turn the light on', or enable the
'turning on a light on in the hall automatically with a PIR during an
overcast day and before sunset' ... sort of thing. ;-)

Cheers, T i m

T i m wrote:
I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

Failing something that one might have kicking about, I wonder if the
local glazing shop would be able to cut a bit of picture frame glass
and if that is likely to be transparent for this role?

Cheers, T i m

[1] How likely is it that the fairly old DG has some sort of 'Solar
glass' in it and is already impacting the readings somewhat?

https://www.flir.com/products/ir-windows/

"crystal window" but it wouldn't say what it was made of

https://flir.custhelp.com/app/answer...9-transmission

"The material of the FLIR IR Windows is CaF2 (Calcium Fluoride Crystal)."

Your device probably doesn't need something that exotic.
The question would be more one of "what fitting is on your project",
like a C mount or whatever. If there was a standard mounting, you
could go to a photography shop and look at their filters. Maybe
there are filters they use for spacing or something, elements
that aren't supposed to have optical effect, and have an AR
coating as well. The AR coating might be MgF2
on some sort of glass (crown glass).

For exotic windows, be aware that some are water soluble,
some have ammonia or alcohol sensitive surfaces. Using
a less performant material might give a more environmentally
robust solution (can be wiped with a damp cloth).

Just a flat glass with an AR coating might be enough.
And the AR coating is just to avoid surprises if possible
("can't figure out what it sees").

You can also mount the material at an angle, like a Brewster window.

Paul

Is there any reason why it cannot be fitted through a hole in the case? One
could bond it in with epoxy or something, just a thought.
Brian

--

This newsgroup posting comes to you directly from...
The Sofa of Brian Gaff...

Blind user, so no pictures please
Note this Signature is meaningless.!
"John Rumm" wrote in message
o.uk...
On 27/05/2021 15:47, T i m wrote:
I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

I have not read the spec on the sensor, but normal glass is opaque to some
wavelengths of IR, so that may account for the reduction in level.

(for thermal imaging cameras they often have to use lenses machined from
germanium).


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On Fri, 28 May 2021 08:02:44 +0100, "Brian Gaff \(Sofa\)"
wrote:

Erm, yes, it probably could Brian, but I was thinking it might make it
more vulnerable to dust than if it was behind a bigger window as such?

The tiny sensor is on the surface of a postage stamp sized PCB with a
few other SM components so I'd have to print a funnel / collar shaped
hole in the case to present the actual sensor and not everything else.

Is there any reason why it cannot be fitted through a hole in the case? One
could bond it in with epoxy or something, just a thought.

Cheers, T i m

On Thu, 27 May 2021 22:45:22 -0400, Paul
wrote:

snip

[1] How likely is it that the fairly old DG has some sort of 'Solar
glass' in it and is already impacting the readings somewhat?

https://www.flir.com/products/ir-windows/

"crystal window" but it wouldn't say what it was made of

https://flir.custhelp.com/app/answer...9-transmission

"The material of the FLIR IR Windows is CaF2 (Calcium Fluoride Crystal)."

My DG is 30+ years old so I'm not sure how fancy it was in those days,
however, I've been thinking ...

Your device probably doesn't need something that exotic.

No, I don't suppose it does. ;-)

The question would be more one of "what fitting is on your project",
like a C mount or whatever.

None Paul. It's a 'bare bones' opto-sensor device mounted on a postage
stamp sized PCB with a few other surface mount components and aimed at
hobbyists and to work with an Arduino or (in my case), an ESP32 WiFi
microcontroller into my Home Assistant HA package.

https://esphome.io/components/sensor...?highlight=tsl

If there was a standard mounting, you
could go to a photography shop and look at their filters. Maybe
there are filters they use for spacing or something, elements
that aren't supposed to have optical effect, and have an AR
coating as well. The AR coating might be MgF2
on some sort of glass (crown glass).

See above ... but yes, maybe a camera addon lens might provide the
sort of window I need for this.

Looking at the spec it suggests the light incidence angle is relevant
to the sensitivity and I have had a think re what I want to use this
for. ATM I have the sensor standing on a shelf just inside the
curtains generally looking at the sky though the South facing window.
If you look at the graph from Home Assistant you can see the light
levels going up and down pretty deep/quickly as clouds pass overhead.

