I've got a low voltage PWM unit which is controlled by a DC voltage. Can
be used for motor or light control etc.

In this case I want to use it to drive high power LEDs. Which it does just
fine.

What I want to do is have it ramp up to max when a switch is made, and
ramp down to minimum when it's opened. I've experimented with RC circuits
which do sort of work - but don't give a linear fade in and fade out.
Any clever ideas? Preferably using just one cheap IC that I'll have in my
box...;-)

It would also be nice to arrange things so it takes no current when the
switch is off. Although I can get round this.

--
*Why are a wise man and a wise guy opposites?

Dave Plowman London SW
To e-mail, change noise into sound.

On Sep 11, 2:32*pm, "Dave Plowman (News)"
wrote:
I've got a low voltage PWM unit which is controlled by a DC voltage. Can
be used for motor or light control etc.

In this case I want to use it to drive high power LEDs. Which it does just
fine. *

What I want to do is have it ramp up to max when a switch is made, and
ramp down to minimum when it's opened. I've experimented with RC circuits
which do sort of work - but don't give a linear fade in and fade out.
Any clever ideas? Preferably using just one cheap IC that I'll have in my
box...;-)

It would also be nice to arrange things so it takes no current when the
switch is off. Although I can get round this.

There are fairly simple methods using opamps. Eg set one to deliver a
constant current, feed this to a cap on the output. Or set one to
deliver a fixed b drop across an output R, which gives constant i,
etc.

Or more basically, stay with a CR and just use a percentage of the v
scale, making it somewhat more linear.


NT

In article , Dave Plowman (News)
writes
I've got a low voltage PWM unit which is controlled by a DC voltage. Can
be used for motor or light control etc.

In this case I want to use it to drive high power LEDs. Which it does just
fine.

What I want to do is have it ramp up to max when a switch is made, and
ramp down to minimum when it's opened. I've experimented with RC circuits
which do sort of work - but don't give a linear fade in and fade out.
Any clever ideas? Preferably using just one cheap IC that I'll have in my
box...;-)

It would also be nice to arrange things so it takes no current when the
switch is off. Although I can get round this.

An integrator is what you're looking for, here's an op-amp based
circuit:

(Fixed width font reqd)
___ ___
| |
| |
| |R1 | | R3 || C1
| | | | ------||---------
| | | | | || |
|_| |_| | |
| | | |\ |
| _______ | | | \ |
----_______|--|-----|-----|- \ |
| | | \_____|________ Vcontrol
| R2 | | /
| |-----------|+ /
| | | /
| | |/
PB1 | | | R4
| o--| | |
--| | |
| o--| |_|
| |
| |
_|_ _|_


The ramp time delay is R2*C1 seconds.

If you chose a single supply fet input op amp then R3,R4 can be 1M, R1
roughly R2/10 to avoid skewing the delay.

Off state current drain should be in the uA range.
--
fred
BBC3, ITV2/3/4, channels going to the DOGs

Actually, for your non critical app, this may well do the job, values
left as an exercise for the reader ;-)

___ ___
| |
| |
| |R1 |
| | |
| | |
|_| |
| | / TR1 (pnp)
| _______ |
----_______|--|--------|
| | |\
| R2 | | \
| | || C1 |
| ----||------|
| || |___________ Vcontrol
PB1 | |
| o--| | |
--| | | R3
| o--| | |
| |_|
| |
_|_ _|_



--
fred
BBC3, ITV2/3/4, channels going to the DOGs

In article ,
fred wrote:
In article , Dave Plowman (News)
writes
I've got a low voltage PWM unit which is controlled by a DC voltage. Can
be used for motor or light control etc.

In this case I want to use it to drive high power LEDs. Which it does just
fine.

What I want to do is have it ramp up to max when a switch is made, and
ramp down to minimum when it's opened. I've experimented with RC circuits
which do sort of work - but don't give a linear fade in and fade out.
Any clever ideas? Preferably using just one cheap IC that I'll have in my
box...;-)

It would also be nice to arrange things so it takes no current when the
switch is off. Although I can get round this.

An integrator is what you're looking for, here's an op-amp based
circuit:

(Fixed width font reqd)
___ ___
| |
| |
| |R1 | | R3 || C1
| | | | ------||---------
| | | | | || |
|_| |_| | |
| | | |\ |
| _______ | | | \ |
----_______|--|-----|-----|- \ |
| | | \_____|________ Vcontrol
| R2 | | /
| |-----------|+ /
| | | /
| | |/
PB1 | | | R4
| o--| | |
--| | |
| o--| |_|
| |
| |
_|_ _|_


The ramp time delay is R2*C1 seconds.

If you chose a single supply fet input op amp then R3,R4 can be 1M, R1
roughly R2/10 to avoid skewing the delay.

Off state current drain should be in the uA range.

Thanks Fred- I'll give that a try.

--
*Eat well, stay fit, die anyway

Dave Plowman London SW
To e-mail, change noise into sound.

On Sep 11, 6:17*pm, "Dave Plowman (News)"
wrote:
In article ,
* *fred wrote:
In article , Dave Plowman (News)
writes
I've got a low voltage PWM unit which is controlled by a DC voltage. Can
be used for motor or light control etc.

In this case I want to use it to drive high power LEDs. Which it does just
fine. *

What I want to do is have it ramp up to max when a switch is made, and
ramp down to minimum when it's opened. I've experimented with RC circuits
which do sort of work - but don't give a linear fade in and fade out.
Any clever ideas? Preferably using just one cheap IC that I'll have in my
box...;-)

It would also be nice to arrange things so it takes no current when the
switch is off. Although I can get round this.

An integrator is what you're looking for, here's an op-amp based
circuit:
(Fixed width font reqd)
* * * * ___ * * * * * ___
* * * * *| * * * * * * |
* * * * *| * * * * * * |
* * * * | |R1 * * * * | | R3 * * * || C1
* * * * | | * * * * * | | * *------||---------
* * * * | | * * * * * | | * *| * * || * * * *|
* * * * |_| * * * * * |_| * *| * * * * * * * |
* * * * *| * * * * * * | * * | * * |\ * * * *|
* * * * *| * _______ * | * * | * * | \ * * * |
* * * * *----_______|--|-----|-----|- \ * * *|
* * * * *| * * * * * * | * * * * * | * \_____|________ Vcontrol
* * * * *| * *R2 * * * | * * * * * | * /
* * * * *| * * * * * * |-----------|+ /
* * * * *| * * * * * * | * * * * * | /
* * * * *| * * * * * * | * * * * * |/
* *PB1 * | * * * * * *| | R4
* * | o--| * * * * * *| |
* --| * * * * * * * * | |
* * | o--| * * * * * *|_|
* * * * *| * * * * * * |
* * * * *| * * * * * * |
* * * * _|_ * * * * * _|_
The ramp time delay is R2*C1 seconds.
If you chose a single supply fet input op amp then R3,R4 can be 1M, R1
roughly R2/10 to avoid skewing the delay.
Off state current drain should be in the uA range.

Thanks Fred- I'll give that a try.


Even simpler is a fet and resistor configured to deliver constant
current, plus a capacitor.


NT

"Dave Plowman (News)" writes:

I've got a low voltage PWM unit which is controlled by a DC voltage. Can
be used for motor or light control etc.

In this case I want to use it to drive high power LEDs. Which it does just
fine.

What I want to do is have it ramp up to max when a switch is made, and
ramp down to minimum when it's opened. I've experimented with RC circuits
which do sort of work - but don't give a linear fade in and fade out.

The solutions posted give you a linear voltage ramp, but I'd have
thought that what you want is something that /looks/ linear given the
eye's logarithmic response. But then, an RC circuit would give an
exponential decay, so I'm confused.

--
Jón Fairbairn
http://www.chaos.org.uk/~jf/Stuff-I-dont-want.html (updated 2009-01-31)

In article
,
NT wrote:

Even simpler is a fet and resistor configured to deliver constant
current, plus a capacitor.

I tried something similar using an LM318 in 'soft start' configuration
which gives a nice ramp in - but would need some sort of constant current
load to give the same ramp out.


NT

--
*Microsoft broke Volkswagen's record: They only made 21.4 million bugs.

Dave Plowman London SW
To e-mail, change noise into sound.

