I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them on
24 hrs to check, I was surprised that they lasted over a month which
makes them very useful as party lights in a damp conservatory, let's say
running them for 4hrs a day which should make them last for 6 months
perhaps.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.
Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and
the time constant of the filament to provide a non flickering and
dimmable light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.
In the interests of battery life, are there any dimmers of this type
available for leds?

Cheers
Don

On 22/11/09 11:03, Donwill wrote:

you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio
Doing that to an led would cause flickering of the light

Not perceptible if you do it at several kHz

"Donwill" wrote in message
...
I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them on 24
hrs to check, I was surprised that they lasted over a month which makes
them very useful as party lights in a damp conservatory, let's say running
them for 4hrs a day which should make them last for 6 months perhaps.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.
Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and the
time constant of the filament to provide a non flickering and dimmable
light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.
In the interests of battery life, are there any dimmers of this type
available for leds?

Cheers
Don

A 555 timer will prolly do the job
http://www.kpsec.freeuk.com/555timer.htm
Of course, the cost of your time in making it would buy a good few batteries

brass monkey wrote:

A 555 timer will prolly do the job
http://www.kpsec.freeuk.com/555timer.htm
Of course, the cost of your time in making it would buy a good few batteries


A bit of slightly opaque paint dabbed over the LED's would be even cheaper.


--
Adrian C

brass monkey wrote:
"Donwill" wrote in message
...

I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them on 24
hrs to check, I was surprised that they lasted over a month which makes
them very useful as party lights in a damp conservatory, let's say running
them for 4hrs a day which should make them last for 6 months perhaps.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.
Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and the
time constant of the filament to provide a non flickering and dimmable
light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.
In the interests of battery life, are there any dimmers of this type
available for leds?

Cheers
Don


A 555 timer will prolly do the job
http://www.kpsec.freeuk.com/555timer.htm
Of course, the cost of your time in making it would buy a good few batteries




Unless you'r an OAP like me. :-)
Cheers
Don

Adrian C wrote:
brass monkey wrote:

A 555 timer will prolly do the job
http://www.kpsec.freeuk.com/555timer.htm
Of course, the cost of your time in making it would buy a good few
batteries

A bit of slightly opaque paint dabbed over the LED's would be even
cheaper.


but a little difficult to adjust the brightness perhaps :-)
Don

On Sun, 22 Nov 2009 11:03:02 +0000, Donwill
wrote:

I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them on
24 hrs to check, I was surprised that they lasted over a month which
makes them very useful as party lights in a damp conservatory,

Sounds a fun place to have a party :-)


--
Frank Erskine

On 22 Nov, 11:03, Donwill wrote:
I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts *presumably.
I was a bit sceptical about the life of the batteries so I left them on
24 hrs to check, I was surprised that they lasted over a month which
makes them very useful as party lights in a damp conservatory, let's say
running them for 4hrs a day which should make them last for 6 months
perhaps.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.
Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and
the time constant of the filament to provide a non flickering and
dimmable light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.
In the interests of battery life, are there any dimmers of this type
available for leds?

Cheers
Don

Pulse width modultaion, subject of several U.S. patents in relation to
LED dimming but thats a rant for another time, 555 as suggested can do
it, fly swimming slowly in ointment is V drop, only have 3V to play
with , drop 0.7V across a semi junction, even low Rds Fet will drop a
bit.

Couple of strings in series to get voltage up and then dim them.

Cheers
Adam

"Frank Erskine" wrote in message
...
On Sun, 22 Nov 2009 11:03:02 +0000, Donwill
wrote:

I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them on
24 hrs to check, I was surprised that they lasted over a month which
makes them very useful as party lights in a damp conservatory,

Sounds a fun place to have a party :-)

maybe it's a BDSM party and the conservatory is the code word for his
dungeon

pretty easy to dim leds as mentioned, but another circuit you could build is
the joule theif, just search google for led joule theif,
these are pretty simple, an inductor transistor and resistor, and you can
drain the very last dregs from a battery lighting the led's up,
common uses are to run a white led off a 1.5 volt practicaly dead battery,

"Adam Aglionby" wrote in message
...

Pulse width modultaion, subject of several U.S. patents in relation to
LED dimming but thats a rant for another time, 555 as suggested can do
it, fly swimming slowly in ointment is V drop, only have 3V to play
with , drop 0.7V across a semi junction, even low Rds Fet will drop a
bit.

At which point maybe the resistor may be as efficient, depending on current.

On Sun, 22 Nov 2009 11:03:02 +0000, Donwill wrote:

I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably. I was a bit sceptical
about the life of the batteries so I left them on 24 hrs to check, I was
surprised that they lasted over a month which makes them very useful as
party lights in a damp conservatory, let's say running them for 4hrs a
day which should make them last for 6 months perhaps.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.
Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and
the time constant of the filament to provide a non flickering and
dimmable light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.
In the interests of battery life, are there any dimmers of this type
available for leds?



