Horror of the Harmonic Percolator Pedal: Analysis (Guitar Amplification and Effects Bonus Lecture)
Вставка
- Опубліковано 7 сер 2024
- Support this channel via a special purpose donation to the Georgia Tech Foundation (GTF210000920), earmarked for my work: • Support This Channel v...
Here is the class playlist: • Guitar Amplification &...
0:00 -- Introduction
0:55 -- Schematic
2:29 -- Input control
3:08 -- Amplifier topology
5:35 -- Biasing
There is a part missing on this drawing. I was not the one to put my name on this schematic. This was done without my knowledge.
That's good to know... I'm going pin this to make sure people see it.
There is a high frequency bypass cap missing that will diminish the highs but also prevent self oscillation at full gain. The diagram is not clear on connections or non-connections. Nowadays if you use nf for caps, all of them should be in nf but this was drawn 15 years ago?
Hey! New to the channel but I wanted to say thanks for taking the time to present this analysis. You're the first person I've seen intuitively break down the perc as two series gain stages-I think because the schematic is drawn in a non-intuitive fashion. Your technique of moving blocks around was also really nice for visualization!
Now, I can't say why the shared AC ground was done in the first place, but I can speak a bit to the nonlinear effects. At very low frequencies, the impedance of the 47uF cap increases enough that the transistors can wiggle each others' bias. This produces some interesting behavior at extreme low frequencies - there's a little bit of a unique sub-octave thing going on. You can also get very interesting results by changing the capacitor value or adding some resistance in series with it.
Anyway, thanks again for making these videos!
Welcome!
Notably there are multiple versions of the original Harmonic Percolator circuit which is interesting not only for the weirdness of the circuit but also the story involved (unobtainable 70's pedal with wax potted transistors made by a mystery guy).
The analyzed Giblet schematic is NOT the same as Albini's pedal. The biasing is different. Different resistors and caps used in key locations.
I became obsessed with this pedal and dreamed up what is basically a sandbox version that allows one to dial in the Giblet or Albini versions and everything in between. Emerald Ox executed my vision (and he made a killer Giblet version of the HP that got me obsessed with it in the first place). The version we came up with is called the Chimeric Percolator.
A couple things I learned from spending time making sounds (as well as building and modding them myself) with both of the popular versions of the circuit, as well as the Chimeric Percolator... 1) the Albini version is great for a smoother sounding HP but his clangy guitar sound really comes from his use of full frequency range speakers, an IVP preamp, and thin copper picks. He BARELY uses the HP, only for short passages. Also he uses a buffered gate pedal before the HP, which will make it sound thin in use (hey, he goes for harshness) even though his HP is actually smoother sounding than the Giblet version. 2) The sub-octave can only be had in the Giblet version of the HP circuit, not the Albini version. I love the sub-octave from the Harmonic Percolator because it is raspy-chaotic and wild. It's a bit magic. With the Chimeric Percolator I can dial in the sub-octave and the range in which it occurs, getting it on the high or low E string with the sub getting as loud as the fundamental frequency. The clipping diodes need to be engaged for the sub-octave to work (some pedals offer diode-lift). Finally, leaky components might help with the sub-octave as well. One HP build version I have goes into self-oscillation when the harmonics (input) is turned down low. In fact, the sub-octave seems to come out the most when the harmonics knob to be within the first 1/4 of its range.
Anyways, cool to learn more about the circuit in this pedal that I've been obsessed with.
Thanks for these insights! Particularly this one: " In fact, the sub-octave seems to come out the most when the harmonics knob to be within the first 1/4 of its range." I can't even begin to wrap my head around that one.
The Jordan Bosstone also uses a PNP/NPN pair of transtors, but I believe the biasing is a little more conventional. Still, it's not very often you see both varieties of transistor used together.
Albini is def a top-shelf reference, highly accomplished recording engineer
Big RIP to Steve Albini :(((
Glad this popped up in my recommended though. Been curious about this circuit a while.
He was so generous with sharing his knowledge, and a tireless advocate for musicians navigating an exploitative industry. A tremendous loss.
I think you nailed it on this one because the flip-flop of the capacitor linking to the bases is acting as a bias and is likely reducing power consumption. It's extremely similar to a circuit I created a few years ago, but I didn't think tying the bases together with a capacitor would complete a voltage divider to bias the bases, and my goal in that project was to try to produce something truly Class-A that was high output but better efficiency.
