Analog flangers and other BBD devices

[POTL]

Hi everybody
I want to clarify some questions about modulation and delay again.
We all know that analog devices are pretty close to each other, but they have some structural differences that I want to ask about.
1) Almost all effects using BBD have Pre Emphasis & De Emphasis filters in their design. I have 2 questions about them
A) DOD and A / DA Flanger and EHX DMM either ignore them, or use asymmetric solutions, I wondered, are they really needed?
B) Gain factor is different for some devices, BOSS and EHX have the same gain, MXR Flanger has more gain. Am I correct in saying that the higher the gain, the better the signal-to-noise ratio? Or MXR has an older design (this is clear from the diagram) and EHX and BOSS solutions are optimal?
2) Noise reduction. Almost all coassic circuits and a modern boutique use NE571 and its analogs, but among some circuits a forging circuit with a detector is embedded, I'm talking about A / DA Flanger and BOSS CE-1, which will be better in 2021, a detector (which I suspect is an outdated solution) or NE571 and analogues?
3) Clock - MN3101 / 3102 or 4047 or 4013.
the first can be seen at BOSS, Ibanez / Maxon, DOD
4047 at A / DA
4013 for MXR and EHX
A) Why is the voltage supplied to the BBD always lower than the signal voltage in Flanger circuits with mn3101 / 3102? (usually 5 or 6 volts versus 9). At that time, the neoiginal clock works in the full range (9 volts or 15 if the circuit requires it).
B) In my opinion, the 4013 has the best and fullest sound, but when the modulation reaches the upper and lower peaks (at great depth), it becomes uneven and harsh, as if the wave is not quite triangular, A / DA on 4047 will learn well, but at high modulation depth, the sound is unusual and a little unpleasant (usually this is characteristic of EHX Small Clone and DMM, which also use 4007).
The MN3101 / 3102 sound very nice, but there is no volume and the modulation depth is very small. I am aware that the problem is long latency and non-original clocks can create less time. I thought, if you add 4049/4069 to the original watch as in the case of 4013, can that reduce the delay time?
Or is it impossible to obtain a beautiful and magnificent flange with the help of original clocks?

[Mark Hammer]

That's a lot of questions, so you'll forgive me if I don't try to answer all of them.

The need for pre/de-emphasis as a means of noise-control will depend on the range of clock frequencies used.  If the clock frequency descends down into the audio range (e.g., 12khz), then some form of noise-control will be needed, even if the sweep keeps the clock frequency well above human hearing most of the time.

"Asymmetric" solutions, that use gates on the wet signal only, presume that no special treatment of the dry signal is needed.

Different BBDs have different passive signal-loss specs, such that some BBD-based circuits (mostly EHX from what I can see) include gain stages between BBDs.

[danfrank]

Ha ha someone beat me to it. Oh well...

Hi!
1a) the faster the clock speed and /or the less BBD stages in an effect, the less need for emphasis/deemphasis circuitry. Notice that most flangers have little signal modification where as analog delays need it. A/DA has fast clock speeds and either a 512 or 1024 stage BBD. The faster the clock speeds and less BBD stages, the less degraded the signal becomes after passing through the BBD. The one caveat though is the faster the clock speed, the more attenuated the signal becomes, passing through the BBD.
1b) the bigger the signal going through the circuit, without overloading the circuit, the better the S/N ratio.

2)
See 1a above. Same principals apply. Remember, the more filtering (noise reduction included) in a circuit, the less "natural" it will sound. All of these things are not perfect. The designer has to do a lot of compromising in order to get to what he/she thinks sound best. Audio is VERY subjective...

4)
Clock signal (the square waves) needs to be at the same voltage as what is applied to the BBD IC for optimal results. IE. If a BBD is supplied with 15 volts, then the clock square waves need to be 15 volts.
The 3101 and 3102 clocks are an expensive way to drive the BBD clock pins. They work well up to 100-200khz, sometimes faster speeds. The 4047 & 4013 are more versital in application but need to have a buffer after it in order to drive a BBD successfully at higher speeds. These ICs have been used along with a 4049 buffer to drive BBDs in the mHz range. They work very well, much better than the 3101 & 3102 ICs. Remember, the clock pulses need to be very square even at the highest clock speeds in order for the BBD to work correctly.
The 4007 is not a clock, it is used as a variable resistance in order to change the clock frequency of the clock IC.

Hope this helps.

