Boss Waza-Craft buffer: what's up with that?

[Fancy Lime]

Hi there,

I've been trying to figure out what the deal is with the Waza-Craft pedals. Apart from custom modes, which essentially activate mods, the big thing seems to be different buffers. I could not find a schematic for these magically better buffers. Does anyone have one?

Also and somewhat unrelated: does anyone have a schematic for the ML-2 Metal Core? Interesting sounding pedal, I'd be interested to see how closely that is related to the HM-2 and HM-3, just out of curiosity.
EDIT: Found some info, finally. The ML-2 is digital, so no point in having a schematic.

Thanks,
Andy

[PRR]

> magically better buffers. Does anyone have one?

https://www.amazon.com/Buffer-Natural-ITAY-Compact-Polisher/dp/B07HXH1XK8?th=1
https://ttte.fandom.com/wiki/The_Magic_Buffers
https://www.promegaconnections.com/what-makes-a-good-buffer/

[bushidov]

Well played, PRR

[PEPPER!]

The engineers who have devoted their lives to building Waza-Craft pedals wake up at 4am each morning to be the first to arrive at the Akibahara electronics market.  They hand-select only the finest components with the merciless exactitude of a master sushi chef, and may reject hundreds of resistors or LEDs before finding the ones that are truly great enough to go into a Waza Craft Metal Zone or Blues Driver.

[alexradium]

in 2019 there are still people who believe in the magical properties of vintage tubescreamers AND that moon landing was a hoax.
i hope to live till 2039 just to see what is next.
Amusing.

[Fancy Lime]

Well, that is kind of the point here. Lots of people try to sell you fairy dust sprinkled Tube Screamers and the like for a premium price. But Boss tends not to go overboard with the voodoo claims. So when they start a "premium" line and one of the biggest changes is a different buffer circuitry, that's enough to make me at least want to have a look at the schematic before deciding if it's hokum or not. Especially since there is not a lot you can do to improve the already pretty good boss buffer of the standard series other than use higher quality transistors or go to opamps. My suspicion is that they did the latter, since people claim that they can here a difference when the pedal is in bypass (yes, I know, "People on the internet are saying" isn't exactly evidence), and the former would probably barely be detectable with a good scope but almost certainly not by ear through a guitar amp. Mostly I am interested if they indeed did something meaningful or if they just replaced the ceramic caps with mica and the resistors with wire wound types and called it an improvement.

Andy

[Ice-9]

I have the Boss Waza craft DM2-W, it is all SMD apart from the coolaudio bbd's and clock chip. 1x4558 smd dual op amp and a 571 compander chip. Looks pretty standard to me.

[Fancy Lime]

Hi Mick,

you don't by any chance have photos of both sides of the PCB from which the buffer circuitry could be traced, do you? In exchange for my eternal gratitude, of course. Sorry, its the best I have.

I'd really like to get to the bottom of the "we have improved the buffer" claim, considering that the standard boss buffers were never something to be sneezed at to begin with. SMD and modern robotized production is obviously going to be an improvement over 80's through hole hand soldering in terms of consistency, parasitic capacitance, and so on. But that would not qualify as "improving the buffer" but as "improving the production process". So I am still curious how much of the buffer-claim is marketing BS or what I might learn about better buffer design. The other improvements to the Waza pedals, like adjusted EQ voicings, better quality components with tighter tolerances, and so on are all understandable to me. But that §%"&§ buffer is driving me crazy.

Also: Interesting that they used a 4558. I cannot think of a single reason to choose that over one of the myriad of lower-noise, better stability modern drop-in replacements. Other than Mojo and I really hoped that the whole Waza thing wouldn't just be Mojo-based.

Thanks,
Andy

[Rob Strand]

"we have improved the buffer"

When you read claims like that you have to ask *what* is better.   
In one man's eye cheaper is better  .

One "flaw" in discrete buffers is the slight gain loss (roughly ~0.075dB for a BJT buffer). 
So if you have two buffers in the effect you might get 0.15dB.   So when you have a number of
effects in cascade the loss adds up, perhaps to 0.5dB to 1dB.   Usually it's not noticed and not a problem.
The JFET buffers can have quite a bit more loss.

One fix for this is to use a buffer with feedback so the gain is 1.
I can't find a good example schematic but you would use a circuit like this with C2 and R4 removed,
then the emitter of the first transistor connects to the collector of the second transistor.
(You can use whatever biasing scheme you like on the base of Q1, like a single resistor
to Vref.)


JFET version:



I'm not saying that's what Boss use.   All I'm saying is there's is a small problem and that's one way to fix it with a discrete design.    Another common method is to replace the emitter resistor of the one-transistor buffer circuit with a current source.

[bluebunny]

"we have improved the buffer"

When you read claims like that you have to ask *what* is better.   

Their profit, perhaps?   

[Rob Strand]

Their profit, perhaps?

I was only using that as an example but yes profit is a better goal.  A lot of companies associate making things cheap as meaning higher profit.  If you spend a lot of time fixing problems in production or looking like idiots because the product is unreliable you don't end up making a profit!

[Ice-9]

Hi Mick,

you don't by any chance have photos of both sides of the PCB from which the buffer circuitry could be traced, do you? In exchange for my eternal gratitude, of course. Sorry, its the best I have.

Also: Interesting that they used a 4558. I cannot think of a single reason to choose that over one of the myriad of lower-noise, better stability modern drop-in replacements. Other than Mojo and I really hoped that the whole Waza thing wouldn't just be Mojo-based.

Thanks,
Andy

Yeah I will get some photo's today of both sides of the board, it is a pain to even look at the pcb as there is so much screen print on it is off putting.

