Anyone have a good reverb pedal design?

[BlackFlag1313]

I just got an old Silvertone 1481.  Super sweet amp (Killer Velvet Underground Sound) but no reverb.  I'm looking to build a reverb pedal of some sort.  Anyone have a good design for a spring or tube reverb in pedal form (it can be a very large or long enclosure)?  I'm also thinking about building a hardwood cabinet for it and would like to know if anyone has an idea about how to build a reverb into the new amp housing/amp circuit and how tough it would be.  I've never messed with any amp mods but I'm comfortable with pedal building.  Thanks in advance.

[Peter Snowberg]

Consider the Stage Center Reverb at General Guitar Gadgets (Link). The only problem with putting a spring reverb in a pedal is that the sound of stepping on the pedal board can crash the springs which isn't too musical. To combat this, I would suggest building the circuit in a small box like a Hammond 1590BB, and then using a set of RCA patch cords to connect the tank. That way you can put the tank on a piece of foam or in a vinyl bag (as is done inside amps).

The sound is a little different if you play loud because the reverb is added before the preamp, but it still sounds great. There's nothing like a spring reverb. If you want to emulate the amp sound a little more, add something like a tube driver (on a mild setting) in front of the reverb.

Take care,
-Peter

[Ansil]

checck out mark hammers page.

also i dont' knwo where you are located but the local hardware store has various sized springs that you could build your own diy reverb tank.

my favorite was to chain two reverb tanks together. for some serious swash.

[ExpAnonColin]

There's always BBDs  :wink:

-Colin

[BlackFlag1313]

What would you recommend for springs (size, characteristics, composition) for the Stage Center Spring Reverb?  It looks pretty easy to build - anyone have any experiences with it?

[Nasse]

I think getting good quality one may do the difference. Maybe Accutronics reverb tanks are most well known ones and available.

[Peter Snowberg]

I've never built that exact circuit, but have built several that were similar. The only tanks I've used have been Accutronics tanks, but that's what just about every amp company uses too.

The only problem with the Stage Center is that it uses a tank with a high impedance drive coil (310 ohms, the 4DB2C1D) which isn't too hard to find, but it's a short tank with a medium delay. While great for little amps, you may want a little more spring in the sound. Finding long tanks with 310 ohm coils is a bit harder.

I've used both the 9AB3C1B and 9AB2C1B tanks with great success, but these have a 10 ohm send coil impedance so you need to add one more chip to the design. Luckily, adding a single LM386 will take care of the impedance change and it's a very easy adaptation. The two tanks I just mentioned are 3 spring long tanks with either medium or long delay time.

This place is a good source:
http://www.tubesandmore.com/ (look under Guitar & Amp Items)

Accutronics Home Page:
http://www.accutronicsreverb.com/

Part Number descriptions:
http://www.accutronicsreverb.com/prodspec.htm

I hope that helps.

Take care,
-Peter

[smoguzbenjamin]

BBD's eh? Sounds like fun, how much does a BBD chip cost?

[V!N]

Is there a known BBD reverb project that works fine and sounds good ?

[Peter Snowberg]

BBDs create echos rather than true reverb. You can make a reverb with BBDs, but you need several of them to make a decent approximation. Any pedalhead needs both springs and BBDs.

The PT-80 at http://www.generalguitargadgets.com uses a digital chip that's similar to a BBD and I've heard good things about it.

The AD-3208 (also at GGG) uses real BBDs, but they're getting harder and harder to find.

Take care,are,are,are,are,are,are,are,
-Peter

[smoguzbenjamin]

I have a spring tank, but I'm not sure I want to us it because of the loud noisy crashing when they're moved... :?

[Peter Snowberg]

One source of BBDs is http://www.smallbearelec.com/

Look about 3/4 of the way down his stock list under the title ICs - Bucket Brigade Delays, Echo Processors, Vibrato

MN3205 for $15.95.
MN3102 for $3.95  (clock generator for the BBD above)

PT2399 for $3.95  (used in the PT-80 delay)


Smoguz,
You can always use 1 or 2 meter patch cables to place the tank out of harms way and on some foam. That should keep it from crashing.