[POTL]

Thanks for answers. ignore 4007, this is my fault, I mean 4047. I will correct the first message so as not to mislead anyone. For 4047 and buffers, EHX ignores them, just like Emphasis filters in DMM. Other than that, the design of the EHX circuits is as weird as the sound of their pedals

[POTL]

in fact I thought the buffer could improve the 3101/3102 range, but it looks like it won't be enough to get to 4047 or 4013. I liked the sound of JAM Waterfall, this is a CE-2 with vibrato and capacitor capacitance mods, I thought it would be cool to build a flanger that is close to it in design, but apparently the native clocks won't help me. Chase Bliss Specter and TC Thunderstorm use native clocks and sound boring. I also heard Chase Bliss Tonal Recall which is a hybrid of DM-2 and DMM, it uses native clocks and its modulation is better than DMM.
It is definitely worth playing with the capacitor values ​​of the 4047 and if its sound continues to be unpleasant for me, then I will dig in the direction of 4013. It definitely works well in flanger and chorus and might work in a delay circuit with modulation (it works separately).

[POTL]

And further about the voltage you understood me not quite right. I meant that the LFO, Clocks and 3007/3207 operate from 6 volts, and the rest of the circuit from 9. this is only found in flangers schemes that use the original clocks. I don't understand why this is done.

[DrAlx]

Regarding running the BBD and clock chips at regulated supply less than 9V, the only reason I can think of is that as battery starts to lose voltage, everything will still work as before without needing to be rebiased. If you are not running on batteries you should run the BBD at as large voltage as it supports so you have more headroom in the BBD, and noise performance of the BBD is typically improved also (see graphs in MN32** datasheets).

[Mark Hammer]

That's why the MN30xx series can make a comeback.  In the early-to-mid '80s we were all still relying mostly on batteries, and no one wishing to try out a BBD-based effect in a music store would be handed a pedal and wallwart and directed to an available outlet.  So, as near as anyone can tell, the MN32xx series was devised to allow for a set-and-forget bias, based on +9v regulated down to a stable +5V.  The regulator needed to get at least +7VDC in order to work properly, but by the time the battery had dropped to that voltage, most of the rest of the audio path wouldn't be working anyway.

[POTL]

Okay, once again, thanks to good people for the answers;) let's summarize a little 1) The greater the Pre Emphasis gain and, accordingly, the greater the De Emphasis attenuation, the better the signal-to-noise ratio? (Of course, we try not to bring the signal to distortion). 2) The difference between the detector and the NE571 will be in sound and it is better to try it out. Do both work and both make sense? 3) MN3102 / 3101 + CD4049 / 4069 won't give me the same effect as CD4013 + CD4049 / 4069 or CD4047? The frequency will be limited to 100-200 Hz and I will not be able to reach Megahertz? Attempts will be in vain, right? Mark, what did you say about the MN30 **? Xvive not releasing them anymore?

[Kevin Mitchell]

XVIVE is the only company that offers newly manufactured -15v type BBDs such as the MN3005. They also make clocks for these (MN3101)
The company coolaudio makes the 9v devices like the V3205 and a V3102 clock. These are more-or-less equivalent to the MN3205 and MN3102.

You're asking about frequencies in megahertz? No sense in that and the devices won't run at those rate.

Not just for me, but for everyone else... can you tell us your intent in a simplified format? Or are you just gathering information here? Are you trying to make your own Chorus/flanger/delay?

[ElectricDruid]

1A) Pre/De-emphasis is the most basic noise reduction you can do, and it's ok for what's it worth. If you've got something better, use it. If you want minimal parts, don't bother. It's a judgement call whether "it's really needed".

1B) Gain. You want to run the BBD at close to its maximum input level for best S/N ratio, but what gain that involves at the front of the circuit depends on the details of the rest of it, so there isn't one answer. It also depends what your approach to distortion is and how much you consider "tolerable".

2) NE571 companders and similar. It's another noise reduction approach, arguably one step up from the basic pre/de-emphasis. While the noise reduction might be improved, there are other side-effects on the sound that might be regarded as problematic, or might be regarded as part of the vintage character of the thing.

3A) +1 what everyone else said. If you're running on batteries, it makes sense to run the BBD on something safely lower than 9V.

