3 Tube Reverb:Is this a Missing Resistor in Your Schematics?

[caphead]

It is the area I circled in the schematic I posted. I have it shown as a trim pot as I actually built it to hear it's effectiveness. The reason I am pointing this LNFB resistor out is for 2 reasons; one is that on the numerous schematics I've seen for the Steel String Singer #002, only the hand-drawn one has this resistor indicated. Second, is how much improvement I heard when this trim pot was dialed to around 1.8k-2.4k. I don't have any test equipment to explain it in graphs or maths, so my verbal description is the best I can do. When the trim pot was at maximum resistance, 51k approximately, with the reverb send and return full up I could hear the sound of the reverb springs being hit too hard, like someone was lightly tapping on the reverb tank while I was playing. With the trim pot set to around 2.2k, that spring sound disappeared while retaining the same reverb depth as before. In conjunction to this test, I also altered the other LNFB loop from the plates to grid on the driver tube (2.2M/.047). All this RC combination did was alter the reverb depth and did almost nothing to combat the spring sound until the resistor value was 330k, at a great expense to depth/volume of the reverb. Doing the opposite, removing the 2.2M/.047 LNFB and using the trim pot set to a lower resistance, I heard a slight decrease in reverb depth but no spring sound. The only issue I had was when I turned the trim pot all the way down (to 820R), I got oscillation.

Like I said before, this was not a scientific test so take any and all opinions just as that, opinions. But I did my best to make it as objective as I could. I also cannot express a 100% certain opinion if that resistor is or isn't in the actual Dumble SSS #002 reverb, since I am not nearly as familiar with these circuits as many of you are.

[wjdunham]

That's a very good point. I've spent a lot of time tuning both that 2.2K resistor and the 2.2M/.05 NFB around the reverb driver to compensate for the tank boominess. I stuck with the 2.2K but lowered the 2.2M down to 1M and kept the cap at .05, which gives a good compromise on the depth while keeping the verb more natural sounding. This with a 9AB2C1B 3-spring medium decay tanks. The 2.2M I found a bit better with the 2-spring tanks. I also like a 180pf or 250pf bright cap on the Reverb Level pot, adds a little sparkle to the verb.

Bill

www.sebagosound.com

[caphead]

That's a very good point. I've spent a lot of time tuning both that 2.2K resistor and the 2.2M/.05 NFB around the reverb driver to compensate for the tank boominess. I stuck with the 2.2K but lowered the 2.2M down to 1M and kept the cap at .05, which gives a good compromise on the depth while keeping the verb more natural sounding. This with a 9AB2C1B 3-spring medium decay tanks. The 2.2M I found a bit better with the 2-spring tanks. I also like a 180pf or 250pf bright cap on the Reverb Level pot, adds a little sparkle to the verb.

Bill

www.sebagosound.com

Hey, you have the same initials I do, I am William J Doran, funny coincidence I suppose. Anyways, my impression of this reverb circuit is that it is overall very natural sounding and all those NFB loops are there to smooth out unnatural sounding elements within it. I used a 2 spring tank (the only good one I have), so my results were a little skewed as this type of tank needs a little more drive for full depth. This circuit I built is one of a few designs I've tested in pursuit of finding the best sounding reverb for my next amp. An observation I noticed with an earlier 3 tube design I came up with is that I got the best results shunting the high end on the send stages and peaking the treble on the return. So putting caps across the return level pot makes alot of sense to me, I will try that before stripping out the circuit and doing something else.

[wjdunham]

Hey, you have the same initials I do, I am William J Doran, funny coincidence I suppose. Anyways, my impression of this reverb circuit is that it is overall very natural sounding and all those NFB loops are there to smooth out unnatural sounding elements within it. I used a 2 spring tank (the only good one I have), so my results were a little skewed as this type of tank needs a little more drive for full depth. This circuit I built is one of a few designs I've tested in pursuit of finding the best sounding reverb for my next amp. An observation I noticed with an earlier 3 tube design I came up with is that I got the best results shunting the high end on the send stages and peaking the treble on the return. So putting caps across the return level pot makes alot of sense to me, I will try that before stripping out the circuit and doing something else.

