LED clipping only one lights up

[duck_arse]

if anyone owns two vactrols and two multimeters, they could try this - replace the clipping leds with the vactrol leds, back-to-back (ohhh, derr). connect the meters across each vactrol ldr, crank the signal, see which/when/how much the resistance drops compared to the oscilloscope showings.

(I did in fact oh, derr myself while typing this. we can't see the leds in the vactrols, which is the whole thread point, nur?)

[karbomusic]

(which is the whole thread point, nur?)

I built this yesterday...




It doesn't clip the signal in any way until those purple LEDS illuminate. I saw several references early in the thread that said they would still act like diodes and clip even if not illuminated, they don't; nor do they clip with any other set of LEDs until they illuminate. Why?

It's not a question that they are clipping when I can't see them, but an observation that they never clip the audio signal until I do see them. I'm not debugging anything, this is an academic disparity between what people I trust say and what I'm seeing actually occur.

[induction]

It doesn't clip the signal in any way until those purple LEDS illuminate.

I assume you mean that you don't hear clipping until they illuminate, or did you scope the output?

The more the signal exceeds the forward voltage of the diodes, the more they will light up, and the more noticeable the sonic effect of the clipping will be. Small amounts of clipping can go unnoticed if you measure purely by ear.

[karbomusic]

I assume you mean that you don't hear clipping until they illuminate, or did you scope the output?

I did it all... Ears, eyes, oscilloscope, FFT and find no evidence whatsoever of any clipping whatsoever until they illuminate...

http://www.diystompboxes.com/smfforum/index.php?topic=109330.msg999645#msg999645

The more the signal exceeds the forward voltage of the diodes, the more they will light up, and the more noticeable the sonic effect of the clipping will be.

That is my understanding of how it should work but others (again whom I trust) mentioned in this thread that it would still be clipping even though the LEDs weren't lighting, I'm trying to find how that would be true so I can get some sleep again in my life. I assume I'm misreading their statements?

[antonis]

That is my understanding of how it should work but others (again whom I trust) mentioned it would still be clipping even though the LEDs weren't lighting, I'm trying to find how that would be true so I can get some sleep again in my life.

Simply by connecting an LED to your multimeter (in the Diode test mode ) in a not absolutely dark room..

[induction]

Sorry Kary, I noticed you said that before and I missed it. Can you post a schematic of the circuit in question and point out where you are taking the measurement from?

As to the original topic, I believe a boosted guitar signal may very often be asymmetric in itself, (with the positive swing usually being stronger than the negative swing in my sims, but the opposite may be possible as well). That could easily account for the asymmetric lighting of leds that I've noticed in my own builds. I also found that swapping the leds didn't change the pattern. The led that clipped the positive signal lit up more, no matter which led it was. If that's the case, then the asymmetric lighting of the leds may be different if you feed it a sine wave instead of a real guitar signal.

[karbomusic]

Sorry Kary, I noticed you said that before and I missed it. Can you post a schematic of the circuit in question and point out where you are taking the measurement from?

It's just the standard LEDs>GND shunt on the output, nothing more. Is that enough? Let me know if not. I had realized this months ago and built a circuit based on it, cool, I'm good. Then I saw this thread saying they don't need to be illuminating to clip, but I could not find that to be true. So in my tests (as per my very first and second posts), I just used an AWG/Scope for the screenshots but its the exact same thing (the pedal above was just another representation of the exact same behavior). It's not like I'm new to the test I performed which is why it caught my eye. But the simple thing is that, I'm not seeing or hearing or measuring it clipping until the FV is exceeded causing the LEDs to illuminate. Sounds fine by me, test prove it but a few others mentioned otherwise so it's important that I know why there is a discrepancy comment wise since I greatly trust those people.

[karbomusic]

Simply by connecting an LED to your multimeter (in the Diode test mode ) in a not absolutely dark room..

These are 'audio' stomp boxes and if the phenomenon cannot be seen nor heard, it is meaningless.   I apologize for being a thorn in everyone's arse but see the first two sentences of the second post in this thread, I can't find how that is the case and I always take Mark's comments to heart and need to understand what I am missing here.

[GGBB]

I could be wrong, but my understanding (assumption?) has always been that a diode does not conduct until it gets enough voltage to move electrons out of the depletion layer. If true, that would explain why the scope grabs show no clipping at lower signal levels. Also, I would expect that the "turn on" voltage is related to the current vs. forward voltage characteristics of the diode, and since Vf is proportional to If, its not going to be a fixed number for every diode in all situations. And since LEDs have higher Vf ratings than e.g. silicon diodes and vary in their efficiencies for light output, a small signal could cause the LED to conduct/clip but not produce light.

So in my mind anyway all of this makes sense of what we are seeing here.

[karbomusic]

I could be wrong, but my understanding (assumption?) has always been that a diode does not conduct until it gets enough voltage to move electrons out of the depletion layer. If true, that would explain why the scope grabs show no clipping at lower signal levels

Agreed and that is exactly what I have been hinting at. I see nor hear clipping before there is enough voltage to light the LEDs.

a small signal could cause the LED to conduct/clip but not produce light.

I'd just like to see evidence of that in a LEDs/GND/Output setup because it was my understanding that a photon must be released. Not saying that I can see a single photon but since I see/hear/test no clipping until I do see them, I assume it must not be clipping until I can see light. Let's not bog down in single photons though, that isn't the concern. The concern is it isn't clipping until photons are being released as best I can tell.

