Blue LED at 3.3V

On a sunny day (Tue, 25 May 2010 19:40:25 -0700) it happened John Larkin wrote in :

I wanted to say: Just use resistor to heat up the LED. Perhaps a NTC could be used, not sure if it would be easy to fidn teh right one, and mount it next to the lED in parallel with the supply...

Now somebody could invent a LED with all that build in, photo sensor too, to keep light output at a progammable (I2C perhaps) level. Little switcher inside? Current limiter... This is the age of integration, logic level I2C input.

Tim Wescott schrieb:

Hello,

they mix the blue light with yellow light from the fluorescent material, but it is a very narrow blue band and a very broad yellow band. If you look at a spectrum diagram of the resulting light it looks very different to white light from the sun or from a ligth bulb with a glowing tungsten wire.

Bye

Some may indeed by bi-chromic but the ones that I've examined recently with a handheld spectroscope have a remarkably even spectrum from about

Hi,

Could you post his circuit on

one,

Nah, A one wire interface with parasitic power.

Maybe a TEC and controller to control the wavelength?

Do you really want LEDs with 80 page manuals (and 5 pages of errata)? ;-)

Best regards, Spehro Pefhany

I usually do multilayer boards with power and ground planes. A few bypass caps will do for the whole board, so I don't put one per chip like in olden days.

The average LED current is set by the R and C values, so you don't need another resistor.

The danger is that the drop of the diode+LED had better be less than the max supply voltage. So silicon might be better in some cases.

R and C have to be selected to get the desired LED current. There's a second invisible capacitor, Cin of the gate, that's involved too.

Hmmm, neither C may matter. It might all cancel out. Somebody should do the math.

John

how

The IC is a schmitt trigger. (And connected to ground also... connection not shown)

George H.

On a sunny day (Wed, 26 May 2010 08:07:44 -0400) it happened Spehro Pefhany wrote in :

one,

Oh well, one gets used to things :-)

one,

We don't yet use any parts with thousand-page datasheets, but we do have a couple over 800. And the longer the manual, the worse it seems to be.

John

. . .

Who better than the designer?

Besides, that circuit won't work as drawn.

Copy of a scheme by Bob Hirschfeld, at National, 40 years ago... called "The Miser"... ran on a 1.5V cell. ...Jim Thompson

Wonder why? Wonder how Larkin will excuse that away ?:-) ...Jim Thompson

-- | James E.Thompson, CTO | mens | | Analog Innovations, Inc. | et | | Analog/Mixed-Signal ASIC's and Discrete Systems | manus | | Phoenix, Arizona 85048 Skype: Contacts Only | | | Voice:(480)460-2350 Fax: Available upon request | Brass Rat | | E-mail Icon at

Wherever you got your 3.3V from, there's probably OTHER power available. Use that.

Regulated 3.3V power is unsuitable for this task. If that's blocking you, re-engineer it so it doesn't.

I have a day job. An idea isn't a design.

Why not? The schmitt needs to be an inverter to oscillate of course, but that should be obvious to anyone who posts to s.e.d. The point is that a clock, a cap, and a diode will boost 3.3 volts nicely, to drive a 3 volt LED. One could even eliminate the resistor!

Better than any ideas you or JT posted. C'mon, do us a 555 thing. Or something with 85 transistors.

John

Excuses! Excuses! Excuses!

Inverter or buffer, it can't work as drawn. Your day job certainly isn't circuit design :-) ...Jim Thompson

Thanks.

Theory of "operation" is left as an exercise for the student :-) ...Jim Thompson

The schmitt has to be an inverter to oscillate, of course. I left out the bubble.

RR could be a short, then it would oscillate faster!

John