In article ,
Jon Fairbairn wrote:
What I want to do is have it ramp up to max when a switch is made, and
ramp down to minimum when it's opened. I've experimented with RC circuits
which do sort of work - but don't give a linear fade in and fade out.

The solutions posted give you a linear voltage ramp, but I'd have
thought that what you want is something that /looks/ linear given the
eye's logarithmic response. But then, an RC circuit would give an
exponential decay, so I'm confused.

Reasonably linear to the eye is what I want. An RC circuit on discharge
gives a steep fade at first followed by a long slow one. As it were.

--
*Parenthetical remarks (however relevant) are (usually) unnecessary *

Dave Plowman London SW
To e-mail, change noise into sound.

On Sep 12, 9:34*am, "Dave Plowman (News)"
wrote:
In article
,
* *NT wrote:

Even simpler is a fet and resistor configured to deliver constant
current, plus a capacitor.

*I tried something similar using an LM318 in 'soft start' configuration
which gives a nice ramp in - but would need some sort of constant current
load to give the same ramp out. * * *

for which an opamp delivering constant current is ideal


NT

On Sep 12, 11:38*am, NT wrote:
On Sep 12, 9:34*am, "Dave Plowman (News)"
wrote:

In article
,
* *NT wrote:

Even simpler is a fet and resistor configured to deliver constant
current, plus a capacitor.

*I tried something similar using an LM318 in 'soft start' configuration
which gives a nice ramp in - but would need some sort of constant current
load to give the same ramp out. * * *

for which an opamp delivering constant current is ideal

NT

and the LM324 possibly a good candidate. There's a P version if you
need to save power.


NT

In article ,
fred wrote:
The ramp time delay is R2*C1 seconds.

Not quite sure what this is, mathematically speaking. Is the * a symbol
which has come out wrong on my browser?

Basically, I'm after a fairly slow fade - perhaps 3 seconds from min to
max.

--
*All generalizations are false.

Dave Plowman London SW
To e-mail, change noise into sound.

On Sep 13, 1:40*pm, "Dave Plowman (News)"
wrote:
In article ,
* *fred wrote:

The ramp time delay is R2*C1 seconds.

Not quite sure what this is, mathematically speaking. Is the * a symbol
which has come out wrong on my browser?

Basically, I'm after a fairly slow fade - perhaps 3 seconds from min to
max.

time constant is R times C


NT

In article
, NT
writes
On Sep 13, 1:40*pm, "Dave Plowman (News)"
wrote:
In article ,
* *fred wrote:

The ramp time delay is R2*C1 seconds.

Not quite sure what this is, mathematically speaking. Is the * a symbol
which has come out wrong on my browser?

Basically, I'm after a fairly slow fade - perhaps 3 seconds from min to
max.

time constant is R times C

Yep, times it is, most of the packages I use denote multiplication with
an asterisk so I techno spake.

For 3s you could kick off with R1 = 10k, R3 = R4 = 100k, R2 = 330k, C1 =
10uF non polar or 2 x 22uF polar back to back to create a non polar cap.
--
fred
BBC3, ITV2/3/4, channels going to the DOGs

In article ,
fred wrote:
In article
, NT
writes
On Sep 13, 1:40 pm, "Dave Plowman (News)"
wrote:
In article ,
fred wrote:

The ramp time delay is R2*C1 seconds.

Not quite sure what this is, mathematically speaking. Is the * a
symbol which has come out wrong on my browser?

Basically, I'm after a fairly slow fade - perhaps 3 seconds from min
to max.

time constant is R times C

Yep, times it is, most of the packages I use denote multiplication with
an asterisk so I techno spake.

Ah. Right. I've never seen it written like that.

For 3s you could kick off with R1 = 10k, R3 = R4 = 100k, R2 = 330k, C1 =
10uF non polar or 2 x 22uF polar back to back to create a non polar cap.

I'll give that a go - the largest non polarized I had was 1µF and space
only allowed two, so perhaps not an ideal ratio. Never thought of using
electrolytics back to back. Although I do have some non polar ones.

--
*Oh, what a tangled website we weave when first we practice *

Dave Plowman London SW
To e-mail, change noise into sound.