Spring reverb

[scaesic]

I'm considering adding a reverb unit to my diy amp.

I was considering the type 8 accutronics model. I was also considering the simple op-amp schematic to drive it, on the accutronics web site.

Has anyone built this? Is the stage centre reverb much better? Is there any other manufacturers of spring verb units? I really need a small one as it's to fit in a 10" combo practice amp.

Also, how do you figure out which impedence to use?

[scaesic]

no one have any experiences?

[Mark Hammer]

There is nothing particularly magical about the SCR except that it was one of the few widely posted reverb projects in the early days of DIY.  I don't know that it is a better design than many others.  It does the basics, providing a way to drive the springs, recover the signal and mix dry and wet together.  I've suggested a number of mods over the years, and so far no one has written back and said they sucked. (High praise indeed, eh?).  If anything, it probably earns a bare passing grade because it doesn't seem to provide for feeding the driver transducer the sort of current it really wants and needs.  The circuit found here seems a much better one to consider than the SCR: http://solorb.com/elect/musiccirc/reverb2/

Obviously the circuits that Accutronics provide themselves will be suitable for driving and recovering the pan.  What probably makes the biggest difference in sonic quality in the pan will be the number of springs.  More springs = a more difuse sound with fewer obvious resonances.  The type 9 has essentially 3 transmission lines, but each line is two linked springs, each with its own characteristics, so it has the "best" tone.  Certainly, the type 8 presents the best compromise in terms of size and tone.  The shorter springs will not sound as rich as longer compound springs, but at least there are 3 of them instead of 2.

If you're building it yourself, then you can adapt the support circuitry to suit the pan, so there is really no basis for selecting one of the type 8 pans over the other.  The impedance properties of the coils at the end of the springs pale in comparison to the mechanical properties of the springs themselves.

The big choice you probably need to make is where to tap the reverb send signal from and where to blend it back in.  Personally, I am a fan of tapping the reverb from the cleanest possible point in the amplifier's internal signal path, and blending it back in at the last possible opportunity before the power stage.  Part of this reasoning is that: a) the cleaner the signal is going to the pan, the cleaner it is coming out, b) the more dynamics you have to the signal, the more you can use the picking dynamics to adjust the feel/tone of the reverb, and c) reverb works best when it sits in the background and does not obscure the primary signal, and having a tonally different reverb signal makes it easier to mentally position in the background.

[scaesic]

There is nothing particularly magical about the SCR except that it was one of the few widely posted reverb projects in the early days of DIY.  I don't know that it is a better design than many others.  It does the basics, providing a way to drive the springs, recover the signal and mix dry and wet together.  I've suggested a number of mods over the years, and so far no one has written back and said they sucked. (High praise indeed, eh?).  If anything, it probably earns a bare passing grade because it doesn't seem to provide for feeding the driver transducer the sort of current it really wants and needs.  The circuit found here seems a much better one to consider than the SCR: http://solorb.com/elect/musiccirc/reverb2/

Obviously the circuits that Accutronics provide themselves will be suitable for driving and recovering the pan.  What probably makes the biggest difference in sonic quality in the pan will be the number of springs.  More springs = a more difuse sound with fewer obvious resonances.  The type 9 has essentially 3 transmission lines, but each line is two linked springs, each with its own characteristics, so it has the "best" tone.  Certainly, the type 8 presents the best compromise in terms of size and tone.  The shorter springs will not sound as rich as longer compound springs, but at least there are 3 of them instead of 2.

If you're building it yourself, then you can adapt the support circuitry to suit the pan, so there is really no basis for selecting one of the type 8 pans over the other.  The impedance properties of the coils at the end of the springs pale in comparison to the mechanical properties of the springs themselves.

The big choice you probably need to make is where to tap the reverb send signal from and where to blend it back in.  Personally, I am a fan of tapping the reverb from the cleanest possible point in the amplifier's internal signal path, and blending it back in at the last possible opportunity before the power stage.  Part of this reasoning is that: a) the cleaner the signal is going to the pan, the cleaner it is coming out, b) the more dynamics you have to the signal, the more you can use the picking dynamics to adjust the feel/tone of the reverb, and c) reverb works best when it sits in the background and does not obscure the primary signal, and having a tonally different reverb signal makes it easier to mentally position in the background.