Take care,
-Peter

[smoguzbenjamin]

Hmmm. The MN3002 looks like fun to play with. But aaaaaargh why do I live in holland?! 512 stage bbd.... cheap enough to replace if I melt it.... $4 clock.... Fun!

Damn. :evil: Why is Holland such a pain in the ass if you're a FX builder?

[Mark Hammer]

What makes BBD's such poor candidates for producing realistic sounding reverb is that they provide a single reflection time.  Turning up the regeneration of a short echo to produce reverb with a longer decay time simply results in emphasis at lower midrange frequencies because of that single reflection/delay time, and it sounds awful.  In contrast, as a mechanical device, springs generate multiple reflection times that are more like real reverberation in a physical space.

In the second and third generation of BBDs produced by Matsushita, they produced two chips (MN3011, MN3214) that were intended to solve this problem by having multiple non-harmonically related taps along the way.  Feed each of these outputs to a mixer stage, feed a little back to the input, and you'd get something a little more like reverb though I suspect the dearth of commercial pedals using these chips is an indication of how authentic or desirable this simulation was.  Most designs I've seen using the 3011 do not fully exploit the possibilities.  I have not seen any designs using the MN3214 so I can't say whether anyone smartened up in that regard.

One of the things that constrains both of these chips is that even though they have multiple outputs, they only have one input.
I started working out the conceptual details of a BBD-based reverb that tries to get around this.  The basic notion is to use a combination of feed-forward as well as feedback.

Think of it like this.  Three medium-delay BBDs (MN/BL3207 is a perfect candidate) are driven by a single ultra high frequency master clock, which is divided down in non-harmonically related ratios.  For example,  BBD-1 uses F-clock/2,  BBD-2 uses F-clock/3, and BB-3 uses F-clock/5.  Between BBD 1 and 2 is a mixer/splitter stage, and same goes for BBD 2 and 3.  Each mixer/splitter stage permits you to send a signal that can go to the next BBD or jump a step ahead.  It also lets you send signals back from the BBD-3 output to each of 3 insertion points.  Work out the math and you'll see that this quickly results in a wide range of multiple reflection and decay times.  A mix-level pot here and there, for both feed-forward and feedback, and a master clock rate pot, and you have a delicious amount of flexibility.  You can also get predelay out of it, which is kinda cool.

Of course, once you add up the cost of the BBDs and all the pots, knobs and chassis space, it isn't really any cheaper than ordering up a reverb pan from Accutronics and building any of the standard reliable designs for these.

[javacody]

I think I'm going to go with a lm386 version like Peter mentioned. One schem that looks nice is this one:



I'm wondering though, if I should use JFET's for buffers rather than transistors? I'm thinking it would get more highs through and give more of a warm, tubey sound?

[smoguzbenjamin]

Guess so, don't see a problem in using JFETs... Socket your transistors and mix 'n' match, see what sounds best

[Elektrojänis]

Think of it like this.  Three medium-delay BBDs (MN/BL3207 is a perfect candidate) are driven by a single ultra high frequency master clock, which is divided down in non-harmonically related ratios.  For example,  BBD-1 uses F-clock/2,  BBD-2 uses F-clock/3, and BB-3 uses F-clock/5.  Between BBD 1 and 2 is a mixer/splitter stage, and same goes for BBD 2 and 3.  Each mixer/splitter stage permits you to send a signal that can go to the next BBD or jump a step ahead.  It also lets you send signals back from the BBD-3 output to each of 3 insertion points.  Work out the math and you'll see that this quickly results in a wide range of multiple reflection and decay times.  A mix-level pot here and there, for both feed-forward and feedback, and a master clock rate pot, and you have a delicious amount of flexibility.  You can also get predelay out of it, which is kinda cool.

If someone decided to build something like that I think it would be interesting to try adding possibility for LFO-modulation of the clock like in choruses and flangers. I think that it might help to combat the separate delays problem if used mildly (even if it does not relly simulate the anything physical in reverb stuff). If it would not help in that then it would probably end up creating some really weird chorus/flange type sounds.

Hmmm... Maybe just modulating the delay time of one of the BBD's would be nice but it would need it's own clock then.

Oh well... It would be over complex circuit even without the modulation stuff.

