Digital/Analog Delay

[analogmike]

Nah.

2) For shorter delays, analog only.  When the BBD is "tapped out" (I'm not Catholic, but if I were, I'd muscle my way to the front of the line and confess this linguistic sin to the Pope's corpse right now! :lol: ), it then gets fed to the digital chip for added delay time.  There would tend to be either a blank spot in the delay range due to the limits on minimum delay possible from the digital chip, or else some overlap.  For instance, if you couldn't achieve a digital delay shorter than 30msec, an MN3005 would take you out to 350msec, and for anything 380msec or more you'd use an additional digital chip, OR, you'd have the choice of setting the BBD to less than its full delay, and adding some fixed amount of digital delay on top of that.

The new Maxon AD999 sounds like that. Too clear/digital sounding at some settings and analog at others. Does not seem to have a digital delay chip in it but I did not look too hard. Anyone seen these?

[puretube]

8 x MN3208, IIRC...

[Khas Evets]

From Maxon's site:

Each AD999 comes loaded with 8 ­ yes, 8 ­ Maxon MC4107D BBD's for a total of 900 milliseconds of pure, rich, organic analog tone.

[puretube]

:oops:  ooops, so my memory and math were offset...

so we could get 1.8 seconds of pure, rich, analog tone out of 8 `08s...

[puretube]

silly question  :oops: :
supposed you have a multiple tap BBD chip for reverb (3011),
and you want to implement a "predelay" option to the circuit...

would you feed the multible branches (taps) back to before the pre-delay section, or after (i.e.: feeding the taps back only to themselves)?

(in fact, this is an algorithm-question, which could also be answered concerning "virtual" software reverbs).

[Mark Hammer]

silly question  :oops: :
supposed you have a multiple tap BBD chip for reverb (3011),
and you want to implement a "predelay" option to the circuit...

would you feed the multible branches (taps) back to before the pre-delay section, or after (i.e.: feeding the taps back only to themselves)?

(in fact, this is an algorithm-question, which could also be answered concerning "virtual" software reverbs).

My sense is that it doesn't really matter.  Once you have the initial time stagger/pre-delay, whatever happens afterwards is simply a "flavour" of reverb.  Indeed, I would argue for a mixing stage between the first pre-delay BBD and whatever multi-tap device follows it, and feed the taps back to BOTH entry points.  If the objective is to mimic reverb, then I would think that whatever diversifies the "reflection times" adds to the realism.  Feeding a small bit of the tapped signal back to both the input and the in-between slot results in all those tapped delay times having iterations which are added to by both the offset stemming from the taps AND the offset stemming from the pre-delay.

Alternatively, depending on a) the amount of pre-delay and b) how complicated you want to get, you might also want to consider where each tap gets sent to.  So, for instance, early taps (e.g., 1-through-4) might get recycled back through the MN3011 with minimal lowpass filtering, while taps 5 and 6 get seriously "sanded down" and fed back through the pre-delay chip.  That would, in theory anyways, result in a burst of early bright reflections, (since the first 4 taps are abrupt and would be recycled through the shortest path to the output), and some longer-lasting mute reflections stemming from the longest-delay taps having the pre-delay time added to their recirculation time.

Incidentally, one of the the advantages of the MN3011 is that you don't HAVE to use all taps, or the longest tap, as the source of the final wet signal.  For instance, nothing prevents one from summing the first 3 tap outputs as the "wet" signal, while at the same time recirculating the longer-delay taps for some nice long lag in the reflections.  Keep in mind that even if it eventually comes out of the shortest-delay tap (396 stages), whatever has passed through 3328 stages of MN3011 PLUS 2048 stages of a BL3208 used as a predelay, PLUS those first 396 stages of the MN3011, has passed through 5772 stages before it ever shows up in the audio output for the first time.

Again, if you can stand the complexity, being able to route those longest taps back to the input to the pre-delay chip OR to the MN3011 itself, can yield a great deal of variation in sonic landscapes and textures.

