mixing dry and wet signals on a spring reverb

[Marcos - Munky]

Last night I built a spring reverb driver (this one by mac) to use with a spare tank I have here (600R input, 2K2 output). It worked as I powered it up, but the sound was a little below unity (which is not really an issue) and I got bass loss.

My first bet on the bass loss was at the mixing part of both signals. So I disconnected them and tested them separately. Clean signal sounds fine, wet signal sounds fine, but when I connect them together I get bass loss. Then I noticed the schematic is missing the mixing resistors, which may be what's causing the bass loss. So I added a 1K mixing resistor to both wet and dry signal. Still got bass loss.

Any clues on what should I try next? I thought on trying larger value mixing resistors, but since I'm already getting a little volume drop, a larger resistor probably would give me a bigger drop.

[Mark Hammer]

One possibility is that the spring itself is not conducting bass.  Another is that the spring and/or piezo are inverting the phase, relative to the dry signal.

[amptramp]

It looks like the piezo is connected to the output of the dry signal directly meaning that it is connected to the low (on the order of an ohm) output impedance of the op amp that drives the dry signal.  The dry output completely swamps the spring output.  1K mixing resistors are way too low an impedance for the output of a piezo (which still sees 2K going to the output impedance of the dry signal output.  A piezo needs a high-impedance buffer and 1 megohm would be the lowest value of buffer input impedance - 10 megohm would be better.

[anotherjim]

Marcos said he has a normal spring line. Even so, it isn't really suitable. You still need a recovery amplifier from the spring output. If you want to persist - reduce the bass in the spring. Try reducing C3 or C5.

[Marcos - Munky]

One possibility is that the spring itself is not conducting bass.  Another is that the spring and/or piezo are inverting the phase, relative to the dry signal.

For the first, yeah, the wet signal isn't very bassy, it's probably related to the 386 driving a 600r coil, but I'm pretty satisfied on how it sounds. The dry signal sounds as bypassed signal. The low cut happens when I try to mix both. For the 2nd, didn't thought on that. It makes sense, signals with inverted phase will cancel, so maybe it's a phase issue. Based on the color of the wires, the tank is correctly connected, but I'll try to reverse the input or output and check how it goes.

It looks like the piezo is connected to the output of the dry signal directly meaning that it is connected to the low (on the order of an ohm) output impedance of the op amp that drives the dry signal.  The dry output completely swamps the spring output.  1K mixing resistors are way too low an impedance for the output of a piezo (which still sees 2K going to the output impedance of the dry signal output.  A piezo needs a high-impedance buffer and 1 megohm would be the lowest value of buffer input impedance - 10 megohm would be better.

Since it's a reverb tank, it's not a piezo like the schematic, but a coil. Still, 2K2 output impedance. As simple as a buffer is, I already have a board for the circuit (yeah, I don't really like breadboards and just go ahead and etch pcbs ), so I'm gonna try other things before making another board with the buffer. Maybe raising those resistors a little will help.

Marcos said he has a normal spring line. Even so, it isn't really suitable. You still need a recovery amplifier from the spring output. If you want to persist - reduce the bass in the spring. Try reducing C3 or C5.

Yeah, it's a reverb tank. If by recovery amplifier you mean a gain/volume recovery, I don't really think it's necessary. The spring output is a little below unity, which is pretty fine. The mixing of both signals is what's causing the bass loss. I'll try to play with values for C3 or/and C5.

[Rob Strand]

Not sure but there might have been a non-piezo version of the schematic, possibly in a another thread.

[anotherjim]

The recovery amp isn't necessarily just to provide gain. It must have a high enough input impedance that it doesn't load down the reverb output coil. If that's 2k2, then you want to feed it into at least 22k. The amp is also an opportunity to set the EQ of the reverb. The amp should have a low output impedance so it can drive the next stage.

[Marcos - Munky]

So it definitively looks like I need the buffer(s). Time to go to the breadboard (was trying to avoid to do that ). Thanks for the answers!