Cool, the thing i noticed about the scr was it had two pots - mix and dwell - the more control over the reverb the better, itd be amazing to control the time delay of the verb, but this is physically impossible with springs, so im looking for lots of control.

i had been thinkign abotu where to take the signal from and where to feed it back in, i think after the initial pre-amp (mosfet boost) but before the tone shaping section, that way id be able to sculpt the tone of the verb.

The schem on http://solorb.com/elect/musiccirc/reverb2/ could presumably be adapted easily for this by removing the preamp schem and the clipping detectors (i dont see the need for these)?

Also, whats the reason for the trafo on that schem? the scr doesnt use one?

pcb or strip layouts of any of these schems would be really useful.

[Mark Hammer]

What has been labelled as the "Dwell" control on the SCR is actually NOT the same thing as the Dwell control on the old Fender Spring Reverb unit.  The true use of the term "dwell" refers to how hard the input to the springs is driven; that is, the gain in the driver stage.  What is shown in the Anderton schematic as "dwell" is simply the reverb level after it is recovered.  When part of an amplifier, the "reverb" control on the front panel is identical to what is called '"dwell" in the Anderton schematic.  It simply blends in some amount of the reverb signal with the primary signal.  Since Craig Anderton had no idea if you wanted the reverb output to be blended with what you fed the reverb springs or with something entirely different (e.g., a fuzzed/wah-wahed signal), or whether you wanted the dry signal on one channel and the wet on another, he included separate dry ("mix") and wet ("dwell") level controls.  If you only want a reverb signal at the output of the SCR, then you just turn down the mix control all the way.

On the classic fender reverb unit, you can control how hard you push the springs (dwell), the tone of what you pick up from the springs (tone), and the blend between what you recover from the springs and the dry signal (mix).
http://www.ampwares.com/ffg/schem/reverb_6g15_schem.gif

In fact, a true "dwell" control is exactly what you are looking for.  The harder you smack the springs, the longer they continue wiggling.  Smack them softly or lightly, and they only wiggle for a brief instant.  By varying the dwell, the gain in the recovery stage and the mix levels, you can achieve the same reverb level, but with different tonal and decay characteristics.  If you were to build the SCR ( http://www.generalguitargadgets.com/diagrams/stage_center_reverb_sc.gif ), then the real dwell control would be something where R7 is.  For example, a 100k fixed resistor in series with a 1M pot, as a replacement for R7, would provide a gain range of just under 5x to 50x, as opposed to the stock gain of 21.4x.  That would enable you to drive the springs softly or brutally.  Note that increasing the value of R7 interacts with C1 to produce a lower treble rolloff.  Anyone using this circuit with the suggested mod may wish to drop C1 down to maybe 150pf or even 100pf.

Also, whats the reason for the trafo on that schem? the scr doesnt use one?

The NE5532 and LM833 are both recommended in the Forrest Cook schematic because they are comfortable driving low-impedance loads such as those provided by speakers and reverb transducers.   Cook uses a pair in parallel to provide greater drive current (which the Accutronics schematics try to do as well). The transformer simply provides a not-quite-so-low load to the chips, but much lower load to the input transducer, so that more modest drive current from the chips is turned/converted into much higher current at the springs.  I imagine he also did this because he was taking his feed for the clipping indicator from the output of the driver chips and using a high-to-low transformer would keep the 1458 half near the top from messing up things.  It also places less strain on the driver chip, whether 5532 or 833.

As always, you are encouraged to read this old, but still pertinent, article by Craig O'Donnell: http://img.photobucket.com/albums/v474/mhammer/Reverb1.png
http://img.photobucket.com/albums/v474/mhammer/Reverb2.png

[scaesic]

using only the reverb driver and recovery part and output level part?

will it be fine to run it off 15V? as long as all the caps are rated well above it? The reason is i already have a +15v in the amp, but +12v would require another reg.

also, before tonestack, or after?