This gets to be fun if you are restricting your supply to 3v. Here's a
dimmer for 12v: http://www.reuk.co.uk/LED-Dimmer-Circuit.htm
You may be able to get it to work using a CMOS 555 (which works down to
about 1.5v) and, maybe, a MOSFET such as 2N7000 instead of the transistor
to give better efficiency at the low voltage. You could decrease C1 to
increase the frequency if you get any flickering. R2 can be omitted if a
MOSFET is used.

--
Mick (Working in a M$-free zone!)
Web: http://www.nascom.info
Filtering everything posted from googlegroups to kill spam.

Donwill explained :
Now you can dim filament bulbs by using an astable multivibrator and varying
the mark to space ratio, relying on the chopping frequency and the time
constant of the filament to provide a non flickering and dimmable light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film movies
which I think have to be in excess of 25 fps in order to fool the eye.
In the interests of battery life, are there any dimmers of this type
available for leds?

It works, but for LEDs you need to increase the switching frequency to
be much higher - say around 1Khz. The filament in a lamp also takes
some time to heat up and cool down, giving it some persistence.

--
Regards,
Harry (M1BYT) (L)
http://www.ukradioamateur.co.uk

On Nov 22, 1:40*pm, "gazz" wrote:
"Frank Erskine" wrote in message

...

On Sun, 22 Nov 2009 11:03:02 +0000, Donwill
wrote:

I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts *presumably.
I was a bit sceptical about the life of the batteries so I left them on
24 hrs to check, I was surprised that they lasted over a month which
makes them very useful as party lights in a damp conservatory,

Sounds a fun place to have a party :-)

maybe it's a BDSM party and the conservatory is the code word for his
dungeon

pretty easy to dim leds as mentioned, but another circuit you could build is
the joule theif, just search google for led joule theif,
these are pretty simple, an inductor transistor and resistor, and you can
drain the very last dregs from a battery lighting the led's up,
common uses are to run a white led off a 1.5 volt practicaly dead battery,

The joule thief pulses, so you can incoporate dimming by delaying its
switching on each cycle.


NT

wrote:

It would be easy to knock up electronics do do so.

A lot of vehicle lights are dimmed by pulsing them. If you observe LED rear
lights and turn your head there is often a flicker. When the same lights are
used as brake lights the flicker disappears as they are then fully on. The
frequency used must vary as some lights are more noticably flickery than
others.

I'd go for at least 100Hz to minimise the flicker. Much higher is easy to
achieve.


Go up in the kilohertz range, or you'll annoy some people.

As many car manufacturers are currently doing with cheap circuits to
save a few pence, sometimes on cars costing 100k.

Andy

On Sun, 22 Nov 2009 17:03:38 +0000, Andy Champ wrote:

wrote:

It would be easy to knock up electronics do do so.

A lot of vehicle lights are dimmed by pulsing them. If you observe LED rear
lights and turn your head there is often a flicker. When the same lights are
used as brake lights the flicker disappears as they are then fully on. The
frequency used must vary as some lights are more noticably flickery than
others.

I'd go for at least 100Hz to minimise the flicker. Much higher is easy to
achieve.


Go up in the kilohertz range, or you'll annoy some people.

As many car manufacturers are currently doing with cheap circuits to
save a few pence, sometimes on cars costing 100k.

Andy

Yes, I'm fed up of having a string of red squares or circles across my
vision as I look from one side to the other behind an LED lit car! The red
man on newer pedestrian crossings can have the same effect.

SteveW

Steve Walker
wibbled on Sunday 22 November 2009 20:47


Yes, I'm fed up of having a string of red squares or circles across my
vision as I look from one side to the other behind an LED lit car! The red
man on newer pedestrian crossings can have the same effect.

SteveW

There's simply no excuse for it. I was messing around with AVR
microprocessors and RGB LEDs, doing PWM on each element with a fixed cycle
period and variable mark/space. No problems running in the many kHz or even
10's of kHz.

Now, if for some reason this is problematic over metres of rough car wiring
looms, then the answer (if they want to be clever) is to build the lamp
controllers fairly local to the lamp clusters and slick them on one of the
control busses. Doesn't have to be one per lamp, could be 2 or 4 for
front/back or corners, plus another set for dash and interior.

I assumed they did something like this anyway to cut down on the number of
heavy copper conductors running everywhere, but I haven't looked hard
enough at my car to inspect the wiring looms to the lights.

If that is undesireable for them, (ie they are cheap or want to be simple)
then they would be better sticking with simple bulbs (which could be LED
bulbs in the same format with the same drive).

--
Tim Watts

This space intentionally left blank...

On Sun, 22 Nov 2009 21:18:20 +0000, Tim W wrote:

Steve Walker
wibbled on Sunday 22 November 2009 20:47


Yes, I'm fed up of having a string of red squares or circles across my
vision as I look from one side to the other behind an LED lit car! The red
man on newer pedestrian crossings can have the same effect.

SteveW

There's simply no excuse for it. I was messing around with AVR
microprocessors and RGB LEDs, doing PWM on each element with a fixed cycle
period and variable mark/space. No problems running in the many kHz or even
10's of kHz.

