Designing a reverb tank that is compatible with ALL amps...

Trying to design a "buffer" circuit that will allow a chosen reverb tank to be driven by any amp's drive/recovery circuit (ie. the big names: Fender, Marshall, Vox, Boogie, Orange, Peavey, etc). I really need this design to be passive if at all possible. If this isn't feasible/realistic I could use a 9v circuit if absolutely necessary.

I'm an amateur so please bear with me, but I'm thinking :

- Start with the most sensitive/easiest to drive tank.
- Use an input and output pot, or resistors selectable via rotary switch to attenuate the signals to/from the drive/recovery circuit as needed for each amp variety.

Obviously the main obstacle would be impedance mismatch, but as long as the signal never gets extremely anemic or extremely distorted I can live with a little extra noise from poor impedance matching. Watcha think? I don't have a shop right now, or a wide variety of amps to test. Would this work in a passive circuit without major problems? Again, I can live with a little noise, heat, inefficiency or even moderate signal degradation for some amp varieties as long as the overall SOUND is still acceptable.

Thanks fellas!!!

  The passive way of doing this would be a transformer, to match what the amp is designed to drive to what the tank is. Think this back and forth, so two different transformers with the required specs.

From accutronics we have
DIGIT #2 - INPUT IMPEDANCE
A = 8 Ohm
B = 150 Ohm
C = 200 Ohm
D = 250 Ohm
E = 600 Ohm
F = 1475 Ohm

DIGIT #3 - OUTPUT IMPEDANCE
A = 500 Ohm
B = 2250 Ohm
C = 10000 Ohm

  So you need a transformer with 14 pins for the input and 14 pins for the output. Ok, let's call 150Ω, 200Ω and 250Ω all 200Ω, that's just 10 leads. Meaning (identical pri and sec) coil start, 8Ω tap, 200Ω tap, 600Ω tap and 1475Ω end of coil. This way you connect the primary to what the amp expects to see, the secondary to what the tank actually is and you have perfect match.

  Simpler for the second transformer, pri:start, 500Ω tap, 2k25 end, sec:start, 100k end. You might need to reverse the transformer if you need step down. Of course if your impedances matches for start you don't need the transformer.

JS

one end of that tank is gonna be near a big power transformer.

if it is the hi Z side, them hum will be a problem.

so a tank with two low z transducers would be cool, but the pickup side of the tank has very little signal to deal with, so a hi z transducer will be required,

opamps have replaced tube drivers in may designs and they seem to sound alright as long as they scale the reverb pot so it does not sound like dick dale when set to 3,

ampeg makes an amp that has no reverb transformer but instead uses a cap,

what was the question? 

CJ, Joauins, thanks!

I considered an impedance matching transformer.....custom....sounds expensive though. Can you recommend someone/somewhere to could create such a thing?

CJ: Thought about using op-amp off 9v guitar pedal type power. But seems like that would get more complex than should be needed to just attenuate the signal. 

Are pots or resistors definitely a bad choice? Why? Again, I can deal with some inefficiency or audible "artifacts" as long as it sounds decent, is cheap and simple. If it's just going to sound terrible though, a custom xformer may be my best bet....especially if the price isn't prohibitive.

  I forgot to mention, while possible it was more like a theoretical solution, it's quite more practical to go with an active circuit, having a proper amp, with high input impedance and low output impedance adding some passive attenuators would make the matching for most cases, maybe the 8Ω needs special attention (extra buffer stage to drive it) And then the same thing at the receiver, from 500Ω to 10k, it's not hard to design a stage to fit there. The output, just an attenuator would meet the 10k:500Ω range without problem.

  Going passive is a problem, maybe you can solve the rid having a few standard transformers by hand and selecting the proper one as needed. For the second transformer finding a nice 600:10k should be easy enough and should work most of the time. For the first one, the 150:600:1350 is 1:2:3 transformer, pretty standard mic could do that part, maybe just a fixed 1:2 and used as an auto transformer if needed. The 8Ω is more tricky, you could probably wind a 600:8 one and call it a day.

JS

Thanks! Very good ideas!

Okay, looks like I may be in luck! I can eliminate quite a few of of the impedances that need to be covered by transformers:

- Eliminate F input since it's very rare, only used on Music Man / General amps.
- Eliminate ALL output impedance needs as all major brands use either 2250Ω or 2575Ω output tanks. I'll just use a 2575Ω tank to cover both.
- Can eliminate whatever the input impedance is of the actual tank itself (ie. the most sensitive/easiest to drive tank) as obviously it won't require a xformer to match up with the an amp that is meant to use that type. I'm guessing the most sensitive tank is probably the one with the highest impedance.