What I ended up doing was voltage divider biases both bases independently, tying the bypass cap into the node where the Q2's collector runs series into Q1's emitter, and ultimately created what I perceive to be a power supply current booster which gave me about another volt of AC gain on the output of Q1, and surely significantly more current gain but I never measured that theory to be sure.
Audibly, there's a richness and density boost when comparing it to other 2 transistor driven gain circuits.
The ultimate goal was to also use an ultra steep low pass so no 60 cycle or other RF ground buzz got through without compromising treble content by the time the signal reaches the speaker, and that absolutely did work. I've been able to use some super steep low-pass filtering in a few preamps and overdrive designs that eliminate 100% of 60, RF buzz and hissing white/pink noise but still have heaps of treble and pinch harmonic sleaze because the harmonic content of the upper bands is still greatly abundant, and I effectively defeated virtually 100% of the noise.
*My current project* is a 3 transistor fuzz that 100% nails the tube distortion with the fizz, zing and note separation we still currently only experience from valves(granted digital is now almost perfectly there) where when I push it hard enough, I even get the synth level soft knee deep tube clipping that sounds like a synthesizer, but it has some distance to grow until I can use it as a standalone high gain lead distortion that provides shredding as well as very bassy and clustered swept arpeggios that elite note separation expected only from tubes and the most recent and expensive digital modelers.
*I didn't finish, but going to college for electrical engineering is one of the best things I ever did for myself.* As my wordsmithing shows, I'm highly passionate about the science.
Never seen anything like this before.. Definitely building this up!
Any luck? I want to build one too.
This is far from scientific, but I've read somewhere/otherwise speculated that it's a play on a push pull amplifier (albeit with the PNP on "top," weird), with the intent to purposefully produce crossover distortion.
The idea is that most fuzzes/distortions etc. affect only the peaks and valleys of a signal, whereas near the bias point it tends to be untouched or otherwise inaudibly changed. With the clippers on the output, the Peculator does both.
(Big Disclaimer: I'm not an EE, don't even play one on TV, apologies if I'm being imprecise if not outright inaccurate with my terminology and/or "analysis" here.)
If I am on the right track though, re: crossover distortion, see also the Zvex Machine, Tim Escobedo's Tripple [sic] Fuzz, which in turn in based on Ken Stone's Simple Wavefolder, which in turn again is based on "Nonselective Frequency Tripler Uses Transistor Saturation Characteristics" by R. Lockhart, Jr." The CGS page (Elby or archived original) talks about some of the theory.
Ah, and also Escobedo's Harmonic Jerkulator, which is a Perculator clone made with contemporary, non-unobtainainium parts.
Thanks for the tip, I just looked up that schematic... it's interesting to see what resistor values they changed to compensate for the switch from germanium to silicon on one of them.
I can totally see why people think the Harmonic Percolator might be a push-pull thing because of the way it's drawn, but I think the way it's drawn misleads people about the structure.
I just took a look at the Zvex Machine -- that's definitely using push-pull structure.
Given I'm so late to the game on this one, I doubt you'll see this, but the DC bias side of things almost looks like a cascode circuit, but the AC side of things looks like a cascade circuit. Definitely bizzare.
I love this! Thanks so much for this beautiful explanation
Thank you for your kind words!
The Crowther Audio Prunes & Custard is a very unusual circuit, definitely recommend taking a peek at the schematic. Some kind of diode ladder thing
Let me take a peek... ohhhhh! the core of that looks like the core of the Yusynth/Haible wavefolder. Veeeery interesting!
Ha. I just noticed in your video that my name “John Greene” is on the schematic you are using.
This was a lot of fun to analyze back on Aron’s form. Very “unconventional”. Perfect example for how you analyze circuits!
Oh wow! I'm so glad you liked it. (And thank you and your fellow stompbox analysts for your work back in the day!)
This is a very interesting circuit! It's always confusing to try to analyse distortion guitar pedals with traditional small signal analysis since the they are meant to be non linear effects. The only obvious difference I can find with this design is the reduced headroom for each gain stage.
Indeed, a proper analysis needs to be a "large signal analysis" and figure out when the various transistors are pushed into clipping. So start we'd need to figure out the bias points, but that was more work than I wanted to put in right now. If I'm lucky someone else will be inspired to tackle it. ;)
My first experience with the HP was a small modified pedal called the Even Anguish by DIRGE Electronics. It's incredibly crazy, and frankly sometime shard to control, but that's what makes it fun. Such an amazing circuit! Hopefully I can make my own
I could be totally off-base here, but it seems like the first two gain stages of the Boss HM-2 (following the buffer) are inspired by the Harmonic Percolator, but NPN into PNP instead of PNP into NPN; there is also more filtering in the feedback loop (like a Big Muff). These transistors in the HM-2 also share a bias current like the HP-1. The HM-2's op-amp seems to also feed back between the two transistors, it's like a shunt within a shunt... SHUNTCEPTION
That's really interesting! There is a difference though -- here the "shared bias current" is *split* between the Q6 and Q7, whereas in the Harmonic Percolator it flows through both transistors.