Now, if for some reason this is problematic over metres of rough car wiring
looms, then the answer (if they want to be clever) is to build the lamp
controllers fairly local to the lamp clusters and slick them on one of the
control busses. Doesn't have to be one per lamp, could be 2 or 4 for
front/back or corners, plus another set for dash and interior.

I assumed they did something like this anyway to cut down on the number of
heavy copper conductors running everywhere, but I haven't looked hard
enough at my car to inspect the wiring looms to the lights.

If that is undesireable for them, (ie they are cheap or want to be simple)
then they would be better sticking with simple bulbs (which could be LED
bulbs in the same format with the same drive).

I must admit that I thought that as more and more cars use canbus the lamp
controllers would all be local to the lamps. As you say, this reduces the
weight of copper that is needed.

As for the the frequency, indeed, there is no excuse for using such a low
rate.

SteveW

"Donwill" wrote in message
...
I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them on 24
hrs to check, I was surprised that they lasted over a month which makes
them very useful as party lights in a damp conservatory, let's say running
them for 4hrs a day which should make them last for 6 months perhaps.

Hang on a minute, many LEDs won't turn on at 3v as that's quite a low Vf.
Not only that but the normal arrangement for multiple LEDs is to put them in
series chains. You could put them in parallel but there are issues. I
suspect then that the 3V from your batteries is steped up to something
rather higher.

LEDs require a current source rather than a voltage and there are some nice
little ICs available which do all of the step up and current control in one
package, you may well have one of these in the lights you've bought. Have a
look near the batteries, if they're cheapo it may be a small black blob
rather than a chip as such.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.

If the lights use an IC to do step up and control then no, a resistor won't
work, or at least not in the way you think.
Even for simple resistor in series with LED circuits you wouldn't reduce the
battery life by increasing the resistor value. You would waste power of
course by heating the resistor but the overall current draw would be less so
the battery would last longer.

Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and the
time constant of the filament to provide a non flickering and dimmable
light.

Again anything like this will screw up an IC controller, if that's what
you've got.

Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.

Movies are 24Hz frame rate *but* they are double or treble shuttered to
avoid flicker. 24Hz is objectionable and unwatchable, 48/50Hz is right on
the limit and anything you'll see in a UK cinema will reach your eye at
72Hz.

In the interests of battery life, are there any dimmers of this type
available for leds?

On 22 Nov, 23:48, "Calvin Sambrook" wrote:
"Donwill" wrote in message

...

I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts *presumably.
I was a bit sceptical about the life of the batteries so I left them on 24
hrs to check, I was surprised that they lasted over a month which makes
them very useful as party lights in a damp conservatory, let's say running
them for 4hrs a day which should make them last for 6 months perhaps.

Hang on a minute, many LEDs won't turn on at 3v as that's quite a low Vf.
Not only that but the normal arrangement for multiple LEDs is to put them in
series chains. *You could put them in parallel but there are issues. *I
suspect then that the 3V from your batteries is steped up to something
rather higher.

Was thinking that but presume OP has red, green and amber chains,
red ,amber and green LEDs will run off sub 3V, blue and white need 3V+
to light and typically have 3 AA batteries.

B&M Homestores have a good range of both mains and battery operated
LED sets this year.

Cheers
Adam


LEDs require a current source rather than a voltage and there are some nice
little ICs available which do all of the step up and current control in one
package, you may well have one of these in the lights you've bought. *Have a
look near the batteries, if they're cheapo it may be a small black blob
rather than a chip as such.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.

If the lights use an IC to do step up and control then no, a resistor won't
work, or at least not in the way you think.
Even for simple resistor in series with LED circuits you wouldn't reduce the
battery life by increasing the resistor value. *You would waste power of
course by heating the resistor but the overall current draw would be less so
the battery would last longer.

Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and the
time constant of the filament to provide a non flickering and dimmable
light.

Again anything like this will screw up an IC controller, if that's what
you've got.

Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.

Movies are 24Hz frame rate *but* they are double or treble shuttered to
avoid flicker. *24Hz is objectionable and unwatchable, 48/50Hz is right on
the limit and anything you'll see in a UK cinema will reach your eye at
72Hz.

In the interests of battery life, are there any dimmers of this type
available for leds?

Calvin Sambrook wrote:
"Donwill" wrote in message
...
I purchased a string of 10 Leds powered by 2 AA batteries connected
in series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them
on 24 hrs to check, I was surprised that they lasted over a month
which makes them very useful as party lights in a damp conservatory,
let's say running them for 4hrs a day which should make them last for
6 months perhaps.

Hang on a minute, many LEDs won't turn on at 3v as that's quite a low
Vf. Not only that but the normal arrangement for multiple LEDs is to
put them in series chains. You could put them in parallel but there
are issues. I suspect then that the 3V from your batteries is steped
up to something rather higher.

