Re: LED impedance

In <pan.2006.04.04.22.42.29.975378@xxxxxxxxxxx>, Rich Grise wrote in part:
On Sat, 01 Apr 2006 18:16:20 -0800, John Larkin wrote:

On Sat, 1 Apr 2006 19:56:00 -0500, "D. G." <abc@xxxxxxx> wrote:

I want to add a red LED directly in series with a phone to show when it
is on use. The current is about 15mA and is not subject to change, it is
an internal phone system.
^^^^^^^^^^^^^^^^^^^^^^

But how about the LED impedance, is it low enough not to disturb the 600
ohm phone impedance?
Or a capacitor in parallel is absolutely necessary?


It won't bother the voice stuff, but it might affect the ring voltage,
and the AC ring voltage could certainly blow the LED. I'd drive it
through a bridge rectifier. It will get *very* bright during ring,
maybe, depending on the phone ringer load.

Maybe if we asked the OP if his internal system has ringer voltage
on it...

In answer to the OP's question, since you're in control of the power
supply, I don't see any particular problem with it - it shouldn't
even need a cap, since it's in a current loop. And yes, your
15 mA supply will need another 1.2V or so of compliance. :-)

A couple points:

1. Most red LEDs at 15 mA drop 1.6 volts (for "original formula"
GaAsP-on-GaAs) to 1.9-2.05 volts (for many InGaAlP types).
But then again, if the LED was in series with a phone on a phone line
where the impedance is often said to be 600 ohms and the open circuit
voltage (for a POTS line) is 40 or 48 volts or whatever, this should not
make a significant difference.

2. I just did some measurements with my POTS line.

On-hook: Voltage across phone 44V
Current through phone less than .1 mA

Off-hook: Voltage across phone 7.2V
Current through phone 22 mA

Phone DC resistance: 33 ohms (approx.)
Line DC impedance: 1670 ohms (approx.)

Voltage drop across red LEDs at 22 mA: Mostly ranges from 1.6 to 2.1
volts, depending mostly on chemistry. Some red and/or reddish orange LEDs
could be considered non-defective if dropping as high as 2.6 possibly 2.8
volts at 20 mA, although 2.3 is really high and 1.5 is really low.

In my experience with a bit of high current pulsing of 5 mm and 3
mm (T1-3/4 and T1) LEDs, I have seen two components to the voltage drop:

1. A resistance of mostly about 10-20 ohms

2. The diode drop - whose voltage drop changes by .1 volt or maybe
sometimes to maybe often some larger fraction of a volt per factor-of-10
change in current. In a few cases of imperfection, the diode drop may
be found to have resistance of a few megohms or 100K's of ohms in
parallel. Static-damaged GaN or InGaN LEDs could have parallel
resistance in or near the the 100-ohm - 1,000 ohm ballpark.

With voltage change unrelated to resistance unlikely to change by .1
volt or more as current changes from 10 to 20 mA (or vice versa), which is
a dynamic impedance of 10 ohms or less, plus resistance of 10-20 ohms, I
would guess that an red LED is typically effectively a 1.4-2 volt voltage
drop in series with a resistance anywhere from 15 to 30 ohms.

3. If you put an LED in series with a telephone on a POTS ine, consider
ringing conditions:

a) Line impedance may be the "traditional" 600 ohms and
I have yet to rule out less, with phone impedance possibly well below 330
ohms due to the ringer being contacted. Peak voltage is traditionally 80
volts, and that may be under load of who knows what impedance.
I did try myself, and peak current through a ringing on-hook phone
surprisingly to me appeared to be less than 22 mA in my particular case.
But I fear much more could be possible! I may want an LED in series with
a phone to be a "Luxeon" or similar one rated at least 350 mA just to be
safe!

b) I have known the peak-to-peak ring signal voltage to exceed twice the
DC on-hook voltage. An LED in series with a phone should have a reverse
diode across the LED. Some LEDs have damage mechanisms beyond merely
overheating if reverse breakdown occurs, with many GaN and InGaN ones
being especially intolerant of any reverse breakdown at all!
Also, wondering if the phone line has any intolerance of especially
unequal conductivity between one direction and the other - could be a good
reason to put an antiparallel diode (or LED) across any LED that is in
series with the phone!

- Don Klipstein (don@xxxxxxxxx)
.

Relevant Pages

  • Re: LEDs - Simple Wireing Questions
    ... Batteries have an internal resistance (that consumes some of the ... voltage as current passes through it and limits the short circuit ... The LED ratings of 3.2 volts and 20 ma implies that it takes ... However, LEDs are not resistors, and their current is very ...
    (sci.electronics.basics)
  • Re: how to control LED array? (follow-up)
    ... Fields also advise to take these 56 ohms base resistors? ... >> They will consume more current than the LEDs. ... The current sink is: ... pin of your micro (it's not uncommon to find output resistance slopes ...
    (sci.electronics.basics)
  • Re: OPA657 50Mhz photodiode preamplifier problems
    ... > I'm using AGC - reducing the bias voltage for the APD and also the local ... > oscilator drive voltage seems to have the least impact on phase, ... hmmm, but if you drive with 25 Ohms, and the receiver end is 50 Ohm, ... > output resistance is generaly equal to the Re of the transistor wich depends ...
    (sci.electronics.design)
  • Re: Sensor/LED circuit-more help needed please-multiVs and dimmer
    ... >> I am still working on this Sensor Circuit with LED indicators. ... >> appropriate voltage. ... >> So I have eight sensors, eight LEDs (some with different ratings, see ... > * You've got a sensor which turns sinks current to GND (low resistance ...
    (sci.electronics.basics)
  • Hysteresis needed in 741 Op-Amp circuit
    ... (determining how the change in resistance ... What that means is that if the thermistor has a resistance of 10000 ... ohms at 25°C, then at 15°C it'll have a resistance of 14,883 ohms, and ... THE VOLTAGE DIVIDER ...
    (sci.electronics.basics)