Clipping diode question

[kirs]

Looking for clipping options for a DIY Rat I just build I put together a Flexi Clip, removed both clipping diodes, and put the input to the clip on anode pad of one of the diodes and the output on the cathode pad of the same diode.  The other diode's pads are empty.



All those switches are really just a DIP.  D1, D2 are 1N914, D3, D4 3mm LEDs, D5, D6 are 1N34a and the last two are 1N34a + BS170.

It seems to work great when any two opposite diodes are engaged, for example both SI diodes, or one SI with cathode to ground and a LED with anode from ground.  However, when only one diode is engaged there's no clipping, as apparent by one LED will not light up unless a complimentary diode is engaged.

Am I missing something basic as in it's not possible as advertised, or should I start trouble shooting?  It's a very simple circuit and all joints look good under inspection.  I'd like to think I have a pretty good understanding of how this stuff works, but for the life of me I can't see why only one diode will not engage.

Thanks.

[ashcat_lt]

My Rat type thing seems to have the same type problem.  I don't have quite so many options.  I just have a DP5T rotary to select  between combinations of one or two silicon diodes on each side.

The position with no diodes engaged sounds exactly the same as the one with only one diode.  I thought I must have wired it wrong somehow.  Haven't had a chance to open it back up, but now I'm waiting with baited breath for someone to explain why this doesn't work.

[kirs]

Hmm, my problem is a bit different.  It works great when two clipping diodes are on, but I just can't seem to get one-diode clipping going for one-diode asymmetric.  

Sounds like yours are not turning on at all or are always on.   The Rat by itself distorts without diodes if the gain is too high.  A lot of the dirt comes from the op-amp which could be what you're hearing.

[kirs]

Trying to debug and finding interesting things.  If I hook a multimeter diode test on the clipper circuit input and ground I'll get a reading when the switch is on.  Seems like the circuit works as advertised from a analytic perspective, but when I throw on just one LED, once and a while it will light up and pulse dimly with the attack then putter out dead.

At this point I am more interested in understanding what's going on than getting one-diode clipping to work.  I suspect it's just a gap in my understand of how this type of circuit would work than anything else.

[ashcat_lt]

Maybe I didn't describe so well.  Mine's got this:

I call it SALWO - Silicon, Assymetric, "LED", Way assymetric, Off

The W and the O sound (and look) the same.  W is one diode one way and no diode the other.  Everything else sounds fine.  I did the same diode test before I installed the switch and had the results I expected.

BTW - Here is my "build report for this pedal over at GN2.

[kirs]

Seems like a similar issue to what I am having.  One diode by itself will not clip at all.  It needs symmetry with a diode of the opposite orientation.  Does not seem to matter which one as I am mixing a LED and Ge diode now.  Sounds great.

If anyone has any insight, I'd love to hear the explaination behind it.

[ashcat_lt]

I was really hoping that somebody would come along and say "Yep, that happens and here's why..."

I can almost accept the fact that mine has a wiring fault, but can't see any reason yours shouldn't work.  Does it matter which side of the waveform you're trying to clip?  I mean, I assume you've tried (for example) each LED seperately and neither one seems to work by itself, right?

[slacker]

I know it's not any help to you but I can't see any reason why it shouldn't work. If the signal is large enough to turn on the diode or LED it will clip, should make no difference whether both sides of the signal are being clipped or one.

If for some reason the signal before the clipper was already very asymmetric then it's possible that the level of one side wasn't enough to turn the diodes on, but then you'd get clipping in one direction but not the other. If you're saying one diode/LED in either direction does nothing then I can't think what it could be.

[boogietone]

Seems like a similar issue to what I am having.  One diode by itself will not clip at all.  It needs symmetry with a diode of the opposite orientation.  Does not seem to matter which one as I am mixing a LED and Ge diode now.  Sounds great.

If anyone has any insight, I'd love to hear the explaination behind it.

I have also experienced this same conundrum and have no explanation for it. Used two dip switches to put various diode types in and out of the clipping section. It did not matter what type of diodes were selected but it did require two in opposition to get clipping.

[teemuk]

What's the source impedance driving the diodes? Is there AC coupling?

Low enough source impedance and capacitively coupled clipping section with no proper tie to a DC reference (resistor to ground) will modulate the DC bias offset when you have just a single clipping diode to ground. Basically, you have just built a "sort-of-a half wave rectifier": You are drawing current fairly asymmetrically through a single diode to ground, and this current draw is now charging the coupling capacitor to a DC potential. As the DC offset thus shifts the clipping diode's forward voltage is no longer exceeded in a manner that would result into clipping of the waveform.

This problem does not portray itself as much with symmetric clipping because the current draw is more symmetric, thus preventing capacitors from charging as much.

Solutions is: First, tie the signal path's node where the clipping diodes connect to a steady potential, usually 0V, with a resistor. Next, increase series resistance feeding the clipping diodes. For example, if the series resistance is 100 ohms, pump it up to 1K. If there's no series resistance at all, the put some there. Note that the lower the capacitance of the coupling caps is the easier they'll also charge and the more you need to increase the series resistance to overcome the resulting RC filter. So, low coupling cap values will require higher series resistance to prevent DC offset shift. Vise versa, you can tamper around by increasing the coupling capacitance.

The following graph illustrates the phenomenon with three different series resistance values of R2 = 100R (green wave), 1K (blue) and 10K (red). The series resistor will likely soften up the clipping a bit, as it is limiting the diode current as well as attenuating the signal a bit. However, with proper design the effect is likely negligible. However, the series resistor also limits the current draw that charges the coupling capacitor so all in all the DC offset modulation has a much lesser magnitude and the diode clipper can perform as intented.

Does it change the overall tone? Damn straight: Now the circuit actually works.  

I throw on just one LED, once and a while it will light up and pulse dimly with the attack then putter out dead

This very same phenomenon is also revealed in the graph: The initial attack of the signal gets clipped but while this happens it also charges the capacitor and modulates the DC offset, "lifting" signal below the clipping point. The DC voltage will eventually settle back to zero at idle but a new signal burst will quickly repeat the phenomenon over and over again.

[ashcat_lt]

Thanks for that.  It's kind of what I was thinking, but I couldn't be sure.  Mine has 1K going into the coupling cap before the clipper.  Is that the series resistance you're suggesting to change?  Won't that change the overall tone of the unit?  Would a resistor to ground right before the clipping section help?  What's a good value for that?

[kirs]

I throw on just one LED, once and a while it will light up and pulse dimly with the attack then putter out dead

This very same phenomenon is also revealed in the graph: The initial attack of the signal gets clipped but while this happens it also charges the capacitor and modulates the DC offset, "lifting" signal below the clipping point. The DC voltage will eventually settle back to zero at idle but a new signal burst will quickly repeat the phenomenon over and over again.

Yes!  An explanation!   Exactly what I needed, a scientific explanation as to what's going on.  I get it and now that I do I can certainly live with it.   Thank you!

This is the stock clipping circuit, BTW.

[Brossman]

I think the easiest way to make asymmetric clipping is to put 2 diodes on one side and one on the other... Maybe this will clip too much, but it's definately asymmetric.

AAANNNND, (It is my understanding that...) Ge diodes arent as powerful as Si diodes. SO when putting, say, 2 - 1n34a's opposite a 1n4001, that will make symmetric clipping of the waveform.  SO THEN, by this logic, putting 1 Si with one 1 Ge diode will make an asymmetric clip...?

[kirs]

Or a LED