LEDs require a current source rather than a voltage and there are some
nice little ICs available which do all of the step up and current
control in one package, you may well have one of these in the lights
you've bought. Have a look near the batteries, if they're cheapo it
may be a small black blob rather than a chip as such.
The leds are wired in parallell and I have looked, and checked again
after your email for a black blob or any indication of some kind of
electronic gizmo that might control current or whatever, I am pretty
sure there is nothing there. the leds have transparent plastic housings
and I can see nothing in the housings that might indicate that they have
anything significant in there either.


I got to thinking about dimming them, a series resistor would surely
do that, but it would reduce the battery life by warming up the
resistor.

If the lights use an IC to do step up and control then no, a resistor
won't work, or at least not in the way you think.
Even for simple resistor in series with LED circuits you wouldn't
reduce the battery life by increasing the resistor value. You would
waste power of course by heating the resistor but the overall current
draw would be less so the battery would last longer.

Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency
and the time constant of the filament to provide a non flickering and
dimmable light.

Again anything like this will screw up an IC controller, if that's
what you've got.
I don't think there is one, it's just a simple translucent battery
holder which includes a switch, no sign of a "black blob" there either.
The possible explanation that occured to me is that they have a
constant current diode built into the leds during the manufacturing
process and is visually quite undetectable.


Doing that to an led would cause flickering of the light but would
the relativly slow response of the eye smooth out the flicker? as in
film movies which I think have to be in excess of 25 fps in order to
fool the eye.

Movies are 24Hz frame rate *but* they are double or treble shuttered
to avoid flicker. 24Hz is objectionable and unwatchable, 48/50Hz is
right on the limit and anything you'll see in a UK cinema will reach
your eye at 72Hz.
OK, thanks, I hadn't realised that, I was thinking to set the multivib
to 1KHz at least.

In the interests of battery life, are there any dimmers of this type
available for leds?

I purchased them on ebay and also bought a string of 20 and the only
discernable difference between them is that the 20 string has 3 AA
batteries in series.
I eventually cut off the battery housings and connected the two strings
in parallell powered by a Nokia mains charger ACP-7X which were being
sold off cheaply, if I remember correctly for 99p, the charger got a
little warm after 10mins use so I inserted a 2ohm power resistor in
series which reduced the current sufficiently to keep the charger
temperature within bounds. The subsequent light output was reduced
obviously but still adequate for my purposes.
This was just an experiment as I did not want to introduce mains
voltages to the still drying out conservatory (damp).

Thanks
Don

mick wrote:
On Sun, 22 Nov 2009 11:03:02 +0000, Donwill wrote:


I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably. I was a bit sceptical
about the life of the batteries so I left them on 24 hrs to check, I was
surprised that they lasted over a month which makes them very useful as
party lights in a damp conservatory, let's say running them for 4hrs a
day which should make them last for 6 months perhaps.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.
Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and
the time constant of the filament to provide a non flickering and
dimmable light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.
In the interests of battery life, are there any dimmers of this type
available for leds?




This gets to be fun if you are restricting your supply to 3v. Here's a
dimmer for 12v: http://www.reuk.co.uk/LED-Dimmer-Circuit.htm
You may be able to get it to work using a CMOS 555 (which works down to
about 1.5v) and, maybe, a MOSFET such as 2N7000 instead of the transistor
to give better efficiency at the low voltage. You could decrease C1 to
increase the frequency if you get any flickering. R2 can be omitted if a
MOSFET is used.


Found this also:
http://www.opensourcepartners.nl/~costar/leddimmer/
which might be interesting to play with.
Cheers
Don

"Donwill" wrote in message
...
Calvin Sambrook wrote:
"Donwill" wrote in message
...
I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left them on
24 hrs to check, I was surprised that they lasted over a month which
makes them very useful as party lights in a damp conservatory, let's say
running them for 4hrs a day which should make them last for 6 months
perhaps.

Hang on a minute, many LEDs won't turn on at 3v as that's quite a low Vf.
Not only that but the normal arrangement for multiple LEDs is to put them
in series chains. You could put them in parallel but there are issues.
I suspect then that the 3V from your batteries is steped up to something
rather higher.

LEDs require a current source rather than a voltage and there are some
nice little ICs available which do all of the step up and current control
in one package, you may well have one of these in the lights you've
bought. Have a look near the batteries, if they're cheapo it may be a
small black blob rather than a chip as such.
The leds are wired in parallell and I have looked, and checked again
after your email for a black blob or any indication of some kind of
electronic gizmo that might control current or whatever, I am pretty sure
there is nothing there. the leds have transparent plastic housings and I
can see nothing in the housings that might indicate that they have
anything significant in there either.

I'm still struggling to imagine how these are arranged. If they are in
parallel then each LED will have a wire going to it so that's 10 wires
running along the "string" plus the return which can be common which makes
11 in all. Each LED will require a current limiting resistance of some
sort. They will not share a limiting resistor as a mis-match between Vf of
individual LEDs leads to runaway conditions and dead LEDs so there will be
10 resistors. This is however a poor way to design battery powered LED
lights as If will vary significantly with battery terminal voltage, bear in
mind that a fully charged AA cell may be anywhere between 1.2V and 1.5V
depending on its chemistry.