Enough lateral thinking for you, Ton? :wink:  This is why I like the MN3011.  It is a wellspring of interesting ideas.

[puretube]

I wasn`t sure whether to ask in this thread, or rather in one of the 3011-threads  
(I could imagine a digi-pre-delay followed by a "reverb-block",
and regard latter as a "blackbox" - no matter if 3011, multi BBD chain or even a spring - for "room/ambience" purposes, and was wondering if it would sound more monotonous, if the reverb-cluster {burst} would repeat with the interval of the additional pre-delay, when fed back to the very input).

I`m afraid to have to breadboard such a scenario  :shock: , to do an A/B comparison - or just add a selector switch for both options (if the difference is drastic enough...).

The other question that occurred to me: since I don`t have an MN3011,
and would use say 4 MN3207s (1 predelay+3 taps) for a reverb:
how might it sound, when the length/time of the 3 taps are being modulated individually (best: asynchronous).
[let`s do the modulation slightly, not to get into "Through-Zero-Reverb" :lol: ].
Would this sound more "lively" or "natural", or would it sound like a crazy
"Dance-Verb"?
And again: yes, immediately an envelope-controlled modulated reverb came to my spoiled mind :wink: .

Anybody knows of such an effect having been done with some software,
or even whether it can be heard in a song?

(BTW: of course I`ve also played with the idea of modulating the relative amounts/amplitudes of feedback and summing the taps on a VCA basis,
for some "dispersion" - but that`d be even more complex, than just a little clock-frequency bending  :wink: ).

Mark: thanks for your thoughts - you done got me thinking!  

[StephenGiles]

I'm working on a through zero fuzz box right now :lol:  :lol:  :lol:
Stephen

[puretube]

here`s the "Through-Zero-Octivider", everybody`s been lurking -ooops-
looking for:
http://diystompboxes.com/sboxforum/viewtopic.php?p=225844#225844


(don`t ask me for a faillproof schem/layout/kit...)

[puretube]

I'm working on a through zero fuzz box right now :lol:  :lol:  :lol:
Stephen

I delayed that project for a while - too little feedback... :?

[Greybeard]

OK, maybe this has already been covered here or in another thread, so either point me in the right direction or something.......... I have been looking at analog (BBD) delays for awhile but just started messing with one yesterday. I usually, due to economic reasons, start by hacking into an old pedal instead of buying new parts. So here is where I am at. I have a Johnson EAD-2 with two 3208's in it. (3102 clock, SA571N compander) Can I feed part of the output from second one back into the first and get sort of a multi-tap effect that way? If so does it need to go through some sort of resistance/capacitence network first? I have data sheets, but the new Chinese devices are in an 8 pin package, where as the older units look to be in a 16 pin package with only 8 pins........... is the pinout the same? I have extended the delay somewhat but it seems to be unstable, some times it is long but the next time I turn it on it is short again.................. Also sometimes it seems to go into a noise feedback (kind of a crackling sound that gets worse) that is not generated by any input from the guitar and at that point the guitar sound gets very garbled, the only way to stop it is shutting down the power and firing it back up again...... could that be because they left out the trim pots?....... there is space and traces for them on the board, but they stuck a cap in there instead. Or is it a timing thing.......... elongating the time delay has made it noisier but not that bad actually. any light someone could shed on this would be helpful.


     I am not really a newbe, but sometimes I feel that way, I am just getting back into this after way to many years off.......... Things have changed so much in 20 years. .........  :wink:

Thanks,
Greybeard

[Mark Hammer]

Can I feed part of the output from second one back into the first and get sort of a multi-tap effect that way? If so does it need to go through some sort of resistance/capacitence network first? I have data sheets, but the new Chinese devices are in an 8 pin package, where as the older units look to be in a 16 pin package with only 8 pins........... is the pinout the same?