Funny, we share two things, the same initials and the relentless pursuit of the best sounding reverb
Bill

www.sebagosound.com

[talbany]

I would call that the Tank Efficiency Control.. By shunting some signal to ground there your basically compensating for tank efficiency..The newer tanks IMO can be too efficient so slamming them with a ton of signal (current) they can get like Bill says real Boomy!!..The older Accutronics tanks are not that efficient part of the reason why they are more desirable IMO..Once you find the sweet spot for that tank (and that AT7) in that amp pop in a resistor and let it eat..Another way of doing the same thing (which I think is what HAD does) is to lower that resistor to ground after the reverb transformer..To me designing a better sounding spring reverb is like trying to design a chuck wagon to be more aerodynamic!..

Tony

[caphead]

Funny, we share two things, the same initials and the relentless pursuit of the best sounding reverb
Bill

www.sebagosound.com

Absolutely, and I don't care how many tubes it takes to make that happen! This amp I'm working on has 8 preamp tube sockets; I need 3 for the preamp, 1 for the FX loop driver and return, and obviously 1 for the PI. That means I've got 3 entire tubes to dedicate to just reverb.

I would call that the Tank Efficiency Control.. By shunting some signal to ground there your basically compensating for tank efficiency..The newer tanks IMO can be too efficient so slamming them with a ton of signal (current) they can get like Bill says real Boomy!!..The older Accutronics tanks are not that efficient part of the reason why they are more desirable IMO..Once you find the sweet spot for that tank (and that AT7) in that amp pop in a resistor and let it eat..Another way of doing the same thing (which I think is what HAD does) is to lower that resistor to ground after the reverb transformer..To me designing a better sounding spring reverb is like trying to design a chuck wagon to be more aerodynamic!..

Tony

If I got you correctly you are saying that tanks are more efficient today than they were before? Do you think it is better transducers or better physical design? The way I see it is that since the reverb driver is basically a single ended power amp, the same rules would apply as if you were powering a speaker. An amp without negative feedback is naturally going to sound both boomy and shrill at the same time. While alot of different reverb designs I have seen compensate for the excessive high end (usually via shunting caps), I've never seen one that keeps the low end in line aside from just using a small input capacitor. This I don't think really fixes the problem of too much low end reaching the tank. Think of the bright input on a Vox AC-30, it uses a 500p coupling cap after the input stage, but when the signal goes through the power amp section it certainly doesn't lack low end despite such extreme frequency attenuation early on. I think the concept of the NFB loop connecting from the reverb transformer secondary to the input stage cathode is really genius. As opposed to the numerous LNFB loops which couple from the plate to the input (or preceding stage's cathode) using a capacitor and large resistor, the 2.2k from the 4 ohm secondary feeds back practically all the way down to DC, where as the cap/resistor method has a roll off dependent on the size of that capacitor; everything below the AC frequency roll off of that cap does not get fed back. So thus unless an extremely large cap is used, the boomy low end frequencies could not become as well stabilized using that method alone.

I also experimented with the reverb return stage's input resistor, as shown in my drawing as a 1M trim pot. I found that it didn't take much resistance from ground to obtain full reverb depth, not much higher than 50k I guess (due to the parallel resistance of the tank's output transducer, I could not measure the pot in circuit). The only thing that happened when I turned that pot up higher was I got more noise. When I actually build this amp, that resistor will be somewhere in the range of 56k and 82k (I actually have a single 75k resistor in my collection I may use for this just to make things interesting lol!)

Although I can't say I am completely finished in my pursuit of the best sounding spring reverb, doing this circuit certainly pointed me in the right direction on where to go. So my hat goes off to HAD for this and many other things as well