Revisiting old failures – Roland AD-50 Double Beat Fuzz section

[bmsiddall]

Ever since I'd first heard them, I've loved the old Roland 1970s fuzz/fuzz-wah pedals.  Bee Baa AF-100, Double Beat AD-50 and Bee Gee AF-60 are all extraordinary and unique fuzz pedals.  My favourite of these is the AD-50, which at full bore accurately replicates the sound of an amplifier on the verge of exploding.  Dial back your guitar volume to 70% or so and you have a kick arse, aggressive rhythm tone to play around with.  The full glory H-bomb tone is just a short twist of the dial away!   

When I "graduated" from building pedal kits back into 2016 or so, I started using veroboard and naturally tried my hand at the fuzz section of the AD-50, with a layout from guitar fx layouts.  It always seemed a little fizzy/nastier than the YouTube clips (on all settings), so I ended up buying a vintage pedal to AB.  No comparison- mine was not that great, so I put it aside (I'd found by then I didn't really enjoy troubleshooting veroboard!)  and used the vintage unit until I got around to redoing the circuit on perfboard with a layout by dmk2113 on the other forum.  Nailed it this time  . 

Same as my earlier failure, it has volume and fuzz pots and a three-way switch giving the classic "triangle", "square" and "sine" wave tones.  As I imagine most people do, I leave it on square wave for the full bomb blast tones this beast is capable of.  Here's the fuzz section schematic I've used:



And so the failed version sat in its enclosure in the bottom of my drawer until a few weeks ago.  Wanted to do another one but also didn't want to duplicate the one I'd built.  I decided to get rid of the restrictive three-way tone switch for a real tone stack, with the tone pot replacing the fuzz (I prefer to use guitar volume to control fuzz for this unit).  I also wanted to keep the three-way switch and find some other use for it.  I briefly considered the stupidly wonderful tone control, but I really wanted to have control over bass and treble with a single control, so I settled on a nice tilt control I found here: https://www.diystompboxes.com/smfforum/index.php?topic=89717.0/.  Works really well!

As per the schematic below, I used the three-way switch to give stock (high gain), lower gain/crossover distortion, and bass cut on the emitter of Q2.  With the diode crossover circuit (am I actually achieving crossover distortion here?), I wasn't sure if I needed both diodes but checking them individually in circuit didn't really make an audible difference until both were used.

Anyway, here it is, my version of the greatest/most underrated fuzz pedal circuit of all time 

[Mark Hammer]

It's a great fuzz, stock.  I only had 6-pos switches, and couldn't stand to see three rotary-switch positions wasted, so I thought up a few other sounds not readily achievable with a continuous tone control like you posted.

Here you see the three traditional sounds as the first three settings.  Next (D) is a 2-pole lowpass filter. E adds crossover distortion for another layer of clipping.  Finally, F is a midscoop filter.  I tried to achieve a similar output level from each, though not perfectly.



Here is what I ended up with, although I'm not sure what I'm hearing corresponds to the drawing.  I may have altered one or more of the presets.  Not able to confirm or deny that at the moment, because I ended up selling it to a guy on the coast who just HAD to have it.

The simplicity of the "voice selection", however, could easily be extended to a 12-position switch, or one could use the second pole of a 6PDT rotary to select something else for more complex voice construction, perhaps by changing the bias of one of the transistors, or swapping the 1uf emitter cap for a different value.

[bmsiddall]

Hi Mark.

I'm glad you posted (and greetings from Australia).

I read your earlier thread on the six way tone mod ages ago and the crossover distortion idea I stole from you actually    I might end up adding the additional tone options to my other one sooner or later.

While we're on the subject, the original tone switch idea has a 6.8nf capacitor in circuit providing some level of high cut/roll off at all times, right? I notice your 6 way stack adds some additional capacitance. Any chance this might affect the tone on the squarewave/330 K output? I'm guessing it's probably not audible but not sure how to calculate this?

Cheers,
Brett

[Mark Hammer]

Are you talking about the second 6n8 cap on option C?  I'm assuming that Roland's intent was to make a crude bandpass filter, with the low-end rolled off by the series 6n8, and the top end rolled off by the one going to ground.

Looking at the 3 drawings (original, your redraw, my redraw), I'm not seeing what I think you are referring to.  Near as I can tell/see, the 330k path simply passes the full bandwidth of the fuzz, without any additional filtering.

If one was concerned about any unwanted effects from all those 6 different path inputs being tied together, you can always use the second pole of a 6PDT rotary to provide true bypass for each of those little networks.  That is, the rotary switch selects between connecting each path with both the 0.1uf/1M junction and the volume control.

[bmsiddall]

Yep, I was talking about option C, with options D and F seemingly adding a little more capacitance than the original cct.  Probably doesn't make any audible difference, but good suggestion about using the second rotary sw. pole.

