My black box incorporates this ability for the prop to use leds in low resistance operation.
The leds rely on the sense ability of a very low (considered "dangerous") resistor to sense using the led wilist also using it as acitve led.
Forrest Mimms III showed this ability in the radio shack electronics book.
It works well when us use in bright environments, or provide a light source for leds to detect. (i have a picture using a parallax ccfl tube as light source too)
I am working on a new design that might not use resistors at all, much like you mention here. Internal resistace with dual pin in out, along with bandgap junction fall off rate sense, if all goes well, no led should ever pop or go into any kind of thermal runaway.
Even this can be measured and caught, using current sense INTO the prop itself, because the prop is doing all the current source sink work.
So you can have theoretical max current draw, use rctime on another pin to calculate total prop current draw wilist in operation, if any led (transistor) latches up, prop will catch it pretty quickly, and can turn off circuit. This would require adjustment of program to not allow too many leds to turn on at once, if you are running multiple leds. And don't forget rctime will use current also from the cap drain/charge.
Package max on prop is 200ma. 3.3v @ 200ma = 16.5 Ohms, clamp props supply voltage @ 16.5 ohms use rctime.
Also, I just took a look at the datasheet for the 160-1058-ND and it shows the current going down with temperature...
Where ?
The Light output decreases, but that is not the same as Vf
Osram says this
["The forward voltages of the LEDs have a negative temperature coefficient. (Tk = −3.7mV/K) "]
Infineon says this
["LEDs have a negative temperature coefficient for forward voltage as regular PN functions do - but frequently higher in magnitude (e.g. -4 mV)"]
Both the LED and the Prop have a series resistor component, which needs to be added for sensitivity checks.
That 4mV/'C over an appx 50ohms total will add 800uA to the LED for a 10'C rise
and Vf on a LED has a broad spec on the data sheets
Avago Vf specs
Color @i Typ. Max.
AlInGaP Red 20 mA 1.9 2.4
AlInGaP Amber 20 mA 1.9 2.4
AlInGaP Red[2] 10 mA 1.8 2.3
AlInGaP Amber[2] 10 mA 1.8 2.3
InGaN Green 10 mA 3.4 3.8
InGaN Blue 10 mA 3.4 3.8
HER 20 mA 2.1 2.6
GaP Orange 20 mA 2.2 2.6
GaP Yellow 20 mA 2.1 2.6
GaP Green 20 mA 2.2 2.6
Note they do not even give a Min Vf, - so you have to hope the LEDs you have, are much more average than max (or min).
For those on a Tape and Reel, you will be much better within a batch.
If you can vary Vcc, to fine-adjust LED drive, then 'Direct' LED connection (just the two internal resistors) is viable for simple systems, if you choose your LED Process first.
Given most LEDs are 'too bright' anyway, the InGaN give lower drains, direct connected.
Blue LEDs are no longer expensive, I see Digikey starts from 6c/4000 for Blue, and stock ones at 12c/2k
Comments
The leds rely on the sense ability of a very low (considered "dangerous") resistor to sense using the led wilist also using it as acitve led.
Forrest Mimms III showed this ability in the radio shack electronics book.
It works well when us use in bright environments, or provide a light source for leds to detect. (i have a picture using a parallax ccfl tube as light source too)
I am working on a new design that might not use resistors at all, much like you mention here. Internal resistace with dual pin in out, along with bandgap junction fall off rate sense, if all goes well, no led should ever pop or go into any kind of thermal runaway.
Even this can be measured and caught, using current sense INTO the prop itself, because the prop is doing all the current source sink work.
So you can have theoretical max current draw, use rctime on another pin to calculate total prop current draw wilist in operation, if any led (transistor) latches up, prop will catch it pretty quickly, and can turn off circuit. This would require adjustment of program to not allow too many leds to turn on at once, if you are running multiple leds. And don't forget rctime will use current also from the cap drain/charge.
Package max on prop is 200ma. 3.3v @ 200ma = 16.5 Ohms, clamp props supply voltage @ 16.5 ohms use rctime.
http://forums.parallax.com/showthread.php?115258-TheBlackBox-Release-v2.0-Propeller-HSS-FX-Sequencer-with-Digital-Audio-SPDIF
http://forums.parallax.com/showthread.php?115258-TheBlackBox-Release-v2.0-Propeller-HSS-FX-Sequencer-with-Digital-Audio-SPDIF
Where ?
The Light output decreases, but that is not the same as Vf
Osram says this
["The forward voltages of the LEDs have a negative temperature coefficient. (Tk = −3.7mV/K) "]
Infineon says this
["LEDs have a negative temperature coefficient for forward voltage as regular PN functions do - but frequently higher in magnitude (e.g. -4 mV)"]
Both the LED and the Prop have a series resistor component, which needs to be added for sensitivity checks.
That 4mV/'C over an appx 50ohms total will add 800uA to the LED for a 10'C rise
and Vf on a LED has a broad spec on the data sheets
Note they do not even give a Min Vf, - so you have to hope the LEDs you have, are much more average than max (or min).
For those on a Tape and Reel, you will be much better within a batch.
If you can vary Vcc, to fine-adjust LED drive, then 'Direct' LED connection (just the two internal resistors) is viable for simple systems, if you choose your LED Process first.
Given most LEDs are 'too bright' anyway, the InGaN give lower drains, direct connected.
Blue LEDs are no longer expensive, I see Digikey starts from 6c/4000 for Blue, and stock ones at 12c/2k