Experimental Reverb Unit

[Mellotron Man]

I'm in the process of designing and (eventually) building an experimental Delay unit, based on a string of 3 or 4 MN3207 Bucket Brigade Devices (BBDs). It is usual to use an MN3102 Clock Generator chip to drive the BBDs, but it is extremely difficult to reliably vary its clock frequency, so I intend to have a triangle wave oscillator driving the VCO in a 4046 then creating the 2-phase clock using a 4013. I'm looking at a load of old circuits in some British electronics magazines, trying to find an optimum circuit. I've had a look at a "BBD SOUND EFFECTS UNIT" in the April 1990 edition of ELEKTOR magazine via the worldradiohistory website, concentrating on the clock-driving section of the circuit diagram. The pdf file on the worldradiohistory is of poor quality, and I cannot make out the values or type numbers of some of the components in and around the clock. Is there anyone who has a copy of this magazine that can help me out?

Ref - https://worldradiohistory.com/UK/Elektor/90s/Elektor-1990-04.pdf The circuit is on page 22.

[Mark Hammer]

I suspect one of the reasons for the existence of the MN3011, beyond the simple convenience and space-saving aspects of having multiple taps from a single chip, was addressing the challenge of heterodyning.  Having multiple clocks on the same circuit board makes layout even more critical than usual, in order to avoid the various clocks interacting with each other.  The MN3011 only required one clock to generate a variety of time delays, circumventing that challenge.

In theory, if one was aiming for reverb, rather than discrete audible echoes, the PT2399 presents a more usable option.  This is probably why they are used in the various Belton Brick modules to achieve that goal, instead of multiple BBDs.  One of their advantages is that the clock is internal to the chip itself, rather than requiring clock paths to be laid out so as to avoid coming too near each other, making layout less of an issue.

Though the Belton Bricks don't explicitly address it or permit tinkering, in principle, use of feedback paths from different PT2399s - conceivably from different sources and to different mixing nodes - can create a complex set of reflections.  Beltons use 3 delay chips, apparently.

One of the things to keep in mind is that reverberation in the real physical world loses high-frequency energy over time, as things bounce off imperfectly reflective surfaces and objects.  That's not to say that a reverb signal that conserves such frequency content would/could not be musically useful or even pleasant.  But if the goal is mimicking a real physical space, keep that progressive filtering in mind.

[ElectricDruid]

There's all sorts of interesting things about that circuit, thanks for finding that!

(The +/-8V power supply, the LC filters, the chip itself  - MN3011 isn't that common)

There's a unusual "varicap" component used in the clock circuit, the BB204. I've never heard of it, or any of the other replacements suggested which seem too be from the same series. The circuit shows a way to add modulation to a MN-series clock, so I don't know why you say that's so difficult - you're looking at an example! The Boss CE-2 chorus is another that comes to mind, and there are lots more.

I think the reason for the MN3011 existing is that the taps are at unrelated points. Perhaps this is what Mark means about heterodyning - the only way to achieve this with fixed delay line lengths is to run multiple clocks, which causes problems. The MN3011 allows you to run a single clock and gives you six outputs that provide unrelated delay times.

That circuit is pretty versatile. Depending which tap you choose and how much modulation you apply, you'd be able to get a variety of chorus, slap-back, reverb-ish, and echo sounds.

One modification which I've seen on some delays from this period is to use variable filters so you get more bandwidth at shorter delays, and only reduce the bandwidth where necessary. In the 1980's this was done with voltage controlled filters like the CEM3320 (which is now available again in multiple versions, so this is an option again) but a more modern way would be to run switched capacitor filters from the same clock that the delay line clock is derived from.

[Ripthorn]

I designed a discrete pt2399 reverb circuit using for pt2399 and it sounds rather good, in my opinion. I spent a bunch of time tweaking it. BBD circuits are good and wonderful, but that seems like an overly complicated and expensive circuit nowadays. Back in 1990 it was probably state of the art for DIY.

[Mellotron Man]

Thanks for the responses, but I was just looking at the clock circuit - the rest of the circuit isn't of THAT much interest! I've looked at other BBD circuits, and none of them have inductors or varicap diodes.
Taking some of the points mentioned chronologically...
1) I am aware of the PT2399. I am also in the process of making 3 reverb units based on this device - one for my Mellotron and 2 for mates in Stafford. The circuit came from 320volt.com. I have to say that I don't fully understand what a PT2399 does nor how it works. It has a 44kB memory, but all the documentation bangs on about modulation and de-modulation. I don't get it! Surely, if you want to interface with digital memory, you use ADCs and DACs. As a side issue, there is a new PT device with 48kB of memory, the PT2396.
2) I'm aware of CEM3320 VCF chips. They, along with other synth chips (VCA , VCO & others) were made by Curtiss, but are now knocked out by a company out of Latvia called Alfa. Oddly enough, I bought some AS3320's at the same time I got the PT2399's.

I've had some people saying I should take a look at Digital Signal Processing (DSP) - not a chance! From what I can understand of DSP, it's all under the control of a microprocessor (or micro-controller - I don't know the difference). I was introduced to MPUs (specifically the Motorola 6802) in the early 1980's and I was bloody useless with them. I still am. I'm not interested. I want to stick with Electronics, not computing. </rant>

[Rob Strand]

There's a unusual "varicap" component used in the clock circuit, the BB204. I've never heard of it, or any of the other replacements suggested which seem too be from the same series.

They are basically reverse diodes, often in dual packages to reduce distortion.   The varicaps in the project are FM types.  You might be able to replace those with normal diodes or zeners .   Do a google search,  some of the RF guys have pages showing capacitance vs voltage for common diodes and zeners aimed at replacing varicaps.    The hardest parameter to match for varicaps is the capacitance ratio over a given (and smallish) voltage range.  You can parallel diodes to up the capacitance.  (There's varicaps for AM applications as well, these are impossible to replace with common diodes.)

The use of LCR filters in at that circuit is a bit funky.

As far as getting a better copy of the schematic.   The project may have appeared in other European versions of Elektor, like Elektuur.   You might be able to dig-up a copy of those.   You might have to search a few months before and after the English version as the projects don't always appear in the same month (sometimes they don't appear at all).


March 1990, see magazine page 72, pdf page 50,
https://archive.org/details/Elektuur31719903/page/n1/mode/2up

[Mellotron Man]

Rob Strand, thanks for the schematic - that's exactly what I wanted. I did do a search and found Elektuur, but I didn't look at other months, later thinking that it was unrelated to Elektor. I also found the Indian version of Elektor, but I couldn't find the issue with the schematic.