Thx!

[snk]

It's a great fuzz, stock.  I only had 6-pos switches, and couldn't stand to see three rotary-switch positions wasted, so I thought up a few other sounds not readily achievable with a continuous tone control like you posted.

Here you see the three traditional sounds as the first three settings.  Next (D) is a 2-pole lowpass filter. E adds crossover distortion for another layer of clipping.  Finally, F is a midscoop filter.  I tried to achieve a similar output level from each, though not perfectly.



Here is what I ended up with, although I'm not sure what I'm hearing corresponds to the drawing.  I may have altered one or more of the presets.  Not able to confirm or deny that at the moment, because I ended up selling it to a guy on the coast who just HAD to have it.

Hi,
Thank you Mark for you update on this vintage Fuzz!
I built it today, using DirtBoxFX layout. It works as expected, except that the output volume is much below unity gain (and i have read the output volume should be massive).
Am I right assuming that Q1 & Q2 are here to shape the distortion, and Q3 is responsible for the output level?

[Mark Hammer]

Yep, it's loud, and Q3 is jointly responsible for making it so.  I know on the one I made, I used a 6PDT rotary switch, which disconnected the unused filters.  I can't speak for the original, but I wonder if the low output you are experiencing comes from passive bleed arising from multiple parallel paths to ground.

[snk]

Hi, Mark. Thank you for your quick answer. I am also using a 6PDT, fwiw.
I'll check the HFE of my transistors. I have both TO92 and metal can 2N222A : does it make any difference?

[iainpunk]

(am I actually achieving crossover distortion here?)

technically, no, but sonically, yes you are.
to be honest, i'm somewhat of a crossover distortion connoisseur, and have tried lots and lots of ways to create crossover-ish distortion, but never have i seen that particular method, ill try it the first thing when i get home!

ow, a cool trick with corssover distortion is adding a little bit of 180 degrees out of phase clean signal, it gives a really wonky feel to chords, almost like ring-mod.

I have both TO92 and metal can 2N222A : does it make any difference

no, technically yes, but the variances between individuals in both types greatly overlap in terms of gain.

cheers, Iain

[Mark Hammer]

The can-vs-plastic thing doesn't make much difference in this application, BUT, as always, one needs to verify pin assignments.  I'm constantly amazed at how many different ways there are to have the wrong pins when there's only 3 of them.

[snk]

I have measured the HFE of my 2N222A : they were all around 270.

I have also audio-probed some parts of the circuit : wow, this thing can sound amazing! I managed to hear a loud, fuzzy sound, while running my audio probe on the layout.
Here is what i found so far :
- I tried to turn my transistors 180° : it sounded roughly the same 
- When I put the audio probe at the emitter of Q3, i hear what i guess the circuit should sound like (a loud, warm, delicious fuzzy sound!)
- After the 100nF cap, after the diodes, i can hear the same loud fuzz sound.
- In fact, the layout sounds fine "nearly everywhere", except where the circuit goes to the rotary switch : there, i hear the same fuzzy sound, only much, much quieter.

I am using this layout (posted below) :

• If I put my audio probe on the right side of the lines A, C, G, it sounds loud
• If I put my audio probe on the right side of the lines B,D,E,F,I,J, it sounds quiet

https://dirtboxlayouts.blogspot.com/2020/10/roland-ad-50-double-beat-w-mods-by-mark.html

[Mark Hammer]

1) Thanks for giving me credit.  Much appreciated, but it's easy when you start with a great initial design and simply tack on a few mods..

2) Building on, and comprehending, stripboard is one of those things that has eluded me, so I'm going to bow out of deciphering and debugging your posted layout.  I don't speak "that language" very well, but I'm sure there are others here who do, and I look forward to their insights.

3) Why Boss has chosen to rerelease other of their pedals in "Waza" revisions, and neglected this gem, is beyond me.

[snk]

1) Thanks for giving me credit. 

I didn't create and published the layout  I just linked it from this great blog featuring a huge collection of veroboard layouts.

Yes, veroboards can be trickier to debug, as each one is unique and more prone to errors. But it's a great way to try, discover and experiment


Well, I have compared my build with the layout for the 10th times, made a picture of my build and zoomed it (to find traces, cuts, etc), and found nothing more... I don't get what's going on... Around Q3 and after the 100nF cap, the sound is very loud, but when it "goes" to the rotary switch, i get this volume drop.
It's like some of the sound is going to ground, but i don't really know what else to look for...

On a positive note, for anyone willing to build this excellent fuzz : I tried with BC547, and found that it sounds very good (full, bodied) with it.

[snk]

Well, it' still puzzling me...
I checked, and checked again, and didn't notice a suspect trace or cut in the circuit. I checked the connections one more time, and from what I see everything seems fine.

