5
\$\begingroup\$

I'm working with a commercially available circuit that can optionally be configured to disable its internal LEDs and use two external LEDs. This post discusses the use of external LEDs. One LED or the other may be activated periodically; simultaneous activation of both LEDs never occurs. Both external LEDs connect to a common 5V source, and the cathode of each connects to one of the circuit's two LED output pins. When not activated, the output pins are left floating. When a pin is activated, it pulls its connected cathode to ground. The output pins each accept a maximum voltage of 5V, a maximum sink current of 20mA and cannot source current. I think the circuit, configured to use external LEDs, looks something like the following. (The mfr. documentation calls the output pins "open collector outputs", so I've shown them that way. But I think this is questionable, because their operation is described as though input signal is not inverted. It could also be that my understanding of open collector outputs is incomplete.)

Schematic showing how LEDs are connected to the circuit.

If possible, I would like to replace the two LEDs with a single, 2-pin, bicolor LED (in effect, 2 LEDs connected in reverse parallel). The new circuit would behave as follows: when neither of the circuit's output pins are grounded, no power is supplied to the LED; when one of the pins is grounded, power is supplied to the LED with polarity oriented one way or the other, depending on which output pin is grounded. The power source is a battery whose voltage will range from 11V to 13.6V.

Additional requirement: Relays are not an option. In its default configuration (using internal LEDs), the circuit's current consumption is advertised as 0.125mA. If a solution that keeps power consumption similarly low isn't possible, I'll stay with two external LEDs, or I'll go with the internal LEDs and a light pipe if possible. As the enclosure might need to be opened by the user, an electrical connection to a panel LED is preferable. Also, zero (or as close to zero as possible) RF and conducted noise is a requirement.

Related threads: I think the answer in this thread might come close to what I need. If this solution could meet my requirements, I would need help with how to adapt it for two inputs and to allow both LEDs to be off. Answers in this thread also seem applicable, but again, I would need help in adapting a solution to my needs. This one contains a number of interesting solutions, but again, one logical input, not two.

\$\endgroup\$
  • \$\begingroup\$ does the replacement LED have to be a 2-pin device? \$\endgroup\$ – jsotola Sep 26 '18 at 21:56
  • \$\begingroup\$ The short of it is this: You want to use a single bicolor LED (3-pin is commonly found) with two open-collector outputs of a COTS device where it is known in advance that one or the other or neither LED is to be lit, but never both. That about it? \$\endgroup\$ – jonk Sep 26 '18 at 21:58
  • \$\begingroup\$ operation is described as though input signal is not inverted .... what is the manufacturer's description? \$\endgroup\$ – jsotola Sep 26 '18 at 21:58
  • 3
    \$\begingroup\$ then replace the two LEDs with a common anode RGB LED .... you will have a choice of any two of three colors .... you could even use a diode matrix to mix colors and have a choice of two of 7 colors ( instructables.com/id/… ) \$\endgroup\$ – jsotola Sep 26 '18 at 22:16
  • 1
    \$\begingroup\$ there is no mention of the input signal in that statement .... all it talks about is the output ...... they are saying that there is a connection to ground when the LED is on \$\endgroup\$ – jsotola Sep 27 '18 at 1:15
9
\$\begingroup\$

Given that the pins are open drain, the following circuit should work quite nicely:

schematic

simulate this circuit – Schematic created using CircuitLab

You'll get some wasted power as one of the resistors will be pulled straight to ground when an LED is turned on, however if you go for a low-current LED, you can increase the resistance to reduce the amount of wasted current.


Alternatively, the simplest option would be to use a common-anode 3-pin bi-colour LED. This would then be wired up exactly as the diagram in your post, with the common anode via resistor to +5V, and the two cathodes to your outputs.

\$\endgroup\$
  • \$\begingroup\$ Yeah. That's where my head was at! That includes your final comment, as well. Same thoughts, exactly. \$\endgroup\$ – jonk Sep 26 '18 at 22:02
  • \$\begingroup\$ Of course, a common-anode LED will work. I'm not an engineer and have little experience with electronics, so although I knew about common-cathode LEDs, I didn't know about common-anode LEDs. \$\endgroup\$ – dcorsello Sep 26 '18 at 22:30
  • \$\begingroup\$ It seems to me that when both inputs are high (both LEDs off), this circuit will deliver the full battery voltage to the inputs, which will exceed the stated maximum of 5V. \$\endgroup\$ – caf Sep 27 '18 at 0:19
  • \$\begingroup\$ @caf I was assuming from the earlier description that a 5V supply was available from the device. \$\endgroup\$ – Tom Carpenter Sep 27 '18 at 6:56
  • \$\begingroup\$ Sorry for being unclear about the power supply. During normal operation, system voltage will range between 11V and 14.4V, but under a particular error condition, it will be as high as 21V. Using input voltage = 21V, Vf = 2V and If = 20mV, I calculate a resistor value of 950 Ohms. Will this work? \$\endgroup\$ – dcorsello Oct 15 '18 at 23:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.