Switch LED on blinking signal/input

Thread Starter

tunafish24

Joined Oct 21, 2011
28
I started a thread a while ago, but the thread is old and the system won't let me post to it now. Plus I have a different setup in mind, so starting a new thread.

Disclaimer: I'm a Computer Science guy with only basic understanding of electronics, so would appreicate detailed replies (THanks!)

Problem: My PC case has a power LED, which is really bright and when I put the system in standby, it starts blinking - which is very annoying when I'm trying to sleep. What I want is a small circuit (with no external power because system is in standby), that is essentially self-driven in everyway. I will attach it to the power LED wires and also replace the power LED with either two LEDs or a bi-color LED. So that when the output is continous (system is on), green LED will light up, however, when the system goes into standby - the input signal will be blinking then the circuit should turn on (blinking is fine as well) the orange LED.
 

Thread Starter

tunafish24

Joined Oct 21, 2011
28
yea...but that means hooking into the ATX power connector and I don't wanna go that route.

I have been able to fade-in/out LED - which is obviously very easy. I think there should be an easy way to detect blinking signal and power different LED. I was just checking the voltage across RC Led circuit during blinking and I noticed that the voltage never reached peak...perhaps the Capacitor never got fully charged? OTOH, while the current was constant, the voltage across circuit was stable 4.7V. Can we take advantage of this in some way?
 

#12

Joined Nov 30, 2010
18,224
Not reaching 4.7V while blinking opens up options that you measured wrongly.
If you used an analog meter, the needle might not be fast enough to register the highest voltage and a digital meter might not refresh often enough. Just checking the details. You got this covered?

Making the LED dimmer is as simple as a 10 cent resistor. Reversing polarity without using any external power takes a lot more brain work and a lot more parts. How hard are you willing to work, considering you aren't willing to splice into a 5 volt supply wire?
 

Thread Starter

tunafish24

Joined Oct 21, 2011
28
I used a digital mulimeter and I did suspect that the meter might not be reporting the peak, so I kept looking at it for quite a bit (10-15 seconds) and didn't see the voltage hit 4.7V once. Unless, it was some freakish concidence that the meter's refresh rate was perfectly in sync with the blinking signal, I think the reading should be valid. I'd check it out again today...just to confirm.

We don't have to reverse polarity per say, could we use an RC filter along with an op-amp or comparator and then switch output between LEDs?

Basically, I'm looking for something standalone/non-instrusive. I have a new PC and don't wanna cut PSU wires.
 

#12

Joined Nov 30, 2010
18,224
If you use a diode and capacitor as a crude peak detector, that could solve the measuring uncertainty. Maybe a 1N4001 and a hundred microfarads? Then add .6 volts to what the meter says.

PS, Thanks for the ideas about how to do this.
I think cmos op-amps waste very little power.
I'll go look it up.

It would help immensly if you could provide a schematic of the "RC LED circuit".
 

Thread Starter

tunafish24

Joined Oct 21, 2011
28
The RC Led circuit I have is pretty simple. It's a diode connected in a series with the resistor, then capacitor in parallel and finally the LED. This gives a very smooth fade-in/out, as opposed to putting reistor After the capacitor, which gives a smooth fade-out but rapid fade-in.

I like the trickery this guy has done to make an LED flasher...I'm looking for simple circuit like that (http://wild-bohemian.com/electronics/flasher.html)
 

wayneh

Joined Sep 9, 2010
17,495
How does all this solve the problem of the light keeping you from sleeping? So far all the discussion is on replacing one flasher with another. Maybe I missed something?
 

Thread Starter

tunafish24

Joined Oct 21, 2011
28
It started off with the annoying LED, now I know that I can add a resistor and it'll dim the LED, if I add an RC circuit then I can also have it fade-in/out. What I'm trying to do now is to have the LED switch colors. Agreed it has nothing to do with the original problem, I'm just trying to improve on the existing solution and learn along the way :)
 

Thread Starter

tunafish24

Joined Oct 21, 2011
28
FWIW, I really like the way the power LED "throbs" on my PowerMac G5. Slooow in an out fades.
I've got that throbbing effect going as well...it looks really cool. Still I was wondering if I could get an orange/amber color LED to throb on standby, instead of the green one...it'll look nicer.
 

Thread Starter

tunafish24

Joined Oct 21, 2011
28
thanks...have you tested it in a simulation software or it's just a schematic? btw, can you briefly describe how it works?
 

#12

Joined Nov 30, 2010
18,224
I don't simulate. It's just a schematic that I presented. I only tested it by building sections of it, then seeing what they do on my oscilloscope.

It works by rectifying the off transitions to make a negative voltage that turns off the jfets. If the fets are off, the orange LED lights. If the fets are on, the green LED lights up.
The 1uf cap and the 2.2 meg resistor set the time the fets stay off after pulsing stops. You will have to adjust that. The Big uf capacitor has to hold enough charge to light up the orage LED between "on" pulses...or not. You can leave it out and let the orange one pulse.

You might have to make the pnp into a Darlington. Depends on whether you can get a transistor with enough gain and how much idle current your particular jfet uses.
 
Last edited:
Top