It turns out there are actually a few more impedances for type 8 & 9 tanks, but they are very close to and basically follow the numbers you posted for type 4. (useful info here: http://amprepairparts.com/reverb_numbering.htm )

So with a Fender type 8EB2C1B (800Ω input 2575Ω output) and a 2P3T switch:

Pos. 1 - For 600Ω-800Ω input range goes directly to the tank, no transformer needed.
Pos. 2 - 8:800 transformer to cover the 8-10 input range, 800Ω output goes to tank input.
Pos. 3 - 250:800 transformer to cover the 150-310 input range, 800Ω output goes to tank input.
              - Tank output is always wired directly to the recover circuit, no transformer needed.

Or is it possible to have a single transformer with 8Ω and 250Ω taps on the primary which would yield around 600-800 on the secondary? Would be sweet if I could get away with just one transformer, but that may be asking too much.

Does this all sound correct? Again, thanks VERY MUCH for your help with this! I feel like this is exactly what I needed!!!!

You could do a 8:800 and have a 250Ω tap in the 800 winding, so you use it as an autotransformer if needed.

JS

joaquins said:

You could do a 8:800 and have a 250Ω tap in the 800 winding, so you use it as an autotransformer if needed.

JS

Click to expand...

Do you mean a 250 tap on the 800 side? After 8:800, the only other one I'd need is 250:800. Do you think 250:that ratio would load down the autotransformer secondary too much or cause a voltage drop? That's not a huge step up, but I know autotransformers can be picky.

Thanks again for the great help!

jdurango said:

Trying to design a "buffer" circuit that will allow a chosen reverb tank to be driven by any amp's drive/recovery circuit (ie. the big names: Fender, Marshall, Vox, Boogie, Orange, Peavey, etc). I really need this design to be passive if at all possible. If this isn't feasible/realistic I could use a 9v circuit if absolutely necessary.

Click to expand...

This is simply NOT realistic. You could use a set of custom-designed multi-tap transformers, but that would cost more than a new reverb tank, and you already have enough magnetic hum-sensitive s**t in your guitar amp.
A very simple solid-state circuit would cater for most possible configurations.
On the receiving end, a 5534 works fine with the range of impedances.
On the drive side, you would need a "power" amp, capable of driving the 600  ohm coil with 20Vrms, which implies using a bridge configuration, and driving the 8 ohm coil with about 30 mArms. That could be achieved by implementing two circuits, on ewith a 5532 in bridge mode, the other with a single opamp with a pair of transistors, or a dedicated power IC. You would need a +/-15 V PSU.
That would leave the 1475 ohm out of the picture, since it requires a drive voltage that's out of range for most low-level SS designs.

Obviously the main obstacle would be impedance mismatch, but as long as the signal never gets extremely anemic or extremely distorted I can live with a little extra noise from poor impedance matching.

Click to expand...

There could very well be cases where the impedance mismatch is such that it won't be acceptable. Let's say the recovery circuit is a TL071 and the pick-up coli is 600 ohms, there's a loss of about 12dB in S/N ratio. Now if you drive a 1475 ohm coil with the 8 ohm output of a Twin-Reverb, that's a signal loss of about 30 dB! Clearly this wouldn't work.

Would this work in a passive circuit without major problems?

Click to expand...

You may get away with just connecting indiscriminately the pickup side, since the signal degradation is "acceptable", but using transformers is the only passive way you can drive with as much current or voltage as necessary.

Finally, are you aware of the Accutronics digital reverb modules?
http://www.accutronicsreverb.com/
They are designed exactly to do what you want. They require a 5V supply, that's all.
The only thing wrong is the "kick the amp" effect is missing. 

I'm not sure why you started a new thread? Look at the Gibson GA-1RT-1 circuit for basic. Look at guitar amp line out circuits for basic gain reduction.

> I'm not sure why you started a new thread?

+1

> I'm guessing the most sensitive tank is probably the one with the highest impedance.

No. They all have the same POWER sensitivity. They just split the copper winding to different ratios. You need to think Power. You need a Power Match (or design mis-match) to work well. 2X difference is no big deal.

As you deign to do your own amplifier, you must look at a transformer.

RadioShack Audio Output Transformer Catalog #: 2731380 $3.89

This is 1KCT:8 (they don't bother with specs anymore). It is ultra common in older pocket transistor radios, tho I'd hate for you to cut-up vintage radios. It is one of a line that is still in production in Asia.