[Mark Hammer]

Javacody,

I built that circuit and it works fine, although don't let the 9-12v supply info fool you.  This one needs more than a 9v battery for acceptable performance.  Think wall-wart and you'll be happier.

One of the things that can often improve the performance of a spring system is careful toneshaping, especially of the upper range.

You have a variety of points for modification in Cook's design.  One fairly conventional one is to simply drop the value of the various series caps leading exclusively to the springs, such as the various 0.1uf caps along the way.  You can drop the 100uf cap a bit too, to perhaps 47uf.

Another way to shift the frequency response upwards and keep more brightness than sproing is to use the pin 8-to-1 feedback loop in the 386.  The 10uf cap in there now provides gain across the full spectrum.  Dropping its value substantially will apply increased gain across only the upper portions of the spectrum.  Should one do this, I suspect that altering the 0.1uf caps along the way is less necessary, although if you look in your parts bin and only find .047 or .033, that shouldn't bother anything.

In general, one sets the "drive" control for a maximum distortion/sproing-free tone and simply uses the "wet" control exactly the way you would use the "reverb" control on an amp.

I strongly doubt that use of FETs would add anything useful to the design.  Th driver is a 386 so whatever any FETs ahead of it might contribute is washed ou by the 386.  The output impedance of most reverb pans is also low enough that the high input impedance of a FET won't really benefit anything.

Elektrojänis,

In the early 80's, A/DA produced a rackmount unit called the Stereo Tapped Delay (STD-1) which did exactly that.  It could be used for stereo flanging or chorus, with each of the taps from the MN3011 assignable to the left or right channel.  The only commercial endorser/user I know of it that I know of is Allan Holdsworth, if that means anything to you.  I never saw anyone else exploit that aspect of the various Matsushita multi-tap BBD chips.  Pity.

With respect to the hypothetical master-clock+divider arrangement I described, there is no reason why the divisors have to be fixed.  As long as they are essentially prime numbers or non-divisible by any of the others (e.g., 3, 5, 7 instead of 2, 3, 5) then the standing-wave and tin-can tone problem should be reduced.  Having the option to switch divisors would open up a broader range of tone colours.

[V!N]

Of course, once you add up the cost of the BBDs and all the pots, knobs and chassis space, it isn't really any cheaper than ordering up a reverb pan from Accutronics and building any of the standard reliable designs for these.

True .. but that 'pedal' you describe will have sounds in it you can't get from a reverb pan (and vise versa).

If only I had the knowledge to put the things mentioned by you into practice .. You always come up with the greatest ideas !

By the way, how is your leslie simulator coming along ? That most certainly is one of the best ideas ever.

[Mark Hammer]

Thanks V!N.  I'm actually quite intrigued with the multi-stage BBD thing.  If all a person wants is a little standard-sounding reverb (and who doesn't?), the reverb-pan is clearly the fastest route to it.  If you want to produce ambient spaces you've never heard before, then the multi-stage BBD thing is more in one's league.  

It strikes me that some "reverse reverb" could be achieved with it by feeding early reflections through pots to a mixer stage at the final output.  For example, let's say the three BBD's, through master-clock division, generate delay-times of 20, 30, and 50 msec.  In series, they produce a total delay of 100msec.  If you tap at each of the two interstage junctions, you also get delays of 20 and 50 (20+30msec).  Now send those taps to a final mixing stage but set the level of the earliest reflections lower than later reflections.  To "dress it up", stick a slow attack pedal like a Boss Slow Gear, Maestro Envelope Modifier, the E&MM String Damper, the Barbarello Envelope Modification Unit, or PAiA "Gator" in front of it and you can likely generate some of those nifty ghost-like sounds where the reverb seems to sneak up backwards.  Stick just a hint of feedback from the final mixer stage back to the input, and a but more to maybe the stage B/C junction to have it trail away.

Again, it is the option to have both selective feedforward and feedback that provides such flexibility.

As for any Leslie simulation, I scored a real Leslie mechanism on Saturday (8" full-range speaker with a single styrofoam "cheese wheel" rotor) so I'm going to wait for a bit until it is properly housed.  I'll listen closely and see what that can suggest in terms of how much and what to adjust to fake it by analog means.