The two multi=tap BBDs that Matsushita/Panasonic produced had taps that were NOT harmonically related.  That is, none were the sum or product of any of the others.  In the case of single-tap BBDs, when you use a short delay and add some regen to try and mimic "reverb", you will note that the tone gets kind of boxey.  That boxeyness is a result from the emphasis produced when all repeats are multiples of the same basic delay time, and specific frequencies are reinforced.  When there are multiple unrelated taps, the repeats start to be more like true reverb where, unless one is sitting smack dab in the middle of a room with concrete walls, no furniture and no carpeting, the reflected copies of the original arrive at many different times, with few standing waves and points of emphasis (and please note the use of *few* as opposed to *none*).

This is the longish way of saying that if one used two successive BBDs that were ganged to the same clock and took a tap from their midpoint for half of the regen path, you would have a situation more comparable with the single tap scenario than with the multi-tap scenario, because a) there aren't nearly as many taps, and b) the delay time at the final output is a multiple of the midpoint delay.

Now, having said that, it strikes me that it MIGHT be theoretically possible to reduce the boxeyness by creative ue of filtering in the feedback path.  for instance, if the inter-chip tap underwent some highpass filtering to remove much of the fundamentals, and the output regen tap underwent some lowpass filtering to make it audibly different than the interchip tap, perhaps you *could* combine the two into a single regen signal to feed back to the starting point, without it sounding quite so boxey.  It wouldn't be as good as the nonrelated taps IMHO, but it would be better than simply feeding back the *entire* signal available at each BBD output to a common point.

The subject of "heterodyning" has come up here and/or elsewhere.  Ideally, if the two BBDs were run at different clock rates, then one could use a couple of identical chips to achieve multiple non-related delay times.  Here is where the heterodyning comes in.  Think of it like ring modulation at higher frequencies.  If one was clocking chip A at 20khz and chip B at 22khz, their (clocks) sum and difference would be 42khz and 2khz.  Should any of this additional clock signal be picked up in the audio path, the 42khz signal will obviously be outside of hearing range and likely well above the cutoff point for any filtering in the pedal and further along the signal chain.  The 2khz result, however, may be VERY audible, and the filtering needed to remove it would be too extreme to permit a decent audio signal at the output.  Or at least would restrict the bandwidth too much to provide much sonic flexibility.

As Ton drew to my attention, if the sum and/or difference is outside of the audio range, then the filtering burden is lifted.  So, if chip A was clocked at 23khz and chip B was clocked at 39khz, their sum would be well out of hearing range, and while their difference (16khz) would be within hearing range, it would likely be well above where the filters are set (most analog delays try to roll off highs above 4-5khz), making it much less of a potential problem.

As far as I know, the pinouts are the same for 8-pin and 16-pin packages.

[vanhansen]

I like original idea of this thread.  Worth trying.

When it comes to delay (which I rarely use now anway), I usually set it to 1 or 2 different settings, anywhere between 200ms to 350ms, low feedback.  For those settings, my old trusty DD-2 is perfect.  I never really messed around with "true" analog delay but love the sound of the DD-2.  I also have a Rocktron Intellifex LTD rack that has 24-bit sampling.  The delays and reverbs in it are outstanding.  I need to see if I can find the schematic to see what is in that thing.

So, yeah, I'd probably try something different with blend pot of some sorts (digital/analog or whatever it would be called).

[Greybeard]

Well that gives me something to think about, thanks. First time I looked at this Johnson unit I just thought..... all the right parts, just not the right sound. I'll see what I can come up with, the multi tap thing just sounded like a possible idea.

Greybeard

[puretube]

http://diystompboxes.com/sboxforum/viewtopic.php?p=236254#236254

[puretube]

http://www.toadworksusa.com/rd/whatsogreat.html
:?:

[cloudscapes]

I totally agree with you.

Interesting first post.

Interesting site in your signature too. If I visit meilleurtest.fr, it just looks like a generic referral storefront used by spam bots..