Spring reverb tanks - general questions

[überfuzz]

This weekend I gutted an amp thats been sitting on my shelf for a while. It turns out that there's a 3 springed reverb tank in it. I haven't really given it mush thought up until now, always gone for 2 springs. But as always, at bedtime my head was filled with thoughts and questions.

1. Tone
   Is it possible to hear a difference in different reverb tanks. I mean setup like 2 springs or 3 springs.
2. Schematics
   Every tank I've tinkered with has an output and an input. Why is that necessary? The signal is just jammed through the springs, why does the direction need to be arbitrary.
3. Sustain
   What factor(s) affect or determines the sustain of the reverb tank.

I seem to remember having more questions, but I guess it'll be easier to fade away this evening... If someone can enlighten me.

Edit: corrected some strange sentences

[teemuk]

Is it possible to hear a difference in different reverb tanks. I mean setup like 2 springs or 3 springs.

In short, yes. Generally, the more there are springs in the tank the "fuller" the reverberated tone is.

On the other hand, it's not as overly simple. Many of the 2-spring reverbs actually have 4-springs, two of these in series (so it easily looks like a two-spring reverb at first glance). There is a slight difference in their tensions, which then adds slightly different reverberation characteristics and creates a fuller sound. Then, some companies like Baldwin used just a single spring but they tapered it so that the tension was not uniform and again the tank had different characteristics than a tank with a plain single spring.

On top of that, the actual electronic driving arrangement can make a huge difference.

Every tank I've tinkered with has an output and an input. Why is that necessary?

My wild guess is that the driving transducer optimally has to have different characteristics and design than the transducer picking up the vibratation. Sort of like a loudspeaker can also act as a dynamic microphone, but it's not very good in that application - and vice versa.

What factor(s) affect or determines the sustain of the reverb tank.

Spring lenght, spring tension, the amount of springs, the driving electric current.

[tubeswell]

Every tank I've tinkered with has an output and an input. Why is that necessary? The signal is just jammed through the springs, why does the direction need to be arbitrary.

The input transducer is set to a particular input impedance, and the output transducer is set to a particular output impedance. Usually, you want the input impedance of the pan to give the optimum reflected load from the reverb transformer.

Or in the case of a transformerless reverb, you want the input impedance on the pan to be greater than (AFAICT = about at least 5 x the output impedance of) the stage that is driving the pan (so you don't loose too much signal strength when inputting a signal into the pan).

Similarly you want the output impedance of the pan to be about 5 x less than the input impedance of the recovery stage so you don't loose too much signal going out of the pan into the recovery stage.

[Ronsonic]

This weekend I gutted an amp thats been sitting on my shelf for a while. It turns out that there's a 3 springed reverb tank in it. I haven't really given it mush thought up until now, always gone for 2 springs. But as always, at bedtime my head was filled with thoughts and questions.

If these are the things keeping you up ....

Well, you found the right bunch of geeks to hang with.

[*]Tone
Is it possible to hear a difference in different reverb tanks. I mean setup like 2 springs or 3 springs.[/LIST]

Yes. The better the drive and recovery circuitry the more clear the difference. But there is indeed a reason for those different models.

[*]Schematics
Every tank I've tinkered with has an output and an input. Why is that necessary? The signal is just jammed through the springs, why does the direction need to be arbitrary.

It is a slightly different sort of transducer at each end. It would be possible to design a reverb driver / recovery transducer with matching impedances, etc. just wouldn't be as good. You could also use a phone cartridge as a mastering lathe head ... IYKWIMAITTYD.

Part of how the signal gets "jammed through the springs" is that they are twisted instead of pulled. That was part of the invention, rotating the springs gives a far better reverb simulation than just tugging on them.

[*]Sustain
What factor(s) affect or determines the sustain of the reverb tank.

The design of the tank itself, length of the springs, self-damping, how hard they're driven. The common "Dwell" knob varies the signal strength driving the pan.

I seem to remember having more questions, but I guess it'll be easier to fade away this evening... If someone can enlighten me.

Sleep well.

[R.G.]

There is a critical hidden item in reverb tanks. The damping material supporting the springs in the ends of the spring holder is critical.

Accutronics went through a disaster period back when I was having my amp mass produced. The Accutronics tanks sounded ugly, very ugly. Some heart to heart talks with them revealed that they could no longer get the material they had used for decades for this. They were working on a good replacement, but shipping tanks with a different material in the meanwhile.

We changed vendors. It was better.

Accutronics may be past this issue now. Don't know.

[Enzo]

Why are there separate input and output ends? Someone mentioned it above. The two ends do different jobs. Your PA system has two ends - ione end drives the speaker, and the other end accepts the signal input.

The input transducer has to drive the signal into the springs, much as a speaker drives the signal into the air. The output end transducer merely has to pick up the signal from the spring, much like a microphone. The two ends do very different jobs, so they are made different.

[überfuzz]

Thanx all!

I understand that the output and input does different jobs. The thing I have harder to grasp is why the spring tank has an out/in as well. I guess I just have to accept that, for now.

[Steve Conner]

I think the springs themselves aren't directional. You could swap the input and output transducers, and it would sound the same.

It's just that the two transducers need to be different to interface with the electronics.

[Enzo]

Keep in mind your wires do not connect to the springs, a transducer at each end interfaces the mechanical motion of the spring to the electronics of the amp. The different inputs and outputs I mention ARE the ends of the spring assembly.

One end of the spring has to be shaken around, the other end has to sense that motion. Just like microphones and loudspeakers have different finctions, so do your spring transducers. As someone mentioned, you can use a loudspeaker as a microphone, in fact some pros mount a big speaker in front of their kivk frum and use the signal it generates as if it were a mic. And while I fon;t recommend it, if you run the output of an amplifier into a microphone, the poor little diaphragm in it would try to make sound as if it were a small speaker.

Each of those things can do the other's job, but they mostly do it poorly that way. The transducer designed to shake the springs doesn;t do as good a job sensing their motion, and the transducer that senses the spring motion doesn;t do a very good job if used to drive the spring.

Steve is right, the spring itself doesn;t care which end is which, but the transucers at each end of the spring are different.

[Wombaticus]

Or in the case of a transformerless reverb, you want the input impedance on the pan to be greater than (AFAICT = about at least 5 x the output impedance of) the stage that is driving the pan (so you don't loose too much signal strength when inputting a signal into the pan).

Similarly you want the output impedance of the pan to be about 5 x less than the input impedance of the recovery stage so you don't loose too much signal going out of the pan into the recovery stage.

I could be mistaken, but when you wish to maximize *current* transfer, you *match* the impedances between stages (i.e. make them the same), but when you wish to maximize *voltage* transfer, then you really want to do impedance bridging, which is the proper term for making the input impedance much larger than the output impedance of the driving stage (as in the 5x rule you mention).
Since reverb tanks are current controlled devices, while valve grids are voltage controlled, it makes sense to me that you would match the driver circuit's impedance to that of the tanks, but then do bridging to the recovery stage.... someone please correct me if I don't have this quite right.

[Wombaticus]

I could be mistaken, but when you wish to maximize *current* transfer, you *match* the impedances between stages (i.e. make them the same), but when you wish to maximize *voltage* transfer, then you really want to do impedance bridging, which is the proper term for making the input impedance much larger than the output impedance of the driving stage (as in the 5x rule you mention).
Since reverb tanks are current controlled devices, while valve grids are voltage controlled, it makes sense to me that you would match the driver circuit's impedance to that of the tanks, but then do bridging to the recovery stage.... someone please correct me if I don't have this quite right.

Er, strictly speaking, impedance matching maximizes POWER rather than current transfer, but I think the general idea still holds....

[martin manning]

I understand that the output and input does different jobs. The thing I have harder to grasp is why the spring tank has an out/in as well. I guess I just have to accept that, for now.

I'm not exactly sure what is not clear yet, but the thing is just an electro-mechanical analog simulating "real" reverb. Think of the input transducer as your voice, the springs as the air and surfaces in a room, and the output transducer as your ears. The echo effect you hear in room reverb is a result of the time delay, multiple copies, and damping of the original sound wave as reflected in the returning waves. The springs, the air surrounding them, and their attachments mimic these effects using mechanical vibrations in place of sound waves, and the transducers are just the driver (voice) and receiver (ears).

[tubeswell]

I could be mistaken, but when you wish to maximize *current* transfer, you *match* the impedances between stages (i.e. make them the same), but when you wish to maximize *voltage* transfer, then you really want to do impedance bridging, which is the proper term for making the input impedance much larger than the output impedance of the driving stage (as in the 5x rule you mention).
Since reverb tanks are current controlled devices, while valve grids are voltage controlled, it makes sense to me that you would match the driver circuit's impedance to that of the tanks, but then do bridging to the recovery stage.... someone please correct me if I don't have this quite right.

yes I think that's right for the input impedance of the pan, but the output impedance is just driving a straight voltage signal at the recovery stage's grid, so bridging would be more important AFAICT. No doubt someone will chime in