It's just never done this way, in fact the only case where I've seen LEDs
controlled with a resistance commercially is in keyring lights where the
relatively high internal resistance of two button cells is relied on to
limit the current through a single LED. Your chain simply can't be using
the internal resistance of AA batteries as that can be very low in some
cases. In fact as a fun experiment try fitting NiCd or NiMh cells, either
it won't light because they have a lower V or the LEDs will pop because they
have a very low R.

I got to thinking about dimming them, a series resistor would surely do
that, but it would reduce the battery life by warming up the resistor.

If the lights use an IC to do step up and control then no, a resistor
won't work, or at least not in the way you think.
Even for simple resistor in series with LED circuits you wouldn't reduce
the battery life by increasing the resistor value. You would waste power
of course by heating the resistor but the overall current draw would be
less so the battery would last longer.

Now you can dim filament bulbs by using an astable multivibrator and
varying the mark to space ratio, relying on the chopping frequency and
the time constant of the filament to provide a non flickering and
dimmable light.

Again anything like this will screw up an IC controller, if that's what
you've got.
I don't think there is one, it's just a simple translucent battery holder
which includes a switch, no sign of a "black blob" there either. The
possible explanation that occured to me is that they have a constant
current diode built into the leds during the manufacturing process and is
visually quite undetectable.

Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool the
eye.

Movies are 24Hz frame rate *but* they are double or treble shuttered to
avoid flicker. 24Hz is objectionable and unwatchable, 48/50Hz is right
on the limit and anything you'll see in a UK cinema will reach your eye
at 72Hz.
OK, thanks, I hadn't realised that, I was thinking to set the multivib to
1KHz at least.

In the interests of battery life, are there any dimmers of this type
available for leds?

I purchased them on ebay and also bought a string of 20 and the only
discernable difference between them is that the 20 string has 3 AA
batteries in series.
I eventually cut off the battery housings and connected the two strings in
parallell powered by a Nokia mains charger ACP-7X which were being sold
off cheaply, if I remember correctly for 99p, the charger got a little
warm after 10mins use so I inserted a 2ohm power resistor in series which
reduced the current sufficiently to keep the charger temperature within
bounds. The subsequent light output was reduced obviously but still
adequate for my purposes.
This was just an experiment as I did not want to introduce mains voltages
to the still drying out conservatory (damp).

What voltage is the mains charger? If you have access to a multimeter why
not measure the current.

Calvin Sambrook wrote:
"Donwill" wrote in message
...
Calvin Sambrook wrote:
"Donwill" wrote in message
...
I purchased a string of 10 Leds powered by 2 AA batteries connected
in series giving 3 volts or thereabouts presumably.
I was a bit sceptical about the life of the batteries so I left
them on 24 hrs to check, I was surprised that they lasted over a
month which makes them very useful as party lights in a damp
conservatory, let's say running them for 4hrs a day which should
make them last for 6 months perhaps.

Hang on a minute, many LEDs won't turn on at 3v as that's quite a
low Vf. Not only that but the normal arrangement for multiple LEDs
is to put them in series chains. You could put them in parallel but
there are issues. I suspect then that the 3V from your batteries is
steped up to something rather higher.

LEDs require a current source rather than a voltage and there are
some nice little ICs available which do all of the step up and
current control in one package, you may well have one of these in
the lights you've bought. Have a look near the batteries, if
they're cheapo it may be a small black blob rather than a chip as such.
The leds are wired in parallell and I have looked, and checked again
after your email for a black blob or any indication of some kind of
electronic gizmo that might control current or whatever, I am pretty
sure there is nothing there. the leds have transparent plastic
housings and I can see nothing in the housings that might indicate
that they have anything significant in there either.

I'm still struggling to imagine how these are arranged. If they are
in parallel then each LED will have a wire going to it so that's 10
wires running along the "string" plus the return which can be common
which makes 11 in all. Each LED will require a current limiting
resistance of some sort. They will not share a limiting resistor as a
mis-match between Vf of individual LEDs leads to runaway conditions
and dead LEDs so there will be 10 resistors. This is however a poor
way to design battery powered LED lights as If will vary significantly
with battery terminal voltage, bear in mind that a fully charged AA
cell may be anywhere between 1.2V and 1.5V depending on its chemistry.

It's just never done this way, in fact the only case where I've seen
LEDs controlled with a resistance commercially is in keyring lights
where the relatively high internal resistance of two button cells is
relied on to limit the current through a single LED. Your chain
simply can't be using the internal resistance of AA batteries as that
can be very low in some cases. In fact as a fun experiment try
fitting NiCd or NiMh cells, either it won't light because they have a
lower V or the LEDs will pop because they have a very low R.
The leds have a thin cable, transparent insulation, and having two
conductors in it; each conductor is connected to a leg of the 1st led
and then carries on to the next led, so in effect they are in parallel.
I found the invoice, and have a ref to the advert on ebay, you can see
on the pic how they are wired.

http://cgi.ebay.co.uk/10-LED-BATTERY...em3ca7 9c497d
I tried using my own camera but it wouldn't focus closely enough to
show the detail of the led connection.