What I can say is that I audio probed the circuit, and the sound is plenty loud after Q3, before going to the rotary : after the 100nF capacitor, the sound is loud, but at the "last" lug of the 150k resistor going to the 250pF cap (and then to rotary 2), the sound is very soft.


My knowledge of electronics is quite limited, sadly, but I am trying to understand : if the volume is much lower, it might be because some of it is going to ground, right? What else could it be? What to look for, and where to look at?
I am trying to understand the reasons why the sound is getting weak (if it is not a transistor issue).
I have checked the 150k resistor value, and it is fine. At one end of it (coming from the 100nF cap), it's loud, and at the other end (going to the 1.5nF cap), it's weak...
I think my wiring is ok, I don't see any trace or cut on my veroboard, and re-counted every components and they are all here... Do you know what could be the culprit?

[iainpunk]

What I can say is that I audio probed the circuit, and the sound is plenty loud after Q3, before going to the rotary : after the 100nF capacitor, the sound is loud, but at the "last" lug of the 150k resistor going to the 250pF cap (and then to rotary 2), the sound is very soft.

well that resistor forms a voltage divider together with the volume pot, essentially it does what its supposed to... its just not a loud pedal i guess.
you could add a make-up-gain transistor to give a bit more volume if you want, just add any singe transistor gain stage, like the last transistor in a big muff does. (yes, you can just steal that gain stage if you want.)

cheers

[snk]

Thank you, Iainpunk, much appreciated !

you could add a make-up-gain transistor to give a bit more volume if you want,

Yes, I had considered it (adding a LPB, for instance).
But first, I wanted to be sure that I hadn't done anything wrong and that the circuit behavior is quite optimum;)

its just not a loud pedal i guess.

This is what I thought first (I've tried several vintage Fuzz or wah with significant volume dropout). But from what I could read elsewhere, the circuit should be very loud, so it's more likely that something is wrong with my build.

well that resistor forms a voltage divider together with the volume pot, essentially it does what its supposed to...

Ok, so by changing its value, I could try to get back to unity gain, then? Would it affect the sound color in any other way?
I guess that, since the 150k resistor is just the beginning of one of the 6 parallel "tone/modes" path, I would have to adjust also the values of the others paths (A...F)?

[snk]

Ok, assuming that I built the circuit properly, and that the volume loss is caused by the voltage divider resistors in each parallel circuit, I'm trying to figure out the best way to compensate for this volume loss (which is around 6 to 12dB from my estimation).

• 1- adding a booster after the circuit. It should be the fastest, easiest way to do (since i wouldn't have to change or tweak anything on the D50 circuit). But, to me, it seems that it would be also the laziest solution, not leading to the best results : it adds another amplification stage, resulting in a noisier circuit, wouldn't it? Wouldn't it be smarter to lower less the volume in the earlier stage (after the transistors, before the rotary switch), and not having to boost the signal accordingly afterwards?
• 2- tweaking the voltage divider resistors, so that they do not lower the signal that much. Having 6 parallel circuits, it means that i would have to tweak them all. Is there any online calculator available so I wouldn't do it blindly and know which ballpark value to reach for?
• 3- boosting even more the signal at the early stage, so it compensates for the voltage dividing afterwards. It sounds like a faster way than solution 2, but it may also make the signal to noise ratio worse?

Which solution would you find the best?

[duck_arse]

is your signal only softer at point B, or at each of those switch contacts? you should probably be "listening" to the switch positions by monitoring at "fuzz tone select", the switch common, as it's only when the filter string is loaded by the volume pot that any of this matters. and - simplest, non-lazy way would be to add a last stage transistor, as in either the superfuzz or the big-muff pi circuits, as they both do exactly the thing you are trying.

[snk]

Thank you, Duck Arse.

is your signal only softer at point B, or at each of those switch contacts?

At each of those switch contacts.

simplest, non-lazy way would be to add a last stage transistor, as in either the superfuzz or the big-muff pi circuits, as they both do exactly the thing you are trying.

Perfect : I came to the conclusion that it might be wiser to add a small booster daughterboard than tweaking 6 resistors on a busy circuit at the risk of creating other issues, and while i'm replying to this message, i have the LPB layout in front of me, along with a blank veroboard

[snk]

Hi,

- The LPB daughterboard did the trick perfectly ! It allows me to have the LEVEL knob not at the maximum all the time and have plenty output!

- So, I played a bit with every modes, and I would like to thank Mark Hammer for the added modes : it adds even more versatility to an already well featured pedal.
I even like mode 4 better than mode 3 : I find the mode 4 to be warmer and richer, while mode 3 sounds quite similar but a bit duller... Is there any suggestion to spice it up a bit, and make it sound slightly different? I am thinking maybe some added resonance around the cutoff point, maybe, or some added filtering?

I am going to build an enclosure for this little gem tonight, thank you again all for your help