Whilst the 'best case' sunlight levels available might be of interest
from a weather POV, for lighting automation I think I need a more
'ambient' light level indication from say inside the window but not
facing the sky at all, just looking horizontally at the general
reflected light levels?

For exotic windows, be aware that some are water soluble,
some have ammonia or alcohol sensitive surfaces.

Noted.

Using
a less performant material might give a more environmentally
robust solution (can be wiped with a damp cloth).

That was my thought. Long term stability against UV and easily
dustable. If the sensor window was vertical so the sensor can look
horizontally that might also help stop the window gather dust. ;-)

Just a flat glass with an AR coating might be enough.
And the AR coating is just to avoid surprises if possible
("can't figure out what it sees").

I think anything that lets the widest spectrum of light though would
be good but maybe a UV filter might help against damage from UV to the
sensor or other components (whilst not impacting the 'light
readings')?

You can also mount the material at an angle, like a Brewster window.

Stuck vertically on the widow reveal behind the curtains may well give
me the result I want.

Maybe a little glass box (jewellery display case top) might work and I
can just print the base to hold the PCB and provide a wall /
windowsill mounting?

Cheers, T i m

On 27/05/2021 20:04, John Rumm wrote:
On 27/05/2021 19:39, nightjar wrote:
....
Clear cast acrylic sheet is what I would try - inherently UV resistant
and with up to 92% visible light transmission - highest when the light
strikes perpendicular to the surface. I can't find values for its
transmission in the IR range though.

I think, go with whatever is to hand and re-calibrate the thresholds
expected in the software to allow for the chosen window material.

You do, however, need to be sure that, whatever you use, is not degraded
by UV. That would lead to a progressive loss of transmission over time.
That is why I like acrylic sheet - it is completely UV resistant without
the need for any coating.

It is also likely to be what I have to hand :-)

--
Colin Bignell

On 27/05/2021 20:26, Jeff Layman wrote:
Have you considered an old camera lens glass for the cover? If you don't
have one, a camera shop selling secondhand film cameras might have some
old partially-broken lenses which could serve as a source, or even an
old UV filter. However, depending on what spectral range you are trying
to measure the light level of, you might want to have a look at this first:
https://petapixel.com/2020/06/04/why-uv-filters-are-basically-useless-on-modern-cameras/


bog standard UV filters might work - use em as lens protectors

random example gave me a clear glass protector

https://www.pricerunner.com/pl/282-3...Compare-Prices

--
New Socialism consists essentially in being seen to have your heart in
the right place whilst your head is in the clouds and your hand is in
someone else's pocket.

On 28/05/2021 10:13, nightjar wrote:
On 27/05/2021 20:04, John Rumm wrote:
On 27/05/2021 19:39, nightjar wrote:
...
Clear cast acrylic sheet is what I would try - inherently UV
resistant and with up to 92% visible light transmission - highest
when the light strikes perpendicular to the surface. I can't find
values for its transmission in the IR range though.

I think, go with whatever is to hand and re-calibrate the thresholds
expected in the software to allow for the chosen window material.

You do, however, need to be sure that, whatever you use, is not degraded
by UV. That would lead to a progressive loss of transmission over time.
That is why I like acrylic sheet - it is completely UV resistant without
the need for any coating.

It is also likely to be what I have to hand :-)

Yup acrylic and proper glass should be fine. Polycarbonate might not be
so good in this case. Having said that it can still get dirty outside
regardless of material - so a bit of slack built into the thresholds is
not a bad idea.


--
Cheers,

John.

/================================================== ===============\
| Internode Ltd - http://www.internode.co.uk |
|-----------------------------------------------------------------|
| John Rumm - john(at)internode(dot)co(dot)uk |
\================================================= ================/

On 27/05/2021 23:27, T i m wrote:

According to the spec I believe it covers the range of 300-1000 nm
(even though I understand we can only see around 380 to 750 nm etc)
but what I would describe as 'daylight' is all I'm really interested
in re managing general light levels indoors. Given this sensor seems
to be designed to 'see' the sort of thing we see, I think that as long
as whatever light it gets isn't filtered at all, it should be ok.