The Shack part is good for 100mW above ~~300Hz. Most reverb drivers cut bass, and few are over 100mW, so this works OK. The standalone reverb box (and at least one production amp) do use 6F6/6V6/EL84 class drivers which may overload this part (garbled midbass). There are 400mW versions of this part.

> autotransformers can be picky.

Do not agree.

if you are going to design a tank then you will need nickel laminations and tiny magnets that sit on steel wires inside the flux area,

these laminations will have a coil for the input and a coil for the output,

the output coil is easy, just about every tank uses 2250 ohms as the output Z.

input Z?  we revisited the revisit reverb tank website and found that input Z might be

8  10  150  190  250  310  600  800  1475  or 1925 ohms.

you could eliminate some of these by grouping together like terms, so we might have

8  200  600 1600   

so the input transducer coil could have taps, but wait,  a Z of 1600 will take hair wire, will it handle the high current from an 8 ohm driver?  probably not. especially if you run a 058" E string.

so what you do is put two coils on the same leg of the nickel core,  coil 1 would cover 8 to 200  coil 2 would cover 600 to 1600. each coil would have a tap brought out to a jumper board. you woul need to have a custom lam stamped to make room for the extra coil but this would eliminate any add on transformer or active circuit solutions.

so now you have a tank that would replace 30 tanks.

maybe. what if they want 3 springs? or a short box for an Orange Rockereverb?  life ain't perfect, but hey, jus sayin, wtf, over?

CJ said:

if you are going to design a tank then you will need nickel laminations and tiny magnets that sit on steel wires inside the flux area,

these laminations will have a coil for the input and a coil for the output,

the output coil is easy, just about every tank uses 2250 ohms as the output Z.

input Z?  we revisited the revisit reverb tank website and found that input Z might be

8  10  150  190  250  310  600  800  1475  or 1925 ohms.

you could eliminate some of these by grouping together like terms, so we might have

8  200  600 1600   

so the input transducer coil could have taps, but wait,  a Z of 1600 will take hair wire, will it handle the high current from an 8 ohm driver?  probably not. especially if you run a 058" E string.

so what you do is put two coils on the same leg of the nickel core,  coil 1 would cover 8 to 200  coil 2 would cover 600 to 1600. each coil would have a tap brought out to a jumper board. you woul need to have a custom lam stamped to make room for the extra coil but this would eliminate any add on transformer or active circuit solutions.

so now you have a tank that would replace 30 tanks.

maybe. what if they want 3 springs? or a short box for an Orange Rockereverb?  life ain't perfect, but hey, jus sayin, wtf, over?

Click to expand...

All great ideas, but to make a long story short and spare you many reasons why, I really would prefer to stay passive if at all possible. Noise isn't an issue.

So I like where CJ is going with the custom transformer. Noise won't be a problem since this will be an external tank sitting far from the amp and housed in a steel box.

Also, again I don't need to accommodate the very rare F input impedance range (1475 &  1600)

So im just left with 3 ranges. One can be covered by the tank itself. The other 2 would need to be handled by the transformer.

Also, I don't quite understand what PRR is saying. All tanks have the same POWER sensitivity, but with different impedances? Wouldn't the impedance level impact signal "power" /load drawn from the driver circuit? Maybe I don't understand this relationship correctly. I know an 8 ohm speaker isn't necessarily more "sensitive" than a 2 ohm speaker, but it draws less wattage to move the same amount of air. Maybe the technically correct term I should use is most "efficient" instead of most "sensitive"? Either way, you catch my drift right? The highest impedance (600-800) will require the least power to drive (ie will be the easiest to drive), correct?

I just figured I should use the most sensitive tank since I can always pad the signal with passive components, but can't amplify a signal using passive components to drive a tank that isn't sensitive enough.

CJ is taking about a custom TANK, not transformer.

> I know an 8 ohm speaker ...than a 2 ohm ... it draws less wattage to move the same amount of air.

No.

Incandescent lamps all have (nearly) the same power sensitivity (light per watt). But different impedances, aimed for different voltages, all the same power. Say a 12 Watt light. This could be 12V 1A (12 Ohms), 120V 0.1A (1,200 Ohms), 240V 0.050A (4,800 Ohms), etc. Internally they all have the same weight of Tungsten. But the 12V job has a short fat filament, the 240V has a very long thin filament.

PRR said:

CJ is taking about a custom TANK, not transformer.

> I know an 8 ohm speaker ...than a 2 ohm ... it draws less wattage to move the same amount of air.

No.