[Ben N]

Interesting that they used a 4558. I cannot think of a single reason to choose that over one of the myriad of lower-noise, better stability modern drop-in replacements. Other than Mojo and I really hoped that the whole Waza thing wouldn't just be Mojo-based.

Andy, it's a bucket brigade delay, a severely limited, thoroughly obsolete technology that wouldn't even exist any more if not for OCD guitarists with their vacuum tubes and their LM308s and such, and their reverence for instruments made by hand in the 1950s. I'm not saying the sound isn't worth it, but if it's not about mojo, really, what is it about? And in the larger scheme of things, does the difference between a 4558 and a 5532 matter all that much in an analog delay pedal?

[Ice-9]

Ok so here are some pics of the bottom in full and a section of the top of the PCB around the buffer in, op amp and compander sections. I doubt there is any real difference in the buffers to be honest, the 4558 looks like it is used in the exact same way as on the original DM-2. The main difference for me on this waza craft is that the delay has been extended with 3x BBD chips.

Sorry the pics won't give a great way of tracing anything, without fully removing glued in wires etc this is the best I can get.

Top right of the top side is the 4558 area.

[amptramp]

My bet would be some sort of Sziklai Pair follower with perhaps some gain added to compensate for the small loss in a follower.  Also called the complementary Darlington, it is common in output stages of transistor amplifiers.  Another configuration might be the National Semiconductor LH0002CH complementary buffer with four transistors but I doubt they would use that.

[Rob Strand]

To me it looks like a JFET buffer with a current source load then that feeds an NPN transistor buffer.  The current source is formed with a transistor.

So they solved the JFET buffer attenuation problem I mentioned above by using current source load.  The JFET gives the high input impedance and the BJT gives the low output impedance.




Boss Waza Input Buffer V1.0.png

R1   don't know: 1k or 10k                                 ;10k seems more probable
C1   don't know: unpopulated, 10pF, 47pF, 150pF   ;?
C2   don't know: 47n, 100n, 470n, 1uF              ;470n, 1uF more probable

[Fancy Lime]

Hi Mick, hi Rob,

thanks for the pics and analysis! After staring at the thing and at the datasheets way too long, I concur  that the topology that Rob suggests seems the most likely although I wouldn't have been able to trace it myself from this glorious mess.

So, long story short: the only "improvement" over the standard JFET buffer is that the gain is almost 1 instead of 0.9 or whatever. That would show up as an obvious improvement when A-B testing Waza vs non-Waza, as I suspect is done excessively on Youtube (did not bother to check). In the real wrld, however, this does not solve any problems that would not be just as easily and probably better solved by putting a good transparent booster at the front of your pedal chain. So while I think its a good improvement in a commercial pedal, where components bought in boxes of tens of thousands are cheap and robotic population of boards is too, it's probably not worth the hassle for diy pedals. I'm going to breadboard it anyway and compare to a single-JFET buffer. This may become a nice little project for beginners. Not extremely necessary but certainly educational.

Thanks again for scratching my curiosity-itch! Also: I'm glad to see that Boss still keeps their innovations rooted in the real EE world, even if they are of minor benefit. They might as well just use Mojo components, charge twice the price and call it a day. That seems to be how most (or at least too many) companies do it these days.

Cheers,
Andy

[Fancy Lime]

Interesting that they used a 4558. I cannot think of a single reason to choose that over one of the myriad of lower-noise, better stability modern drop-in replacements. Other than Mojo and I really hoped that the whole Waza thing wouldn't just be Mojo-based.

Andy, it's a bucket brigade delay, a severely limited, thoroughly obsolete technology that wouldn't even exist any more if not for OCD guitarists with their vacuum tubes and their LM308s and such, and their reverence for instruments made by hand in the 1950s. I'm not saying the sound isn't worth it, but if it's not about mojo, really, what is it about? And in the larger scheme of things, does the difference between a 4558 and a 5532 matter all that much in an analog delay pedal?

Hi Ben,
you are not wrong, of course. On the one hand: 4558 or 5532 will make no real difference here. But on the other hand: 4558 or 5532 will make no real difference here. And that is kind of my point. A BBD delay is at least noticeably different from most digital delays (better or worse is a matter of taste and debate). Tubes have some genuine advantages, although it is again debatable if they still outweigh the drawbacks in 2019. An LM308 slammed into the rails sounds distinct and some people like that. But if you use a 4558 or any number of other opamps in a DM-2 does not really matter. Therefore, if I were in charge of buying parts for Boss, I would opt for buying a larger stock of an opamp that outperforms the 4558 in other applications, where it does matter. 1.010.000 5532's are probably cheaper than 1.000.000 5532's plus 10.000 4458's, no? Then again, what do I know. I have no experience in this sort of operation and there may just be a Roland warehouse somewhere with enough 4458's from the 80's to last them 'till doomsday and then some.

Andy

[Rob Strand]

only "improvement" over the standard JFET buffer

The other thing it does is it makes the buffer more linear.  Using a current source load on a buffer to get less loss and more linearity is an old trick.  You will see it used in HiFi stuff.

FWIW, the SD-1W uses a slightly different circuit but the main ideas are still present.

[Fancy Lime]

only "improvement" over the standard JFET buffer

The other thing it does is it makes the buffer more linear.  Using a current source load on a buffer to get less loss and more linearity is an old trick.  You will see it used in HiFi stuff.

FWIW, the SD-1W uses a slightly different circuit but the main ideas are still present.

You don't happen to have a schematic for that as well, do you?

Andy