The sensor you've chosen can see out to 1000nm. Most glasses and
plastics will transit light to near IR

The advantage of using a UV filter is the surfaces will be optically
coated to minimise reflection and maximise transmission over the
wavelengths of interest.

snip

I'm not trying to actually 'measure' anything (although the sorts of
levels are interesting if they are correct and when compared with the
stated light output levels of domestic lighting) but to just get the
comparable levels that could be useful to automate what might equal
'it's a bit dark outside, better turn the light on', or enable the
'turning on a light on in the hall automatically with a PIR during an
overcast day and before sunset' ... sort of thing. ;-)

Interesting. Given everything is relative and open to calibration, crude
or otherwise, I would say any old piece of glass or plastic window would
be fine.

On 27/05/2021 23:27, T i m wrote:

snip


I'm not trying to actually 'measure' anything (although the sorts of
levels are interesting if they are correct and when compared with the
stated light output levels of domestic lighting) but to just get the
comparable levels that could be useful to automate what might equal
'it's a bit dark outside, better turn the light on', or enable the
'turning on a light on in the hall automatically with a PIR during an
overcast day and before sunset' ... sort of thing. ;-)


For that purpose I think you can safely forget about losses from
anything that starts off looking clear and is stable. The odd 20% is
neither here nor there. The difference in illuminance between direct
sunlight and an overcast day is around a factor 100. It's another
factor 10 or so down to a typical living room at night.

JOOI does the home automation software have built-in facility to ignore
short-term changes so e.g. the headlights of a car arriving don't turn
off the lights on the stairs?

--
Robin
reply-to address is (intended to be) valid

On Friday, May 28, 2021 at 2:31:28 PM UTC+1, Robin wrote:
On 27/05/2021 23:27, T i m wrote:
snip
I'm not trying to actually 'measure' anything (although the sorts of
levels are interesting if they are correct and when compared with the
stated light output levels of domestic lighting) but to just get the
comparable levels that could be useful to automate what might equal
'it's a bit dark outside, better turn the light on', or enable the
'turning on a light on in the hall automatically with a PIR during an
overcast day and before sunset' ... sort of thing. ;-)


JOOI does the home automation software have built-in facility to ignore
short-term changes so e.g. the headlights of a car arriving don't turn
off the lights on the stairs?
--
Robin
reply-to address is (intended to be) valid

A friend of mine related a story of a friend of his who had installed a motorised curtain system in his lounge activated by it going dark. He had my friend round along with some other mates to witness the system working. Sure enough as it fell dark the curtains closed only for a few minutes later for them to suddenly start to open and then close again this continued all night every time a car went past.

Fortunately he had not installed the system in the bedroom, I can only imagine the embarrassment being caught €œin flagantre€�

Richard

On Fri, 28 May 2021 14:31:22 +0100, Robin wrote:

On 27/05/2021 23:27, T i m wrote:

snip


I'm not trying to actually 'measure' anything (although the sorts of
levels are interesting if they are correct and when compared with the
stated light output levels of domestic lighting) but to just get the
comparable levels that could be useful to automate what might equal
'it's a bit dark outside, better turn the light on', or enable the
'turning on a light on in the hall automatically with a PIR during an
overcast day and before sunset' ... sort of thing. ;-)


For that purpose I think you can safely forget about losses from
anything that starts off looking clear and is stable.

The bit of very clear blister pack seemed to have quite an impact so
I'll run some tests and compare. ;-)

The odd 20% is
neither here nor there.

True, if I'm only interested in the difference in levels etc.

The difference in illuminance between direct
sunlight and an overcast day is around a factor 100.

Yeah, So I've already observed, or even a cloud going over.

https://ibb.co/QHfKhv9

It's another
factor 10 or so down to a typical living room at night.

Noted.

JOOI does the home automation software have built-in facility to ignore
short-term changes so e.g. the headlights of a car arriving don't turn
off the lights on the stairs?

Hehe ... erm, I'll say 'yes' in as much as you can set thresholds and
delays plus conditions where once triggered (because it was much
darker than usual say 1hr before sunset) then it would trigger the
lights on but wouldn't allow the lights to be turned back off till say
midnight or sunrise etc.

Within each Automation you can have multiple triggers, multiple
conditions and multiple actions (turning on several lights at specific
brightness etc).