Incandescent lamps all have (nearly) the same power sensitivity (light per watt). But different impedances, aimed for different voltages, all the same power. Say a 12 Watt light. This could be 12V 1A (12 Ohms), 120V 0.1A (1,200 Ohms), 240V 0.050A (4,800 Ohms), etc. Internally they all have the same weight of Tungsten. But the 12V job has a short fat filament, the 240V has a very long thin filament.

Click to expand...

Ah, okay, gotcha. So all reverb tank driver circuits can handle roughly the same overall wattage, but the relationship between V and I changes as R (or in this case Z) changes to properly match different circuit designs.....is that about right?

So although the driver circuits can all handle roughly the same power/load, they may operate on different voltages, hence the need to vary Z on different tanks, correct?

The transformer by it's very nature would maintain all these relationships as Z changes, so overall power consumption never really changes, just voltage and current proportionally.....

And as such it doesn't really matter what tank I choose as the "base" because none is really more sensitive or efficient than any other?

Is that about right? Sorry, again, I'm an amateur, still trying to wrap my brain around all this stuff and really fundamentally understand Ohm's Law. 

CJ said:

if you are going to design a tank then you will need nickel laminations and tiny magnets that sit on steel wires inside the flux area,

these laminations will have a coil for the input and a coil for the output,

the output coil is easy, just about every tank uses 2250 ohms as the output Z.

input Z?  we revisited the revisit reverb tank website and found that input Z might be

8  10  150  190  250  310  600  800  1475  or 1925 ohms.

you could eliminate some of these by grouping together like terms, so we might have

8  200  600 1600   

so the input transducer coil could have taps, but wait,  a Z of 1600 will take hair wire, will it handle the high current from an 8 ohm driver?  probably not. especially if you run a 058" E string.

so what you do is put two coils on the same leg of the nickel core,  coil 1 would cover 8 to 200  coil 2 would cover 600 to 1600. each coil would have a tap brought out to a jumper board. you woul need to have a custom lam stamped to make room for the extra coil but this would eliminate any add on transformer or active circuit solutions.

so now you have a tank that would replace 30 tanks.

maybe. what if they want 3 springs? or a short box for an Orange Rockereverb?  life ain't perfect, but hey, jus sayin, wtf, over?

Click to expand...

Ah I see, yeah I hadn't even considered a custom tank with different coil taps for different input Z's. I know winding custom a transformer or two and using a stock tank would be relatively simple. Not sure how comfortable I'd be messing with or modifying a delicate reverb tank transducer though.....hmmmmm. Definitely something I'll research. Thanks CJ!

jdurango said:

All great ideas, but to make a long story short and spare you many reasons why, I really would prefer to stay passive if at all possible. Noise isn't an issue.

Click to expand...

The simplest way would be to use a 70 volt transformer, common for large office and factory paging systems on every speaker to set relative levels. Set to 1 watt in a quiet office, use the 50 watt tap near the loading dock.

Google "70 volt transformers"

As an example, this may be a suitable one for 7 bucks USD. Or get two at that price, for more Z combinations.

http://www.outdoorspeakerdepot.com/70vditr2.html?cmp=googleproducts&kw=70vditr2&gclid=CKqLkNX1otICFQa2wAodduYMKw

Scroll down to "PDF owners manual", it lists all the impedances available in that model. Add a rotary selector switch to select Z to match the amp, done.

Gene

Gene Pink said:

The simplest way would be to use a 70 volt transformer, common for large office and factory paging systems on every speaker to set relative levels. Set to 1 watt in a quiet office, use the 50 watt tap near the loading dock.

Google "70 volt transformers"

As an example, this may be a suitable one for 7 bucks USD. Or get two at that price, for more Z combinations.

http://www.outdoorspeakerdepot.com/70vditr2.html?cmp=googleproducts&kw=70vditr2&gclid=CKqLkNX1otICFQa2wAodduYMKw

Scroll down to "PDF owners manual", it lists all the impedances available in that model. Add a rotary selector switch to select Z to match the amp, done.

Gene

Click to expand...

Wow this look to be EXACTLY what I need! I could use an 8 ohm tank as the base, fed by the  8 ohm tap on the secondary, then use the 250 and 500 (or 1000) taps on the primary to cover the 150-310 and 600-800 inputs! Obviously the 8-10 ohm range would go directly to the tank. Thanks so much Gene! This should do the trick perfectly!

PS: One more question, don't suppose there's any way to utilize both the 500 and 1k (or maybe 250 and 1k?) taps in some way so as to yield somewhere closer to 700 is there? Not a big deal, either 500 or 1k should work for the 600-800 range, but if there's an easy way to split the difference that would be even better! Thanks so much again! This just made my week! Gunna start messing with this ASAP!