I got to thinking about dimming them, a series resistor would
surely do that, but it would reduce the battery life by warming up
the resistor.

If the lights use an IC to do step up and control then no, a
resistor won't work, or at least not in the way you think.
Even for simple resistor in series with LED circuits you wouldn't
reduce the battery life by increasing the resistor value. You would
waste power of course by heating the resistor but the overall
current draw would be less so the battery would last longer.

Now you can dim filament bulbs by using an astable multivibrator
and varying the mark to space ratio, relying on the chopping
frequency and the time constant of the filament to provide a non
flickering and dimmable light.

Again anything like this will screw up an IC controller, if that's
what you've got.
I don't think there is one, it's just a simple translucent battery
holder which includes a switch, no sign of a "black blob" there
either. The possible explanation that occured to me is that they
have a constant current diode built into the leds during the
manufacturing process and is visually quite undetectable.

Doing that to an led would cause flickering of the light but would
the relativly slow response of the eye smooth out the flicker? as
in film movies which I think have to be in excess of 25 fps in
order to fool the eye.

Movies are 24Hz frame rate *but* they are double or treble shuttered
to avoid flicker. 24Hz is objectionable and unwatchable, 48/50Hz is
right on the limit and anything you'll see in a UK cinema will reach
your eye at 72Hz.
OK, thanks, I hadn't realised that, I was thinking to set the
multivib to 1KHz at least.

In the interests of battery life, are there any dimmers of this
type available for leds?

I purchased them on ebay and also bought a string of 20 and the only
discernable difference between them is that the 20 string has 3 AA
batteries in series.
I eventually cut off the battery housings and connected the two
strings in parallell powered by a Nokia mains charger ACP-7X which
were being sold off cheaply, if I remember correctly for 99p, the
charger got a little warm after 10mins use so I inserted a 2ohm
power resistor in series which reduced the current sufficiently to
keep the charger temperature within bounds. The subsequent light
output was reduced obviously but still adequate for my purposes.
This was just an experiment as I did not want to introduce mains
voltages to the still drying out conservatory (damp).

What voltage is the mains charger? If you have access to a multimeter
why not measure the current.
I think I did at the time, but they are disconnected and back on the
battery packs now. I think it was (from memory) approx 350 mA (30 leds)
and the voltage was just over 3volts. Power supply rated at 3.7 Volts,
355mA



Cheers
Don

In message , Calvin Sambrook
writes
I'm still struggling to imagine how these are arranged. If they are in
parallel then each LED will have a wire going to it so that's 10 wires
running along the "string" plus the return which can be common which
makes 11 in all.
Don't be daft. It's two wires running from the battery holder.

Think simple, the guys that make these things just use one series
resistor for all the LEDs, it's not good practice but it does work after
a fashion.
Each LED will require a current limiting resistance of some sort.
If you want to do it right, yes they do. If you want to sell the lights
for a quid in the local pound shop (where I got mine) then they don't.
They will not share a limiting resistor as a mis-match between Vf of
individual LEDs leads to runaway conditions and dead LEDs so there will
be 10 resistors.
In practice, I bet the mismatch between LEDs of the same batch is tiny.

It's just never done this way,
It is. As you say, it's not the right way but it is done.

--
Clint Sharp

"Clint Sharp" wrote in message
...
In message , Calvin Sambrook
writes
I'm still struggling to imagine how these are arranged. If they are in
parallel then each LED will have a wire going to it so that's 10 wires
running along the "string" plus the return which can be common which makes
11 in all.
Don't be daft. It's two wires running from the battery holder.

Think simple, the guys that make these things just use one series resistor
for all the LEDs, it's not good practice but it does work after a fashion.
Each LED will require a current limiting resistance of some sort.
If you want to do it right, yes they do. If you want to sell the lights
for a quid in the local pound shop (where I got mine) then they don't.
They will not share a limiting resistor as a mis-match between Vf of
individual LEDs leads to runaway conditions and dead LEDs so there will be
10 resistors.
In practice, I bet the mismatch between LEDs of the same batch is tiny.

It's just never done this way,
It is. As you say, it's not the right way but it is done.


Uk, I must remember to get out into the real world occasionally (apparently
the Real World is where the pizza delivery man comes from). I fear you are
right having seen the ebay item referred to.

Nasty but cheap.

On Mon, 23 Nov 2009 10:16:57 +0000, Donwill wrote:

mick wrote:
On Sun, 22 Nov 2009 11:03:02 +0000, Donwill wrote:


I purchased a string of 10 Leds powered by 2 AA batteries connected in
series giving 3 volts or thereabouts presumably. I was a bit
sceptical about the life of the batteries so I left them on 24 hrs to
check, I was surprised that they lasted over a month which makes them
very useful as party lights in a damp conservatory, let's say running
them for 4hrs a day which should make them last for 6 months perhaps.