3B) I think you're getting caught up in "which chips". More important is "what shape modulation". It doesn't really matter which chips you use - you're only producing square waves, after all - but what matters is the frequency range (lowest clock to highest clock freq) and the control voltage response to the LFO input. Is it linear, exponential, based on clock frequency or clock period? These factors have a *huge* difference to the eventual sound, since they change the way the delay time (and hence the pitch of the signal passing through it) responds to the modulation.

The original clock chips don't output enough current to drive the BBDs at high speeds, and were never intended to have modulation applied either, so they're far from ideal for the job of building the perfect analog flanger. Probably there's a way to make them do it, but I doubt it's worth it in terms of circuitry or work. Other equivalent implementations are probably simpler.

Xvive are still making clones of some of these chips:

https://www.xviveaudio.com/chips-p0064.html

I doubt you'd get them to clock at MHz. My experiments with the Panasonic originals suggested that the output levels drops off alarmingly above about 500-600KHz or so (which messes up the balance of wet and dry signals and kills the notch depth in a flanger) and I doubt the Xvive clones have improved on that, but then again, you never know. I'd love it if someone *did* produce a BBD with genuine "modern" specs, but I haven't seen much sign of it yet.

[Mark Hammer]

XVIVE is the only company that offers newly manufactured -15v type BBDs such as the MN3005. They also make clocks for these (MN3101)
The company coolaudio makes the 9v devices like the V3205 and a V3102 clock. These are more-or-less equivalent to the MN3205 and MN3102.

You're asking about frequencies in megahertz? No sense in that and the devices won't run at those rate.

Not just for me, but for everyone else... can you tell us your intent in a simplified format? Or are you just gathering information here? Are you trying to make your own Chorus/flanger/delay?

They CAN, and I've witnessed it.  The hurdle to clear is the input capacitance of the clock pins on the BBD, and having enough current drive in the clock to overcome that.  The stock 3101/3102 cannot, so other clock circuits are required if one wishes to reach those clock frequencies and still have any hope of decent audio passing through the BBD.

All of EHX's analog delays (Memory Boy, Memory Toy, et al.) rely on the Xvive chips as do many other commercial pedals, though I wouldn't hold out much hope for any of the various Asian-made under-$50 1590A pedals, that likely use a V3205 or equal.

[POTL]

Thank you all for your answers, once again I received a lot of necessary information. Speaking about Xvive, there are no microcircuits on their website for several months, the column is always "0" and the price is also "0", in Russia all suppliers do not have MN30 ** microcircuits, only MN32 **. I've heard a couple of times that Xvive is no longer making these ICs, Small-Bear's stock shows the minimum amount of MN30 ** versus MN32 **. I am inclined to assume that these chips are indeed part of history now, just like reticon.
Hopefully Behringer (CoolAudio) will reissue these chips and reticon. On the other hand, all modern pedals are based on MN32 ** and IMHO the same Jam Waterfall sounds better than the BOSS CE-2, maybe not everything is so bad.

[ElectricDruid]

All of EHX's analog delays (Memory Boy, Memory Toy, et al.) rely on the Xvive chips as do many other commercial pedals, though I wouldn't hold out much hope for any of the various Asian-made under-$50 1590A pedals, that likely use a V3205 or equal.

This is an interesting comment. I haven't directly compared the Xvive BBDs and the Coolaudio ones (I really should!) but the implication of this statement is that the clock input capacitance of the Xvive clones is lower/closer to the originals than the Coolaudio ones.
It's certainly true that the better you can get the clock waveform, the less losses you get through the delayline. If your clock starts getting rounded at higher frequencies, you'll see more losses, because the signal no longer cleanly switches on/off and passes to the next cell. Instead, it's a kind of crossfade, which degrades the signal as it goes along. So more clock drive is pretty much always a good idea.

I have compared the V3207 and original MN3207 directly, but only up to 500KHz, which is not really high enough to see the worst of the effects we're talking about. It's the outer limit of "normal" behaviour, but what we're discussing here is the "way off the datasheet" behaviour. Lol, typically... Trust us lot to push it to breaking point...

I did also once try building a analog delay with 4x V3205 (so 16K stages, and about 1.2 seconds max of delay) which demanded serious paralleled buffers from the clock to make it run. The clock input capacitance on so many stages of those BBDs is getting big! The noise performance was very poor with so many BBDs in series. Perhaps with even *more* serious clock buffers I could have made it better, but I wanted lower noise than the PT2399, and I wasn't getting it, so I gave it up. In the end, that project turned into the digidelay instead.