LED clipping only one lights up

[morten_beta]

Hey All,

I'm baffled. I'm building a Distortion + circuit and tried LEDs for clipping what puzzles me though is that only one lights up. Does that mean that only one is clipping/engaging or is it only the one going in the "right" direction that will light?

Been googling but no dice. Maybe I'm asking the wrong question but I have not been able to figure it out.

Cheers,
Morten

[Mark Hammer]

It is not necessary for LEDs to light up.  They are serving only as diodes.  If the current they conduct is enough to cause illumination, great, but if not that doesn't stop them from being diodes.  Perhaps one of your diodes is more efficient than the other (I.e, lights up in response to less current).

[karbomusic]

If the current they conduct is enough to cause illumination, great, but if not that doesn't stop them from being diodes.  

I didn't know that. I've ran some clipping diodes several times and using the oscilloscope they never clipped the signal before the light turned on. I couldn't find any evidence (in the ones I was testing) of clipping before I could see the light. Notwithstanding being in a darker room of course. So, that makes me curious about the above from an academic standpoint. I was just using some purple clear ones yesterday (higher FV) to act as a clip detector in a boost circuit (and when it did happen to clip it would be little more pleasant), that's why I had to verify the clipping  (and lack of) via oscilloscope and FFT. Now I'm confused again.  

Also, I have not seen them mismatched on flashing in general that is a curiosity as well.

[anotherjim]

Something isn't symmetrical.

Which LED isn't lighting?

Is the amp output about half supply voltage with no signal?

Is the 1uF coupling cap between amp and diodes leaky? The right way around (if polarised) - positive to the amp output?

If you fit a higher value cap - does that make both light?

[morten_beta]

The LED with the positive lead to ground is not lighting.

Yes, amp output is about half.

Coupling cap is right way around. Putting a bigger on in parallel did not make the other LED light (I double checked that the LED works and it does).

I don't know how to check for leaks ... But it's brand new. Not that that is any guarantee.

[Caferacernoc]

Switch the LED's to see if it's the diodes or the circuit causing the problem.

[Mark Hammer]

Keep in mind that the superbrights so many of us buy for mere pennies are often 15 times more efficient than the ones found on Boss pedals, or in clipping circuits in some amps.

[anotherjim]

If the cap was leaky, putting another across it won't cure it.
Test (no signal) that the voltage on the -ve of the cap is zero. If it was leaking, it will act like a resistor in a potential divider with the series resistor and the volume control pot. If good, it will be charged up to equal the reference voltage level of amp output on the positive  and 0 volts on it's negative plate.

On another matter, in the stock circuit, the resistor and pot are both 10k. So they divide the output of the amp by 2 before the LED's can act. The usual amp isn't a rail to rail type (have you used a 741?), so maybe you get 3 to 4 volts peak to peak swing at most out of it due to the divider. That's 1.5 to 2 volts at most for each LED, so they won't get a lot of time when they can turn on enough to be visible. In fact, it's surprising even one can light!

Only slight asymmetry could be making the difference here. Either in the amount of voltage swing the amp can do above or below the 1/2 supply bias voltage, the charge versus discharge rate of the polarized cap isn't quite the same, or the LED's are a bit different in threshold and efficiency. You might reduce the value of the 10k series resistor to get more consistent clipping.

[morten_beta]

@anotherjim

Your'e a genius! I feel my noob-ism shining through.

So the cap is not leaky.

Yes I'm using a 741 opamp. I have more and it's socketed so I can swap it out and try another one. However putting a 10k in parallel with the series resistor made a difference. The secon LED is lighting up now, not as much though as the other one — but I like to know that it's engaging. Swapping them gives the same result so same difference.

That probably still means asymmetry but I wouldn't know how to fix it. But then again it doesn't have to be a problem I just really wanted to know that it's actually working properly.

[GibsonGM]

Generally, you have no need to fix "asymmetry"...most people like the sound, WHEN they can hear the difference.   To move more toward a 'symmetrical' clipping, you can find 2 diodes and measure their Vf's (forward voltage drops, or "turn on voltage" for the layman).   Ones that match will be more symmetrical.   

