Build Your Own Clone Message Board

Putting together the Mouse pedal and paying a lot of attention to the schematic and what it's actually doing (especially with the different types of diodes) has been huge for me in my understanding of hard clipping.

I have one nagging question about how diodes work.

My understanding (which may be wrong) is that in a hard clipping circuit, there's an amplifier of some sort that's pushing the guitar signal up into significant voltages, then there are a pair of diodes after that can shunt signal to ground. They don't start working immediately though, and it takes a certain amount of voltage for them to "kick in," and this is the forward voltage spec that pedal people are talking about when they compare Ge to Si to LED or whatever. This has the practical effect of clipping the tops off of the waves, which is what we hear as distortion.

So in this crappy drawing, the blue line is what be what the amplifier would be doing if the diodes weren't there, and the red line is what happens when there are clipping diodes in there:


*not to scale

If this is correct, my question is: in the time between t1 and t2, why is the signal-flat lining and not going to zero? Like once the threshold has been met for the diode to engage, why doesn't ALL the signal go to ground?

Yes, your interpretation is basically correct, though there is a time factor in the turn-on that rounds off the corners of that plateau somewhat. But the diode, once the voltage threshold is reached, is acting like a dam, not an open door--only the voltage above the threshold is shunted to ground, not all of it.

_________________
“My favorite programming language is SOLDER” - Bob Pease (RIP)

My Website * My Musical Gear * My DIY Pedals: Pg.1 - Pg.2

Awesome, thanks so much.

I'm not an engineer, but when I was in college (25 years ago) I hung out with mostly engineers. One of them was a guitar player, and we would occasionally go into his lab and fire up an oscilloscope and zoom in on that part of the wave you're talking about, trying to get a solid state thing to sound more like a tube thing by softening that knee.

Is that "speed to engage" a spec that's typically in a diode's data sheet? It would be cool to be able to compare them when diode shopping. I've tried googling for it, but I don't think I know enough of the correct jargon to get there.

I've googled all of this, by the way. I know way more about P-N junctions than I ought to just make a guitar chug, and if anything, I think it was making it worse.

I really appreciate you entertaining these questions.

When I was in college (49 years ago), I hung out with mostly co-eds.


Hmm, I think I'd search using the terms you mentioned. Not sure if there's a way to find that out on a datasheet, but it's certainly possible. There's some good info included in this classic R.G. Keen article on musical distortion (along with a lot of other useful stuff): http://www.geofex.com/effxfaq/distn101.htm

_________________
“My favorite programming language is SOLDER” - Bob Pease (RIP)

My Website * My Musical Gear * My DIY Pedals: Pg.1 - Pg.2

There is in fact an amount of time that it takes for the innards of the diode to slosh around into its new configuration. I've encouraged my students to study the difference between rectifier diodes (1N4001, say) and signal diodes (1N914, say) by plotting their I vs. V in real time, as driven by a function generator, and cranking up the frequency until hysteresis loops appear. Sadly, or happily, I am at home at the moment and don't have my notebooks at hand so I can't quote you any frequencies, but the differences are easily seen in the lab. But yeah, never mind my notebooks, what do the data sheets say?

I didn't find anything useful on the first diode data sheet I scared up (1N400x) so instead I went to try to learn about how they are modeled in LTSpice.

https://www.allaboutcircuits.com/textbook/semiconductors/chpt-3/spice-models/

There's a follow-up here:

https://electronics.stackexchange.com/questions/566706/what-is-exactly-modeled-by-transit-time-in-spice-diode-model

I'd read somewhere that LTSpice simulations show the "onset time" effect but that must be when you change some of the default parameters from zero. Perhaps the "transit time" and "junction capacitance" are the ones to look at, but even so I haven't yet found much in the way of actual numbers, much less what they translate into concerning whether they contribute to corner rounding. But there are some--if you scroll down in that All About Circuits page, you'll see a table that shows significant differences between the 1N4148 (signal diode, small capacitance, small "transit time") and the 1N4004 (rectifier diode, larger capacitance, much larger "transit time"). Even so, if the 4.32 microsec "transit time" for the slow one (1N4004) is the order of magnitude time over which something happens, that still corresponds to > 200 kHz, way out of anyone's hearing.

I always thought you could round off the distortion corners by putting a small resistance in series with the clipping diode, but I suspect that for that to work the way you want it to, you need to know a lot about the rest of the design of that stage of the circuit so you'll know roughly about how much current will be flowing.

I had a chance to check my notebooks and I discovered that even at 100-150 Hz there is discernible evidence of hysteresis in the 1N4001 I-V characteristic. So it definitely seems that there could be some audio-band consequences of the sloshiness of diodes when used as clippers.

Cool, thanks. It sounds like the exact curve of when they kick on or off is dependent on a lot of variables, and wouldn't necessarily lend itself to a simple one-liner in a data sheet, but good to know that there is a difference in the audible range.

In another thread, which I did not want to hijack for this purpose, the properties of Zener diodes are being discussed slightly obliquely. So I decided to upload this graph of the I-V characteristic of a 1N4733A Zener diode (5.1 V).

This is an oscilloscope trace which shows, upon proper interpretation of the two axes, the current you get in the diode (vertical axis) as a function of the voltage applied across it (Vf, forward voltage; horizontal axis). Vf > 0 corresponds to forward bias of the p-n junction and, as it is silicon, you really get a substantial current flow once the voltage reaches 0.6 or 0.7 V, and thereafter even a tiny increase in the forward voltage yields a great increase in the current. That's the steep branch in the curve on the right side of the graph.

For a rectifier diode, that's basically the end of the story. For a Zener diode, however, there also can be substantial current flow in the other direction if the reverse bias is strong enough. In the case seen here, once the reverse bias reaches a few volts, there can be substantial current flowing in the reverse direction. That's the downward part of the curve on the left-hand side. Note that for this diode, the onset of the reverse current is more gradual than in the forward direction. This means that there is not such a well-defined threshold voltage for current flow, and use of this Zener diode as a voltage reference requires taking care to insist on a particular current, chosen to match the desired voltage.



The typical "diode test" function of a digital multimeter is the DMM's attempt to answer the following question: "What voltage do I (the DMM) have to apply between my two terminals to get a specific amount of current to flow?" The specific amount is part of the DMM design and is typically on the order of 1.0 mA. Since the forward and reverse conduction characteristics are different, the "answer" will depend on which way around the meter probes are attached to the terminals of the diode.