Driving red/green LED from high/low/tri-state output

I think I know how to do this... just tie both ends of the LEDs together (making a "2 pin" bi-color LED), drive one end from the status pin, the other from a voltage divide formed by 2 resistors, nominally the same values for the same current for both red and green, or somewhat different values to vary it.

Two more resistors and two transistors?

+v | V led +v - | | | in | _R_ |<

--o------------|___|-| | | |\\ | .-. | | | |R .-. | | | | |R | '-' | | | | '-' | _R_ |/ | -|___|-| V led |> - | | GND-------------

(created by AACircuit v1.28 beta 10/06/04

Alternate idea: Connect a resistor from the status line to each of gnd and vcc. Use two comparators to compare this three-voltage state with

1/3vcc and 2/3vcc. This gives you two comparator outputs to work with.

I'm doing this in my furnace controller, except in my case I'm using it as a tri-state booster.

I've tried this before. Base leakage current is sufficient to keep both transistors on when tri-stated. I've even tried putting resistors between the bases, like your drawing. Didn't seem to help; high enough values to stop the leakage were too high to let the signal activate the transistors.

Oops. Duh...

Put your Leds anode to cathode, one end of the pair to the IC and the other to the middle of a voltage divider made by two 500R resistors. It wastes power all the time (5mA@5V) but then so do the LEDs. ;-)

Hi, Joel. Assuming both the green and red LEDs have a forward voltage drop of around 1.7V, and also assuming your tri-state output can source or sink 8mA, you should be able to get what you want with this fairly inelegant but workable solution (view in fixed font or M$ Notepad):

| | VCC | + | | | G V ~ | - ~ | | | D V | - | | | D V | - | |\\ ___ | | -| >O--|___|-o | |/ 270 | | D V | - |D = 1N4148 | | D V | - | | | R V ~ | - ~ | | | === | GND | (created by AACircuit v1.28.6 beta 04/19/05

The forward voltage drops of the two LEDs plus the four small signal diodes will be at least 5.8V. If your supply voltage is 5V, that will keep the LEDs from turning on when the output is tri-stated.

The 270 ohm resistor will lead to a little less than 8mA being driven through only one of the LEDs when the logic gate is on. That should be plenty for high intensity LEDs.

Make sure to look for red and green LEDs with similar typical forward voltages. If the red is still brighter than the green, you may want to place a small resistor in series with the red LED to balance.

Good luck Chris

you're so close to the answer....

.-------------------------------------------------------------. | This is an ascii schematic, if the diagram appears garbled | | try switching to a fixed-pitch font (courier works well) | | pasting it into notepad works well on ms-windows. | | or in google groups "view source" (found under options) | `-------------------------------------------------------------'

green LED in ---+--||---+--[RX]--- gnd red LED

--

Bye.
   Jasen

Thanks to everyone for their solutions; I've learned a lot. I am using the approach that Jasen suggested, and it's working well. I have power to burn here, but if I didn't, I think I'd go with Chris's solution. I'd have a hard time convincing myself to use a comparator, although that seems like the least-kludged approach.

I can see now why people sometimes use one of those 6 pin PICs or 8 pin AVRs to do little more than flash an LED...!

That used to work.

*-Reply-to-author-*-*-*-*+*.*.*.messages.by.this.author+Show.original+view.source "view source" is no longer available.

**Show original** still works.

One of the cheaper, at Digikey, is the PIC10F200. $1 in ones, half that in 25's. You could even set it up to handle the three states differently, on those two LEDs, say blinking the red when nothing is in the charger, setting it solid red when charging, and green when done. All with no additional parts if your voltage supply rail is in range, except possibly the two current limit resistors -- which is only a maybe, as the output drive may self-limit acceptably. Of course, you need the means to program the thing. So there is a big advantage to the discrete approach, if you don't have the tool setup in place. Much easier to make such choices, if you are ready to go, already.

Jon

Good point. For copmarison, ATTiny11's are $0.54/1, $0.41/25, $0.38/100 from DigiKey. I'd imagine that if you're buying, say, 1,000? 10,000? Microchip will happily compete with those prices, however. I also imagine that, in that sort of quantities, you could get your favorite distributor to pre-program the parts for you, which would be even better.

---Joel

Yes. I haven't used the ATTiny11 yet or the PIC10F200, so I cannot speak to either. I'd need to check the data sheets to see what differences may be useful to know about (packaging, oscillator, power, peripherals, etc.) But I note that the prices at 25-qty are quite similar. Close enough to warrant a close datasheet look, I suppose, if I get serious about buying one or the other for 'bubble gum' part purposes. (I usually keep two processor types in tubes, around here, for that and shift to new ones when the old stuff runs out.)

I can only hope my hobby interests ever rise to that level of purchase. ;)

Jon