Has anyone seen inside the new Boss Waza Craft DM-2w???

[Audiotexan]

I know they've only been out a couple of months, but figured possibly someone here might work at/for a Boss dealer and has taken a gander inside....

I only just found out last night that they were at the Summer NAMM show in July, and my curiosity is screaming to know whether they're using SMT, and what BBD as well.

https://www.youtube.com/watch?v=LIuHsUpWQ-s

I absolutely love my Carbon Copy so much that I'm selling off my old AD-9. But I've always wanted another DM-2 that I sold off for 'peanuts' back in the mid-late 80's.

[Mark Hammer]

There were two different issues of the DM-2: one using the earlier MN3005 chip, and a later issue using the MN3205 chip. While manufacturers other than Panasonic have resurrected the 3205 in both through-hole and SMD form, the 3005 is loooooooonnnnngggg gone.

What are the differences?

The 3205 series was essentially engineered to cope with problems arising from battery-power. All BBDs need to have a bias voltage applied to their input, such that the audio signal "rides in on" a DC voltage. If the bias is wrong you ether get a distorted signal (close to optimal bias), or no signal (bias way off). Trouble is that the preponderance of battery-operated BBD-based devices would take their bias from a divided-down version of the present battery voltage. And as the battery got worn down, the bias changed.

Panasonic "solved" the challenge by producing the 32xx series that could be powered off a lower supply voltage. The 9v battery would be fed to a 5V regulator to provide constant power to the BBD chip, and a stable voltage source from which to derive the bias voltage. So, as long as the battery had enough juice to power the entire audio circuit, the bias voltage would remain valid, and audio quality would not diminish. The regulator that dropped 9v down to 5V needed at least 7v going into it to work, but by the time a new battery has drained down to 7v, it won't be able to deliver enough current to the overall circuit anyway.

It was a very smart solution to a problem of its time. Bear in mind that in the early 1980s there were no power bricks or massive pedalboards that required them. So being concerned about problems arising from battery power were very germane.

HOWEVER...

The engineering change also meant that there were some small changes to noise level and headroom. Where a mains-powered MN3005 delay could run nicely off 15V, an MN3205-based delay could not.

Do they actually sound different? I have no opinion, and have never directly compared them. My feeling is that, by the time all that lowpass filtering is applied in the usual way, few audible differences will remain that can be reliably attributed to the chip and not to something else in the signal chain.

But if one is a purist, let it be said that, unless Boss/Roland has been sitting on a hidden secret stash of MN3005s, or unless their pockets are deep enough, and the perceived demand high enough, to justify the resurrection of the MN3005 die, the Wazacraft release is likely using a Coolaudio repro of the MN3205 and the circuit is somewhere between the late-issue DM-2 and the very similar DM-3. There is precious little, if anything at all, different in the audio path between the later-issue DM-2 and DM-3. The primary difference is in the subcircuit used to adjust the clock for setting the delay time. The DM-3 circuit is apparently more dialable.

[Audiotexan]

What has me curious is that the 'new' DM-2w has up to 600ms in 'custom' mode (where the 3005/3205 were limited to 300ms IIRC), and the Carbon Copy has up to 600ms as well -without a silly and inconveniently located switch!

That's what's got me wondering if they're using the same BBD chipset that MXR is using (since they're both claiming analog BBD). If they are, then I'm all set! (FWIW, the CC uses TI CD4047BM {which isn't technically papered as a BBD, but it's listed use is 'time delay'}, and SA571-D compander. Other than that, the only other chips I saw were op-amps.)

There's no need to spend double what I spent on the CC for a 'DM-2 re-issue' that's priced to drop the 'vintage' market, but still fleece the sheep (at $179)! I guess I shouldn't phrase it so harshly, but given they're only 'inspecting them in Japan' -according to the 'special' serial number stickers they took the time to make (assuming they're still being fabbed in Taiwan) then it seems earned.

We all knew it'd be a matter of time before they'd do this (or at least I figured as much) since they were getting to insane prices of $400+ for a 'minty-fresh' DM-2 (3005 chipped).

Is the Coolaudio repro of the MN3205 a 300ms or 600ms or...?

[Mark Hammer]

Back in the day, the chips were pricey. For whatever reasons, these days you can pick up a pair of Coolaudio 3205s for probably less than what a single one would have run you 30 years back. Since, with the advent of cheap digital chips like the PT2399, people have grown accustomed to having more delay capacity for low cost, manufacturers tend to have simply adopted the approach of packing a pair of 3205s, for double the delay time. Admittedly, 600ms is still a little limiting, but it seems to capture 85% of the needs/scenarios of most guitar players. If they need more, they'll opt for something digital.

As a technical aside, when you look at the specs for the corresponding clock chip (MN3102; 3101 if it came with a 3005), there are certain limits identified. A big part of that concerns the clock-pin input capacitance. Huh?

We are all familiar with "loading" of a guitar signal, such that it loses top end on the way to the amp. The high-frequency clock pulse from the clock generator can be considered like any other signal. In order for that clock-signal to turn the little FETs inside the BBD chip on and off with precision, the clock has to provide crisp square waves. Each 3205 has about 700pf of capacitance on the clock input pins. Stick a pair of them in a circuit, driven by the same clock generator, and they pose enough combined capacitance to be analogous to feeding your guitar to your amp through a 100ft cable without any buffering whatsoever. You wouldn't expect great top end for your guitar, and similarly, you would not expect those BBDs to be getting nice crisp square waves.

What happens? If the clock pulse rises a little slower, due to this "treble loss", and does not hit peak voltage right away, then you have tiny gaps between when each cell in the BBD gets told to do this, then do that. Think of it like discontinuous sampling, and the resulting decline in audio quality.

As a result, the datasheets for the MN3102 and 3101 will stipulate that they can only drive up to 8192 stages of delay (i.e., two MN3205s, four MN3208s, eight MN3207s, etc.). They CAN actually drive more if there is appropriate buffering in between the clock generator and BBDs, or if some other HF clock generator is used. That is no different than being able to feed a 100ft cable with no audible loss if you buffer it properly. But, since this would obligate a more complex circuit, and a product other than their chips, the datasheets simply depict the most compact package: one MN3102, and a pair of MN3205s.

And THAT, in a rather obese nutshell, is why you tend to see more and more analog delays with around 600ms delay, but not much more than that.