Cheers, T i m

On Thu, 27 May 2021 15:47:39 +0100, T i m wrote:


I have added an ambient light level sensor to my Home Assistant setup
and it seems to work very well (snapshot taken just now).

https://ibb.co/QHfKhv9

I've currently got it on an internal South facing windowsill so it
will already be looking though the double glazing [1] but I want to
print a case for it with it's own 'window' to keep the dust out I just
tried putting some plastic cut from a blister pack over it to see if
it would have any impact to the reading. It did, it went from about
11,000 lux to about 9,000 so I'm guessing it's only 'translucent' and
whilst I'm really only interested in the change of lighting levels
(for general lighting automation) I'd rather any such window didn't
impact the reading more than necessary.

The actual sensor is a TSL2561:
https://cdn-shop.adafruit.com/datasheets/TSL2561.pdf

And it suggests it combines the output of two sensors ...

"Each device combines one broadband photodiode (visible plus
infrared) and one infrared-responding photodiode on a single CMOS
integrated circuit capable of providing a near-photopic response
over an effective 20-bit dynamic range (16-bit resolution)."

So I was wondering if anyone could recommend a source of a small piece
of something as a 'window' that would be transparent for this role?

Failing something that one might have kicking about, I wonder if the
local glazing shop would be able to cut a bit of picture frame glass
and if that is likely to be transparent for this role?

Cheers, T i m

[1] How likely is it that the fairly old DG has some sort of 'Solar
glass' in it and is already impacting the readings somewhat?


As your window glass will have attenuated the light a lot, the best
you could do, short of putting it outside with some proper covering,
would be to stick it to the wundow without any extra covering, sealed
so that dust can't get in.
--
Dave W

On Fri, 28 May 2021 19:10:38 +0100, Dave W
wrote:

snip

As your window glass will have attenuated the light a lot, the best
you could do, short of putting it outside with some proper covering,
would be to stick it to the wundow without any extra covering, sealed
so that dust can't get in.

Ah, now there is a bit of out of the box thinking. ;-)

Yes, it would be easy to print a small box with a reasonable (~5mm)
flange and simply stick it to the inside of the window (facing out
etc) with some double sided tape. If I leave a some ventilation
underneath it shouldn't fog etc.

I could have two, one 'looking' outside and one monitoring the ambient
light levels in say the hall.

Cheers, T i m

On 28/05/2021 12:11, John Rumm wrote:
On 28/05/2021 10:13, nightjar wrote:
On 27/05/2021 20:04, John Rumm wrote:
On 27/05/2021 19:39, nightjar wrote:
...
Clear cast acrylic sheet is what I would try - inherently UV
resistant and with up to 92% visible light transmission - highest
when the light strikes perpendicular to the surface. I can't find
values for its transmission in the IR range though.

I think, go with whatever is to hand and re-calibrate the thresholds
expected in the software to allow for the chosen window material.

You do, however, need to be sure that, whatever you use, is not
degraded by UV. That would lead to a progressive loss of transmission
over time. That is why I like acrylic sheet - it is completely UV
resistant without the need for any coating.

It is also likely to be what I have to hand :-)

Yup acrylic and proper glass should be fine. Polycarbonate might not be
so good in this case. Having said that it can still get dirty outside
regardless of material - so a bit of slack built into the thresholds is
not a bad idea.


A microscope cover-slip is a good source of thin flat glass.

T i m wrote:

The bit of very clear blister pack seemed to have quite an impact so
I'll run some tests and compare. ;-)

What are your objectives, for this overly scientific sensor ?

I have a sensor which mounts on the carriage lamp out front
of the house. It costs about as much as a light switch at
the hardware store.

Photocell sensor. Mine is brick-shaped, with the CdS sensor
mounted on the side of it.

https://images-na.ssl-images-amazon...._QL70_ML2_.jpg

"Time Delay Function : Accurately judges natural light,
won't be triggered due to occasional light such as
car headlights, lightning or other temporary light.

It's suitable for parking lot, street lighting,
outdoor lighting, factory, indoor lighting"

(Example of a CdS sensor as a component...

https://media.digikey.com/photos/Adv..._PDV-P5001.jpg )

The thing has at least three wires, and there's probably
a Triac in there doing the appliance switching.