I can't wrap my head around VR4. Since the left in of VR4 is connected to the collector I'm not sure we can think of it as a feedback connection; we generally think of collectors as outputs. It's more like a 2nd input to op amp 1b... maybe?
BRAIN HURTING. ;)
Check this out (assuming UA-cam doesn't delete the link): www.diystompboxes.com/smfforum/index.php?topic=114410.0
@@Lantertronics Thanks! Makes a little more sense (in the layman’s Dunning-Kreuger-inflected sense)
Guitarist with electronics degree here, this guy speaks my language
It's interesting to hear your analysis of the circuit. I have a weird sickness where I measure the current requirements of all my pedals, and HPs which don't have lots of mods to them typically run around only 1 ma. (You mentioned in the video the low power requirement of a single bias current.) The original and Theramaniacs versions run off of a battery only... and one of the clones I have (and which is the closest to the Theramaniacs one) added a "normal" power supply plug, but has this weird thing where the ground is positive, so it works fine if you run it off of an isolated power supply, but isn't friendly with daisy chained power supplies... and this pedal measures at 0ma for some reason.
I don't know the first thing about dancing electrons... I just have a compulsion to collect HPs.
In this arrangement, both transistors are effectively running at half of the supply voltage. I've never examined one of these. Might the transistors also be distorting the signal or would their output be too low for that? Q1 might be clipping when the input level is raised, especially with hotter pickups. Upon further thought, Q2 would eventually be driven into clipping as well. An interesting circuit.
I have case for you: the Jen harmon booster. It has a very unique stage for Distortion (i think). It has standard class A gain stages and kind of a big muff tone control, but the middle stage looks unique. The simulation shows a strange asymetrical behavior. This stage has a trimmer to set bias, this looks to have a big impact in the wave shaping.
Nobody seems to copy this topology even it sound like an overdrive before their time.
Dude it's a percolator you got to build it you got to put it on breadboard and you got to play with that big cap . That has a huge effect on the cct. Its only a short to dc... it lets audio kinda "dwell" in that cap almost like an oil can delay.
In some cases, as long as that cap has reached charge, you can pull one of the q's and it will still work. Very similar kinda to the bypass cap on the fuzz with a fuzzface.
Fun video. Subbed.
I remember reading that the percaltor was essentially designed by trial and error, playing around with dead gear in a music store. So that bias sceam is probably just a mistake that worked out.
Hah, that wouldn't surprise me!
its not the same but: phantom powered diff. preamps use stacked jfets and bjts sharing dc current: input jfets has other jfet as current source (which connects to ground through "on" indication led) and drives bjts upstream. I think youre right and this is for power economy.
This kind of looks like a cascode amplifier circuit, where the top transistor is a common base amplifier and the bottom one is a common emitter.
I think it's also worth pointing out that the name seems to be inspired by the similarity in topology to : en.wikipedia.org/wiki/Coffee_percolator
Interesting. I like the way you analyze the circuit separately for DC and AC. Since both transistors share the bias current, the thing that's most interesting is what happens on C4 when signal is applied: will the DC voltage across it change? How will it affect the waveform of the signal coming through?
Great info! Sub'd!
Thanks!
This is similar to a ”Totem-pole amplifier”. Please make an analysis of Catalinbread Formula 55, which emulates the Fender 5E3 Deluxe really good
I'd love a whole series of this. The people that make the clone pcb boards have reverse engineered literally every analog pedal. Do some of the classics like the wooly mammoth, green ringer or the heavy metal pedal.
You've probably run across them, but if not, check out my videos "L43: Diode-Based Overdrive and Distortion Effects" (at 17:50 mark I briefly talk about the Heavy Metal pedal, although it's definitely worthy of a more detailed analysis; and I should note looking at the comments that my analysis of the series diodes may be off target) and "L44: Foundations of Fuzz Effects."
I have some ideas for videos I'd like to do -- not sure when I'll have time to get around to them.
While I'm at it, I highly suggest checking out Arito Suzuki's channel, he does all kinds of deep dives into pedals like this.
