Resistance controlled voltage source (for led array

[whatchitfoool]

hi all,

I'm looking for a cheap (and by extension simple) circuit.

i have a special pot (Piher's 10mm hex PT-10) that can't be changed. the pots can be physically changed but i have to use this type. the series has a 150mw power rating, with resistances of : 100,150,200,250,500,1k,2k,2.5k,5k,10k,20k,25k,50k,100k,200k,500k.

i have a 3.3v source (tight), with 750-1000ma available to play with. i want to drive 15-30 LEDs in parallel, each pulling 19ma. so a total current of ≈ 375ma - 600-650 would be nice; but i could live with a cap of 500ma. i need to be able to take them from (as close to) off to (as close to) full brightness as i get get, via control of the Phier pot. the Phier pots can rotate about 250°, and i can use about 180° (about 75% of its resistive range).

full brightness would be the LEDS seeing 3.1-3.3v

does anyone have any sugestions as to what setup to use to supply the current i need?
id take ideas above 350ma.

EDIT: forgot to mention id like it to in/decrease linearly, like one would see with voltage dividing pot (that was of cource non-logarithmic in style).

 

[MisterX]

Do you have to use a potentiometer with a logarithmic scale? It seems from the data sheet that one may order linear ones.

 

[whatchitfoool]

no, but i could put in either. right now i have some regular linear ones, as that is how i would like to have the LEDs change.

 

[whatchitfoool]

It won’t let me edit my first post...
I made an error when posting some specs. Each led is drawing 6ma. NOT 19ma. 8 leds in parallel pull about 46ma, and at 3.3 volts that takes me to 150mw.
That changes the total current draw to between 100-150ma max. This opens up using an op-amp or such.
Before I realized my error, I was thinking about using a KA317U variable voltage regulator, but I don't know if it will like an input voltage of 3.3v and work how I would like to.

I’m open to another set up eyes looking over me. If someone can see the datasheet on the ka317u and give me there take on it id would be great. Especially their opinion on setting the Vref

 

[MisterX]

I think a linear adjustment of current would produce something closer to a linear adjustment of brightness than a linear adjust of voltage.

 

[vk6kro]

The 3.3 volt power source is really not suitable to use for this.
White LEDs require about this voltage to operate and may need up to 3.5 volts to produce reasonable brightness. LEDs brightness is widely variable with input voltage and it is normal to use a higher voltage and a series resistor to control the current.
So, this doesn't give a good result anyway, but if you want to control the brightness, this will take up more voltage.
You could use a different coloured LED as other LEDs need less voltage to operate.

There is a good chart of LED voltages for various coloured LEDs on Wikipedia:
http://en.wikipedia.org/wiki/LED

You could either get a higher voltage power supply or produce a higher voltage from the 3.3 volts. The first option is much better, but it is possible to get a higher voltage from the 3.3 volt supply.
You could use something like this voltage booster:
http://www.ebay.com.au/itm/DC-DC-Step-Up-Boost-Adjustable-Power-Supply-Module-/120757797729?pt=AU_B_I_Electrical_Test_Equipment&hash=item1c1db9c361
This one works down to 3 volts input and you can get up to 24 volts out. Costs about $5.

Assuming you have a 10 volt or 20 volt supply, it then becomes possible to use a number of LEDs in series with a current limiting resistor and switch the combination with a power FET or transistor.

Controlling the LEDs could be done with pulse width modulation.
You send pulses to the transistor or FET controlling the LED current and the brightness depends on the time the LED is on vs the time that it is off.
There are simple circuits that provide these pulses.
Here is one: