Spring reverb tank

[JonasL]

Hi,

I'm wanting to make an external spring reverb unit with a spring reverb tank i've seen on ebay. (http://www.benl.ebay.be/itm/DIY-guitar-amp-reverb-tank-80s-NOS-Tokai-TA-35-made-Japan-limited-quantity-/250996691938?pt=UK_MusicalInstr_Amplifiers_RL&hash=item3a709193e2) The problem is: this tank has a DC resistance of 200 Ohm, which makes it more complicated. I've seen a drive circuit and a recovery circuit on this site : http://sound.westhost.com/articles/reverb.htm, but i have som questions about the schematics. (it are the second and the third schematics on the page) The drive circuits needs 35V en the recovery circuit needs 15V. How can I make it so the whole reverb unit uses only one power supply? And where goes the 35V and the 15V negative to? Also, in the first circuit two diodes are used, but they have no values. Which values should they have?

Thanx, Jonas

[Mark Hammer]

While Rod Eliot is a thoughtful designer (and patient exlainer), the voltages required may be considered  - while not exactly overkill - furnishing "comfortable" margins of error.  Perfectly acceptable spring reverb systems can be made using a simple single-ended 12vdc supply.  The key thing to remember is that the input transducer on the tank is essentially a low-efficiency speaker, so it needs a reasonable current to drive it.  This is why such circuits invariably use either a 386 (made to drive headphones and small speakers), or an LM833 or NE5532 dual op-amp (both made to handle low-impedance loads...like headphones or speakers).  By "reasonable current", I don't mean an amp or two, but I do mean something more robust than a 9v battery or a 6-pack of AA cells.

Something running off a 12v/500ma supply should be more than sufficient to do the job.

[JonasL]

Allright, and any idea about the diodes?

[Mark Hammer]

None.  And when I note that a decent reverb system can be had with a 12v single-ended supply, I do not mean that you can simply apply +12v to Rod's circuits and they will work.  I mean that there are other decent reverb circuits out there that will run well off +12v.

[JonasL]

Will this circuit work with my 200 ohm input resistance tank?

http://www.generalguitargadgets.com/index.php?option=com_content&task=view&id=43&Itemid=26

[Mark Hammer]

Yes, but you will want to change chips.  To the best of my knowledge, while the 4136 has decent noise specs (or at least what was considered decent for the time - circa 1980), it is not optimized to feed low-impedance loads.

You are recommended to go with either a pair of LM833 or NE5532 dual op-amp chips, or an LM837 quad op-amp chip (a pair of LM833s on a 14-pin dip).  Note that you will not be able to use any existing board layout intended for the 4136, although perfing this to accommodate a chip-change won't be especially difficult..

[Mark Hammer]

And DO NOT drive yourself crazy looking for 39uf caps, for heaven's sake.  A pair of 47Uf or 100uf caps will do just fine.

[JonasL]

Okay thanx, i'll keep that in mind. And i think i will use the LM837, that seems the easiest way to me.

[joelindsey]

The other day I built a reverb unit mostly following the schematics on the first link posted, (http://sound.westhost.com/project34.htm) except I adapted it for use with a single ended supply. If I recall correctly I put another 100k resistor on the non-inverting input going to +9v, making the voltage at that point 4.5v.

I also got rid of the mixing section, as I don't intend to use it as a stompbox in my pedal chain. It would sound very good, except it's terribly noisy with a great deal of hum. I suspect it is coming from the recovery section because adjusting the drive control for the lm386 does not affect the noise. Does anyone have any experience in this regard? The changes I made to the circuit in the link above were changes that I thought would work, not changes I knew would work. Didn't turn out so good.

[JonasL]

Yes, but you will want to change chips.  To the best of my knowledge, while the 4136 has decent noise specs (or at least what was considered decent for the time - circa 1980), it is not optimized to feed low-impedance loads.

You are recommended to go with either a pair of LM833 or NE5532 dual op-amp chips, or an LM837 quad op-amp chip (a pair of LM833s on a 14-pin dip).  Note that you will not be able to use any existing board layout intended for the 4136, although perfing this to accommodate a chip-change won't be especially difficult..

I guess this must sound stupid, but i'm not an electronics expert... Can I just use the LM837 instead of the 4136 (keeping in mind the different pinouts), or do i also need to chance the values of other components?

[Mark Hammer]

No it's a fair question.  Just think of it like switching from a stubby screwdriver to a long-handled one.  Same screws, same project, easier turning.

Pay attention to the pinouts, but everything else is fine.  I've posted a number of potential mods to the circuit which are searchable here, probably under "Stage Center reverb".  They're not compulsory make-or-break, just improvements to find the tone you want.  Other folks have posted them too, and there may even be some build comments over at GGG that can help out.

[PRR]

> drive circuits needs 35V en the recovery circuit needs 15V

Recovery is shown with +/-15V; 30V split in half either side of ground. Likewise the Fig 6 plan uses +/-35V (70V total) which is WAY over the top.

> this tank has a DC resistance of 200 Ohm, which makes it more complicated.

No. Use the Fig 5 circuit, it will be ample. In fact you can omit the D1 D2 R3 R4 Q1 Q2 R5 R6 buffer: '5532 has plenty of current to slam a 200DCR 1.5K nominal tank.

Millions of good guitar-amps use a single chip at +/-15V to drive nominal 1.5K (200DCR) tanks.

> idea about the diodes?

FWIW: the diodes may as well be 1N4007. Or 1N4001; but same-price in DIY and the '07 can be used many more places. You may also use 1N914 1N4842(??) small-signal diodes; but without heavy analysis the large-signal diodes are less likely to over-hot-bias the transistors. But you do not need the transistors (or diodes) to slam a 1.5K tank.

> Will this circuit work with

It's a pretty 1976 plan. Simple. Not fully optimized. The recovery amp is not as low-noise as it could be. The drive gain is high and not adjustable. It "booooings"; but if you turn it to "surf rock" it will be a bit harsh then pushed hard, and will hiss if you ever stop.

'4146 is now a collector's item. It _will_ drive 2K (and 1.5K) decently, as good as TL072 which is common in commercial amps with similar tanks. But unless you "must" use that PCB (laid out for 4146) you want to pin-out for TL072, TL074, some 5532, etc. LM837 is also a good chip, though un-fashionable for no good reason. (It attempted to ape the '5532 but as a Quad; layout is often easier with Duals and the 5532 was a known-good multi-source chip while the LM837 was single-source newbie.)

Also note the TWO 9V batteries. It runs +/-9V.

And unless that tank is exceptionally low-price for you, I'd really favor the "standard" tanks used in 95% of all 'verbbed guitar amps since the 1960s. Every few years someone produced a "non-standard" to shave a few bucks; few of these ever caught on. This crate of tanks seems to have been un-loved since the 1980s; I wonder why.

[JonasL]

Could you explain me what the -15v is and how i must use it? I've never worked with negative voltages before.

[artifus]

*edit* in my (admittedly limited) understanding it's very much like biasing an op amp with a resistor divider network but with a more reliable reference voltage - 0v. hopefully someone more knowledgeable than myself will chime in and correct me if i'm wrong.

[JonasL]

>No. Use the Fig 5 circuit, it will be ample. In fact you can omit the D1 D2 R3 R4 Q1 Q2 R5 R6 buffer: '5532 has plenty of current to slam a 200DCR 1.5K nominal tank.

When i leave the buffer, then to where go pin 8, 4 and 1 of the 5532?