I got to thinking about dimming them, a series resistor would surely
do that, but it would reduce the battery life by warming up the
resistor. Now you can dim filament bulbs by using an astable
multivibrator and varying the mark to space ratio, relying on the
chopping frequency and the time constant of the filament to provide a
non flickering and dimmable light.
Doing that to an led would cause flickering of the light but would the
relativly slow response of the eye smooth out the flicker? as in film
movies which I think have to be in excess of 25 fps in order to fool
the eye.
In the interests of battery life, are there any dimmers of this type
available for leds?




This gets to be fun if you are restricting your supply to 3v. Here's a
dimmer for 12v: http://www.reuk.co.uk/LED-Dimmer-Circuit.htm You may
be able to get it to work using a CMOS 555 (which works down to about
1.5v) and, maybe, a MOSFET such as 2N7000 instead of the transistor to
give better efficiency at the low voltage. You could decrease C1 to
increase the frequency if you get any flickering. R2 can be omitted if
a MOSFET is used.


Found this also:
http://www.opensourcepartners.nl/~costar/leddimmer/
which might be interesting to play with. Cheers



Yeah - I saw that one too. I decided that
a) it was probably a bit of overkill (you don't really need 0-100%).
b) it might be difficult (if not impossible) to get working from 3v.

I also considered using a PIC chip as those will work down to 2v. It
would be possible to use software PWM to drive a LED output MOSFET and an
analogue input from a "brightness" control. The problem is efficiency -
it probably isn't very... ;-) One of the little 8-leg chips like 12F683
would be the animal. This approach has one advantage over the others -
you can make the brightness control linear (to all intents and purposes),
no matter how the LEDs behave. You can also do silly things. :-)

--
Mick (Working in a M$-free zone!)
Web: http://www.nascom.info
Filtering everything posted from googlegroups to kill spam.

In message , mick
writes
I also considered using a PIC chip as those will work down to 2v. It
would be possible to use software PWM to drive a LED output MOSFET and an
analogue input from a "brightness" control. The problem is efficiency -
it probably isn't very... ;-) One of the little 8-leg chips like 12F683
would be the animal.
Why use software PWM when that chip has it in hardware? Trivial to set
up and control with a pot. I reckon you could buy the chip and a capable
transistor cheaper than the equivalent non micro components needed to do
PWM.
--
Clint Sharp

In message , Calvin Sambrook
writes

Uk, I must remember to get out into the real world occasionally
(apparently the Real World is where the pizza delivery man comes from).

I fear you are right having seen the ebay item referred to.
Oh I'm right, I bought a few sets because I can't buy LEDs that cheap.

Nasty but cheap.

Definitely.
--
Clint Sharp

On Wed, 25 Nov 2009 22:19:02 +0000, Clint Sharp wrote:

snip
Why use software PWM when that chip has it in hardware? Trivial to set
up and control with a pot. I reckon you could buy the chip and a capable
transistor cheaper than the equivalent non micro components needed to do
PWM.


Oh yeah... so it has! I should have checked the spec sheet first,
shouldn't I? :-)

I might have a go at building this just for fun!

--
Mick (Working in a M$-free zone!)
Web: http://www.nascom.info
Filtering everything posted from googlegroups to kill spam.

In message , mick
writes
On Wed, 25 Nov 2009 22:19:02 +0000, Clint Sharp wrote:

snip
Why use software PWM when that chip has it in hardware? Trivial to set
up and control with a pot. I reckon you could buy the chip and a capable
transistor cheaper than the equivalent non micro components needed to do
PWM.


Oh yeah... so it has! I should have checked the spec sheet first,
shouldn't I? :-)

I might have a go at building this just for fun!

Did it last night, works well.

Using the internal clock and an external transistor (BC639, first one I
found) it's very simple to code. I reckon the cost of parts is well
under 2 quid and I built it into the battery holder.
--
Clint Sharp

replying to Donwill, tahrey wrote:
That's basically the way most LEDs are dimmed in the first place. And
generally unless you're extremely sensitive, flickering at mains rate (which,
through a bridge rectifier, is essentially 100Hz, thus equal to all but the
most excessively high-refresh CRT monitors of old) should be imperceptible
when the diodes are stationary or only moving slowly relative to the eye. I
was in Homebase earlier looking at some light fittings and only realised that
one stand was displaying a mix of halogen and LED when I went to take a
picture with my phone and saw some of the lights strobing very gently on the
screen. Having been regularly annoyed by low-refresh CRTs back in the day, I
tried to determine whether there was any flicker visible to direct viewing,
but couldn't make any out even with the corner of my eye.

So if you were to use a 555 or other simple vibrator, so long as you tuned it
to operate at least somewhere into three figures that would be more than
enough to exclude flicker. If it's not visible with high brightness picture
tubes of 15 inches diagonal or more, it won't be visible with small, dim LEDs.
My phone itself has an OLED screen which probably refreshes somewhere around
120Hz (it's certainly not 60Hz) and the only evidence of flicker is if I wave
it around rapidly in pitch darkness. Most of the time you completely forget
it's not a regular LCD with a steady backlight.