Curious what you'd make of the Jordan bosstone
I have tip for circuit analysis!
I recently checked schematic for Boss Power Driver PW-2 which is really obscure pedal on its own. There are people who claim that it sound similar to Big Muff. So, despite I'm not good at reading schematic I had to check it for myself if there are some similarities. It seems there are no clipping diods or clipping circuit there. Which is weird for driver/distortion.
Interesting, so they're entirely getting the distortion from overdriving the op amp! I think there was a Craig Anderson design like that.
Great video. Very interesting. You should check out Mask Audio Electronics. Alec Breslow makes a lot of fuzz pedals and has expressed a lot of love for the harmonic percolator. Lots of cool fuzz circuit stacking.
Do the MXR polyblue Octave pedal. And see why it's so fragile. I had one sent from the USA to Australia that was broken on arrival. Like I use the power supply it come with. With a travel adaptor. And it didn't work. So I tryed the power supply from my outlaw tuner Power supply aka iron horse. And still didn't work. Then I used the MXR power supply to run my boss DD7 and it worked fine. I'm no sparky but I'm thinking two things. Something got past Q.C at MXR or they sent the wrong power supply. It says 9v but it struggle to light the led light out of the box.
That's weird -- does the power supply it came with work for other pedals?
If you've got signal ground on the emitters of the two transistors (on top of the fact that one's Ge and one's Si) would that be some sort of pseudo Differential Pair setup? If that's the case, the cap between collectors is just standard for high frequency compensation, right?
The issue isn't that one's germanium and one's silicon; the issue is that one is PNP and the other is NPN. For it to act as a differential amp you need two PNPs or two NPNs. (Going to ground like that they'd need to be NPNs.)
@@Lantertronics But I meant more from the perspective you presented which has that 47uF cap to ground between them. Differential pair I suppose would have one demanding current while the other is lessened to effect a differential signal between collectors, which I suppose isn't the case here, but I just thought it would be one way to look at it. First reaction was "Emitter follower into common base", at least that way it makes sense.
Sorry, not terribly cohesive tonight.
@@kopaka647 "Emitter follower into common base" -- that's EXACTLY what I first thought when I looked at it! :)
I then got confused by the output of the first stage going to the base of the 2nd. It took me ages of just staring at it until I figured out it wasn't.
@@Lantertronics So that would make it... PNP common collector capacitively coupled into the base of the NPN going common collector just like any ol' 2 stage amplifier? I'm putting together in my head how your video went now, it makes a lot of sense to see it how you laid it out here. I guess something that might be fun dumb experimenting is screwing with the value of that 47uF to see what creeps down through the bias. I suppose whatever creeps across would be a lowpass filtered negative feedback since it would be out of phase with the signal. But I guess the signal flips twice, actually, so that would be... odd to experiment with. You can kind of filter out both phases from both stages at... the same time? I suppose the emitter leg with the higher "R" in the RC with that 47uF would be the signal that dominates.
I'm probably getting incoherent here, sorry. It's late.
VENERABLE Prof. LANTERMAN, I IMPLORE YOU, if you are looking for a topic for a video, i have, in my opinion, the single most important topic in pedal history - DEMYSTIFYING THE FETZER VALVE. Maybe academically more known as the Triode Emulator by Dimitri Danyuk, it is a type of circuit that attempts to use a JFET transistor in place of a Vacuum Tube. I think you are the only one equipped to deal with this problem. We should look for answers to a few simple questions and maybe delve deeper thereafter:
Why is it similar to a tube, and why the need to have such specific biasing?
Why is the J201 so special, some say it's used so often in these circuits because it's "not just another regular JFET"
Is it possible to use MOSFETs in place, as some have used them in similar configurations?
DM me for a list of links for research
Bittersweet getting this in my feed so soon after the Master’s demise RIP Senore Albini
It was a tremendous loss. He was always so generous with sharing his knowledge.
Could this be a BJT version of the JFET Mu Amp circuit?
Definitely looks a lot like something along the lines of a mu-amp / totempole / SRPP kind of topology - i.e. a push-pull amplifier.
That's exactly what I was thinking last night, also Sziklai pair sort of like the Axis Fuzz.
@@EversonBernardes I can see why people think that, given the way it's drawn. But the mu-amp (which I was just introduced today) is more of a cascode, with the top transistor being a common gate configuration.