I've used the test equipment I'm using a LOT and my ears a LOT and as it stands at this moment, I trust they aren't lying to me about this. It isn't clipping until they illuminate.

So in my mind anyway all of this makes sense of what we are seeing here.

As far as the stuff I was testing, yes, it does make sense that they don't generally light until clipping starts and that if the conduction is so low that you can't see the light, you can't hear the clipping either (making it irrelevant). Again, assuming a LEDs/GND/Output configuration.

[GGBB]

I'd just like to see evidence of that in a LEDs/GND/Output setup because it was my understanding that a photon must be released.

I think that's correct, and after doing a bit of reading/skimming, I think what could be a factor is the refractive index of the LED.  The photon will be reflected back into the LED material unless it is able to escape the surface.  So not only may it be difficult to see a single photon, but we probably don't even get the opportunity to see every one of them that the LED actually releases EDIT: semiconductor material actually emits.

[karbomusic]

I think that's correct, and after doing a bit of reading/skimming, I think what could be a factor is the refractive index of the LED.  The photon will be reflected back into the LED material unless it is able to escape the surface.  So not only may it be difficult to see a single photon, but we probably don't even get the opportunity to see every one of them that the LED actually releases.

I agree but I'm sort of calling out the opposite, it isn't clipping in any meaningful way until I can see it. It's not a question of its clipping and I can't see it but the exact opposite, it isn't clipping until I do see it. Sorry again for being so pedantic but its important. Big thanks to you guys for being so patient. 

[GGBB]

I agree but I'm sort of calling out the opposite, it isn't clipping in any meaningful way until I can see it. It's not a question of its clipping and I can't see it but the exact opposite, it isn't clipping until I do see it. Sorry again for being so pedantic but its important. Big thanks to you guys for being so patient.  

Clipping occurs if there is any voltage drop across the diode.
Meaningful/audible clipping might be something else (namely a certain level or higher).
Any voltage drop across an LED produces photons.
Visible photons might be something else.

You could very well be right that, for the LEDs you use, "meaningful" clipping is always accompanied by visible photons - only you can decide what level of clipping is meaningful to you.

[samhay]

The emitted photons will all be a similar colour, I suspect. I can't see single photons - I've tried - and I doubt you guys can either.
My 2c - the LEDs are likely conducting a little, but this is not significantly colouring the sound (literally).

[GGBB]

The emitted photons will all be a similar colour, I suspect.

Yes - an LED will produce a range of wavelengths within its colour range. To be clear, when I wrote "visible photons" I meant those that actually escape from the LED (see my previous posts), not those that are in the visible light spectrum.

[karbomusic]

Assuming normal real world visible spectrum LEDs (Let's not overcomplicate)....

Thanks for all the comments. Based on my testing and to keep the original focus here is what I have concluded:

1. Assuming decent viewing conditions such as a dimly lit room, the waveform is not clipping if the LEDs aren't lighting. So much so that hearing it becomes irrelevant, it simply isn't occurring.
2. Due to #1, if one LED isn't lighting, that side of the waveform is not being clipped. The ears, eyes, scope & FFT prove it.
3. The original poster had a concern based on #2 which I at this point consider a "potential" problem.
4. The op also scoped his circuit and only 1/2 of the wave is being clipped which also matches my testing.
5. I can't find where "it's OK, its clipping even though you don't see it" is the correct answer here.

The point here is I don't think the original poster's circuit is clipping properly even though we are saying it most likely is. Maybe VRef is misplaced or it's a disparity between +/- rail swing. If the latter then OK, it is working properly (semantics) but physical conditions are something to live with. The other important academic point is that based on what I'm finding, on LED not lighting at all there is no way to assume it is clipping, there just isn't until one of you guys send me off with my tail between my legs due to some test that proves otherwise.

Why did I go to all this trouble? I think it is important from an engineering standpoint that we don't assume it's clipping when it isn't. I'm the dummy newb here so I still expect someone to show me the err of my ways but until that post appears, if it ain't flashing, it ain't clipping.  

EDIT: Except based on some LED types as slacker so elegantly found below. Thank you for testing.

[slacker]

EDIT: posted this without reading the last few posts, so it isn't a response to them.

In the interests of science I've done the same  experiment as Kary, with two different sets of LEDs. In both cases I used a 1k series resistor and increased the input level until I got either clipping or the diodes lit up.
The first pair were a couple of "normal" green LEDS, VF about 2 volts, the top left image shows these clipping a fair bit with no visible light being emitted, even in the dark looking at them side on I couldn't see anything. Top right is the exact same settings but with the diodes pulled to show the actual level. If I'd have thought about it a better example would have been to use the second channel to show the input and output signals together.
The bottom images are the same test using some clear blue "superbright" LEDs VF about 2.5 Volts. These behave how Kary's LEDs did, they light up before there's any noticeable clipping.



So there you go, some LEDs can conduct before you shove enough current into them to make them light up.

[karbomusic]

@slacker

Thank YOU! Now I need to try the remaining LED types I haven't tried to round this out then scroll up and see what LEDs the OP was using.

[slacker]

Yeah I think the type of LED plays a bit part in this, the green ones I used need a good few mA though them to produce enough light to be used as an indicator in a pedal the blue ones are usable at fractions of a mA.

[GGBB]

So there you go, some LEDs can conduct before you shove enough current into them to make them light up.

That's what I was trying to say. As always, a picture is worth a thousand words. Thanks.