I would question why you want to do this, though. I ran similar sets of lights
over christmas, and found them neither to be too bright for comfort, nor to
have such short battery lives or high voltage demands that running them off
"spare" batteries (non-rechargeables that had run too low to work in at least
one other more demanding device; freshly topped-off rechargeables would have
run a good bit longer) became an annoying performance of repeatedly switching
them out. Are they really too intense, or is changing the batteries every 3~4
weeks (which for christmas lights is essentially once a year) that much of a
hardship that building a dimming circuit is worth the hassle?

FWIW, if the problem is that they produce too intensely focussed little points
of light, I found that with a bargain bucket chain of multicoloured diodes,
which I think were initially supposed to come with star-shaped diffusers on
the end but turned out to lack them. A bit of pocket-money modification
turning them into a lantern string, by buying a pack of white plastic party
cups from Poundland, cutting small holes in the bottom of each cup and poking
a diode through each, proved to have a rather pleasing end result, with the
light being nicely diffused around the entire cone of the cup and giving the
whole string a softer, more traditional incandescent lantern type appearance.

Another mod that may be worth trying, if you have spare battery holders
sitting around, is to just run additional pairs of AAs in parallel, or hooking
the switch up to a pair of Cs or Ds instead. The overall bulk and weight of
the contraption still won't be particularly massive. In extremis a 3V mains
transformer, or a Poundland USB charger with a suitable dropper resistor (or
one charger wired to run two chains in series, though I'm not sure if they
produce much light with only 2.5v applied, my own tests suggest 2.6v is a bare
minimum; alternatively omitting the resistor and seeing if they stand up to
being overdriven to higher brightness, as they're probably underdriven by
design anyway) could be wired up to power it without any batteries, or even a
butchered cheapo solar-charging garden light or two for "free" non-mains juice.

--
for full context, visit https://www.homeownershub.com/uk-diy...es-597246-.htm

On Wednesday, 23 January 2019 13:14:11 UTC, tahrey wrote:
replying to Donwill, tahrey wrote:
That's basically the way most LEDs are dimmed in the first place. And
generally unless you're extremely sensitive, flickering at mains rate (which,
through a bridge rectifier, is essentially 100Hz, thus equal to all but the
most excessively high-refresh CRT monitors of old) should be imperceptible
when the diodes are stationary or only moving slowly relative to the eye. I
was in Homebase earlier looking at some light fittings and only realised that
one stand was displaying a mix of halogen and LED when I went to take a
picture with my phone and saw some of the lights strobing very gently on the
screen. Having been regularly annoyed by low-refresh CRTs back in the day, I
tried to determine whether there was any flicker visible to direct viewing,
but couldn't make any out even with the corner of my eye.

So if you were to use a 555 or other simple vibrator, so long as you tuned it
to operate at least somewhere into three figures that would be more than
enough to exclude flicker. If it's not visible with high brightness picture
tubes of 15 inches diagonal or more, it won't be visible with small, dim LEDs.
My phone itself has an OLED screen which probably refreshes somewhere around
120Hz (it's certainly not 60Hz) and the only evidence of flicker is if I wave
it around rapidly in pitch darkness. Most of the time you completely forget
it's not a regular LCD with a steady backlight.

I would question why you want to do this, though. I ran similar sets of lights
over christmas, and found them neither to be too bright for comfort, nor to
have such short battery lives or high voltage demands that running them off
"spare" batteries (non-rechargeables that had run too low to work in at least
one other more demanding device; freshly topped-off rechargeables would have
run a good bit longer) became an annoying performance of repeatedly switching
them out. Are they really too intense, or is changing the batteries every 3~4
weeks (which for christmas lights is essentially once a year) that much of a
hardship that building a dimming circuit is worth the hassle?

FWIW, if the problem is that they produce too intensely focussed little points
of light, I found that with a bargain bucket chain of multicoloured diodes,
which I think were initially supposed to come with star-shaped diffusers on
the end but turned out to lack them. A bit of pocket-money modification
turning them into a lantern string, by buying a pack of white plastic party
cups from Poundland, cutting small holes in the bottom of each cup and poking
a diode through each, proved to have a rather pleasing end result, with the
light being nicely diffused around the entire cone of the cup and giving the
whole string a softer, more traditional incandescent lantern type appearance.

Another mod that may be worth trying, if you have spare battery holders
sitting around, is to just run additional pairs of AAs in parallel, or hooking
the switch up to a pair of Cs or Ds instead. The overall bulk and weight of
the contraption still won't be particularly massive. In extremis a 3V mains
transformer, or a Poundland USB charger with a suitable dropper resistor (or
one charger wired to run two chains in series, though I'm not sure if they
produce much light with only 2.5v applied, my own tests suggest 2.6v is a bare
minimum; alternatively omitting the resistor and seeing if they stand up to
being overdriven to higher brightness, as they're probably underdriven by
design anyway) could be wired up to power it without any batteries, or even a
butchered cheapo solar-charging garden light or two for "free" non-mains juice.

You're replying to a post from 2009. Do yourself a favour and use a sane interface to this newsgroup rather than that semifunctional website. Google groups works. Here is news:uk.d-i-y


NT