And their claim in that description is accurate. I've
never seen a false trigger on the thing. A car can drive
by, the light stays on. It's not a motion sensor. The light
it controls, is not some kind of strobe show, like all the
neighbours lights. Mine failed at some point, I bought a
new one and fitted the sensor to the pipe-with-a-hole it
mounts inside. And it's been running for probably 20 years
at this point.

Now, the Triac part is no good for your purposes. But
as far as the sensor function is concerned, it accurately
detects dusk. The light will come on whenever there is a
"complete solar eclipse". Any condition which is sufficiently
dark and sustained long enough, will switch it on. And I
don't go out and "clean the sensor face" - there is no
maintenance whatsoever.

It was cheap, and surprisingly fit for purpose (shocker!).
At the price, I was expecting Magic Smoke or Quibbles.

Paul

On Fri, 28 May 2021 22:31:59 -0400, Paul
wrote:

T i m wrote:

The bit of very clear blister pack seemed to have quite an impact so
I'll run some tests and compare. ;-)

What are your objectives,

I don't have one specifically Paul, it's just another opportunity to
try another sensor, see how it integrates into Home Assistant and from
there, see how I can make use of it. One thought of a use was to
'enhance' the electric curtains so rather than just dusk / dawn (+-
whatever offset I set) I could use those + actual light levels, as
some days 'sunset;' is still reasonably light whereas others it's
quite dark'. Apparent you can also use sun angle that takes into
account your height above sea level but I've not tried that yet).

for this overly scientific sensor ?

I'm not sure if it's 'overly scientific', just that it offered
'predictable support / integration into ESPHome / Home Assistant (all
the support libraries in place), cheap, connects via I2C (so simple to
interface) and 'calibrated', as opposed to using a LDR etc.

I have a sensor which mounts on the carriage lamp out front
of the house. It costs about as much as a light switch at
the hardware store.

This probably cost less (inc the ESP32 board that gives wireless
access). ;-)

Photocell sensor. Mine is brick-shaped, with the CdS sensor
mounted on the side of it.

https://images-na.ssl-images-amazon...._QL70_ML2_.jpg

Noted.

"Time Delay Function : Accurately judges natural light,
won't be triggered due to occasional light such as
car headlights, lightning or other temporary light.

It's suitable for parking lot, street lighting,
outdoor lighting, factory, indoor lighting"

(Example of a CdS sensor as a component...

https://media.digikey.com/photos/Adv..._PDV-P5001.jpg )

The thing has at least three wires, and there's probably
a Triac in there doing the appliance switching.

I didn't really want to 'switch' anything, especially as some preset
level (if not adjustable etc), I wanted to dynamically 'read' the
light levels and then I can decide how I want to make use of that.
This was based from previous experiences with 'light sensing' lamps
and lamp fitting adaptors that have no user adjustable sensitivity
settings.


And their claim in that description is accurate. I've
never seen a false trigger on the thing. A car can drive
by, the light stays on. It's not a motion sensor. The light
it controls, is not some kind of strobe show, like all the
neighbours lights. Mine failed at some point, I bought a
new one and fitted the sensor to the pipe-with-a-hole it
mounts inside. And it's been running for probably 20 years
at this point.

Yeah, commercial photo-switches seem to be a strange thing, either
lasting for ever or a very short time.

Now, the Triac part is no good for your purposes. But
as far as the sensor function is concerned, it accurately
detects dusk. The light will come on whenever there is a
"complete solar eclipse". Any condition which is sufficiently
dark and sustained long enough, will switch it on. And I
don't go out and "clean the sensor face" - there is no
maintenance whatsoever.

This was more of an electronics, wireless, Home Automation,
datalogging, 'you can manage what you can measure' sort of project,
rather than a straight light level switching task.

For example, if I wanted I could set all the lights in the house to
come on during a solar eclipse. ;-)

It was cheap, and surprisingly fit for purpose (shocker!).
At the price, I was expecting Magic Smoke or Quibbles.

From my experience of them they normally just die unspectacularly. ;-(

Previously I had the sensor looking direct out and up (towards the sky
/ sun) and logged this:
https://ibb.co/QHfKhv9

Yesterday I turned it though 90 degrees and had it facing horizontally
in the window reveal and have just picked up this:
https://ibb.co/FwXfY0y

And the point of this sort of solution is that you can also go back
over time (not sure how far) and select whatever sensors or switches
you want over whatever period you want and display them side-by-side.

Cheers, T i m