What's the purpose of D2? If it's anode is connected to 0V and there is no negative power rail, then it's never conducting & not doing anything? Perhaps except for some few pF varicap action creating an RC lowpass, but with a cutoff too high to be audible, and perhaps some miniscule reverse leakage which shouldn't be audible either (unless it was some very old and very low quality diode germanium)... so what's the idea behind it? Is it perhaps some kind of protection from stuff connected to the output jack?
It's on the right side of the C6, so at that point the signal is ground referenced through the volume pot (vs whatever DC bias is on the left side of C6). At that point in the circuit whatever the power rails are is irrelevant.
@@Lantertronics thanks, that makes sense... (i get so easily lost in single supply circuits... embarrassing
I'm glad that you did this vid, but...is there no significance to one being pnp and the other being npn? No phase issues or anything? That's what I wanted to know about. What is the significance of one being pnp and the other being npn? Also, the point of the HP was that it was supposed to emphasize, or so I've read, even order harmonics, like tubes. That was the point of it. Is there any truth to that?
I read this too and i wondered if there was any truth to it, i did see a chap on YT doing some signal analysis with something or other (i think maybe a tubescreamer?) in which you could see the odd-order harmonics generated...i could never stand the tube screamer so it made sense to me from that angle 😂
i do have an oscilloscope somewhere so i'll have to investigate.
What i would say about HP type circuits is....i've built a whole bunch of them and have just one bad sounding one. You can mess about with component values and dump most transistor values in and they will still sound at least good, unlike say a fuzz face which in my experience is just swapping parts out til it sounds okay, a total nightmare.
I REALLY like a zonk machine, but that is even finnickier. Honestly all my silicone fuzz builds sound amazing every time, but theres just something about a TB mk1 or 2, zonk machine or nkt275 fuzz face that is on a different level when you get a perfect one :)
There is this wolf computer by mellowtone id love yo understand, sure there are schematics there
Also the fender blender
Those 2 intrigue me a lot
Thanks this was finely educational
It looks like a two-transistor Fuzz Face type circuit (shunt-series feedback configuration) with a common emitter amplifier pasted in front to overdrive the Fuzz Face type circuit more, and the fuzz control is put on the common emitter amp instead of the feedback circuit.
Here's my lecture on common emitter amplifiers: ua-cam.com/video/ipoLVp9TLfA/v-deo.html
Here's my lecture on the Fuzz Face: ua-cam.com/video/xuzT7CSYtyI/v-deo.html
Fender Blender looks very interesting: files.effectsdatabase.com/docs/schematics/fender_blender.pdf
1st stage looks like a Fuzz Face, but the output is taken from the emitter of Q2 instead of the collector. Last stage is Q4 driving back-to-back diodes in parallel to ground, so that's going to be a Boss DS-1/Pro Co Rat type of distortion. What's really interesting is Q3 and diodes D1 and D2; I'd have to think have those, I haven't seen anything quite like that.
You want guitar pedals with weird biasing and cascade setups? Devi Ever's pedals have loads of designs like that.
Hah! You know, I can't make heads or tails of most of her schematics. I have no idea how she comes up with some of that.
They feel... kind of random.
Guitar pedals and modular synths are one of the few areas where a non-rigorous method like throwing components together to see what sticks could produce a desirable output.
What is the most accurate harmonic percolator to Albinis on the market right now?
As a side note, RIP Steve Albini
No idea -- if you find out what the most accurate HP out there is, let me know!
I think you are only scratching the surface on the weirdness of this circuit in you analysis. Depending on the characteristics of the transistors used, and the values of the capacitors; the circuit can exhibit a sort of triggered oscillation. It's almost like an octave divider. I would like to see an in depth analysis of how this could occur.
My first thought would be to dig into C4; I made a pretty standard split between a "DC bias" circuit and an "AC small signal circuit" that assumed the cap was a perfect short at audio frequencies. It may be doing some weird things with the bass.
@@Lantertronics I think that is part of it. If the ac signal squeezes by that bypass cap then it could be reentering Q1. Almost like in a common base configuration. I also think having a leaky germanium transistor for Q1 contributes to this.
@@ianwoodmusic7137 Your common base configuration note is on target -- when I first saw it I assumed it was some sort of mutant cascode until I looked more closely. Maybe there is some mutant cascode action going on...
@@Lantertronics If I was actually an electrical engineer I could probably explain it better, but I'm just a DIY musician who likes to tinker.
@@ianwoodmusic7137 That's how it starts! There is always more to learn! :)
Sim sim. Tem que prender o cabelo e fazer a barba também. Só não vimos você fazer...