Both your diodes are likely going on, just that, as explained, you can't see one of them.   When I work on this stuff, I can usually HEAR if one side isn't turning on because there's almost an odd "clean" guitar tone under the one that's clipping due to one side not being affected...

Typically, the effect isn't noticed much unless you go to more different Vf's, like a Si diode and an LED, or 2 Si and one LED, etc.  YMMV.

[duck_arse]

you can shift the Vref voltage up a little, ?~700mV?, to overcome/offset the lesser downward swing of the opamp.

[karbomusic]

Can you guys help a fellow out?

I'm sitting here with a handful of LEDS, an AWG, a GND, an oscilloscope and an FFT. I cannot in any way get that waveform to clip before I can see the LEDs illuminating. In every instance the LEDs begin to illuminate slightly before the waveform clips. I'm using the FFT to identify any harmonics created due to clipping in the event I don't notice it on the waveform display, no dice, not there before they illuminate.

Reading all the posts above, I'm interpreting the assumption that clipping of the waveform should be occurring before they illuminate. What gives?

[morten_beta]

That is interesting ...

[karbomusic]

I have screenshots if it helps (Red LEDs):

Below turn on voltage




Above turn on voltage




LEDs removed

[anotherjim]

you can shift the Vref voltage up a little, ?~700mV?, to overcome/offset the lesser downward swing of the opamp.

I think that would do it.

In Morten's case, I would want to scope the output of the amp at the point the LED's appeared to clip - to make sure the clipping isn't the amp.

Karbomusic,
I don't think you're breaking the laws of physics.
2 things going on I'd say.
1 - Persistence of vision. The LED's are only actually lighting up briefly on the very tip.
2 - The Tips of a wave are actually already clipped. That sounds stupid I know, but at the tip any wave is DC, rate of change is zero (it has actually, briefly, stopped) -  so no harmonics.

I'm no signal test specialist, but there could probably be some % clipping amount required before measuring harmonics in a lab for it to be meaningful.

[karbomusic]

I don't think you're breaking the laws of physics.
2 things going on I'd say.
1 - Persistence of vision. The LED's are only actually lighting up briefly on the very tip.

I suppose my confusion is that I don't hear it clipping before I see it either, I have found no evidence using any measurements or senses I have to tell me clipping of the audio is occurring in any meaningful form until after the LEDs are illuminating. I'm just trying to find my vast misinterpretation based on that finding and what everyone else say's is occurring. I fully blame me, I just don't get the point if it doesn't affect the audio signal. What the heck am I missing here.

2 - The Tips of a wave are actually already clipped. That sounds stupid I know, but at the tip any wave is DC, rate of change is zero (it has actually, briefly, stopped) -  so no harmonics.
I'm no signal test specialist, but there could probably be some % clipping amount required before measuring harmonics in a lab for it to be meaningful.

Now I really don't get it since the shape of the waveform is what creates the harmonics. Based on general testing I've done over the past year, the scope seems to be well within the limits we are talking about.

[morten_beta]

This is getting a bit above my skill level — but still I enjoy the discussion. I am a stage where I can read a schematic, I can tweak it slightly and I can build it. This is my first opamp build. I'm used to transistor circuits like fuzz face, tonebender, booster-stuff and that whole divided voltage really confuses me. But I'm trying to learn the physics even though I absolutely suck at math.

I assumed that since the diodes are not lighting up equally, they are not clipping equally? That's why I thought that when one is not lighting it's not clipping — as Karbomusic's tests points towards. Again just me not knowing the physics of the stuff. I totally go by ear.

- Morten

[nate77]

Although this is totally unhelpful, largely every circuit I've built using LEDs as clipping diodes, the first one in series lights up much brighter than the other. I foolishly assumed that this was to be expected since I had seen it so often.

[morten_beta]

Nate77 it might be unhelpful but certainly not uninteresting. I just hooked the thing up to a sine wave and an oscilloscope app and I get the same result as the images above show. A clean curve until the LEDs light up I see clipping on one side and not the other. I'm not saying that it's not there but it's not detected. When I use Ge, Si or even rectifiers I see a more symmetry faster. Of course at lower signal input but they also seem to even out faster — does that make sense to anyone or have I lost everybody?

- Morten

[nate77]

I'm following you, I'm just baffled as to why. I'm very stifled as well as piqued by this whole thing.