Chuck D. Bones
Circuit Wizard
I picked up a Belton BTDR-3H from StompBoxParts recently with the idea that I would experiment with it. This breadboard is the result. It's not a mod to any existing design, although it has roots in the VFE Springboard and the EQD Levitation. The Belton reverb brick is a clever design containing three PT2399 delay chips, with feedback arranged to simulated the multiple reflections in a spring reverb. The delay of one of the PT2399s is varied slightly at about a 1Hz rate to give the reverb effect more life and make it sound less sterile. The variable delay creates a sort of chorus effect as well. Some people find the cyclic delay variation to be excessive. I'm one of those people.
U1A is a unity-gain input buffer. PT2399s can be a bit noisy, so I added treble pre-emphasis ahead of the brick and de-emphasis after. Spring reverbs typically have a bass-cut ahead of the reverb pan, which itself is a form of treble pre-emphasis. C2 performs that function here. The BODY pot (VR1) and C3 provide variable bypass around C2, adding some of all of the bass back in. The BODY control is more-or-less copied from the Levitation's TONE control, however the BODY control in this circuit has a much wider control range. U2A works with C2 to provide treble pre-emphasis, it mixes in the AMBIENCE feedback and provides extra gain to improve the overall signal/noise ratio. The two DECAY controls are copied from the Springboard. The LO-DECAY varies the decay of channel 1, the darker of the two delay channels. The HI-DECAY varies the decay of channel 2, the brighter of the two delay channels. The difference is subtle. The AMBIENCE control extends the decay of the treble content, adding some extra shimmer to the reverb. It was copied from the Levitation, however in this circuit it goes right up to the edge of self-oscillation. U2B amplifies the brick's output and performs the treble de-emphasis. The Springboard has Baxandall BASS & TREBLE controls after the brick, which I felt was a bit much. Instead, I opted for a BMP-style TONE control. The TONE control is flat at noon. It can be used to restore some of the bass lost before the brick or add brightness. The REVERB control varies the amount of wet signal fed to the mixer, U1B. The wet and dry mix is about equal around 3 or 4 (10:00 to 11:00). With REVERB at 10, the wet signal almost completely overwhelms the dry signal. U1B inverts the dry signal, which in my mind is no big deal since there is no reason to wrap an external blend control around this circuit. U4 provides regulated +5V to power the brick. Because the gains are low and the TL072 has good DC precision, I was able to DC-couple everything except the AMBIENCE control, saving quite a few 1uF coupling caps compared to the Levitation. The brick is AC-coupled internally at its input and output. Of course, the pedal's input and output have to be AC-coupled since the opamps are all biased to Vref.
I'm satisfied that this circuit does what I intended it to do. There might be some small component value tweaks. Does it sound as good as a real spring reverb? Nope. Is it useful as a guitar effect? I think so.
There are designs out there for digital reverbs using 2, 3 or 4 discrete PT2399s. The T60 by Brian Thorn is a good example. It corrects some of the Belton brick's shortcomings, particularly the delay modulation. There's no free lunch; doing it that way takes up a lot more board space than a Belton brick.
Control (L-R):
REVERB - TONE - AMBIENCE - HI-DECAY - L0-DECAY - BODY
All blue busses are ground except the very top one which is unused.
The bottom red is +5V on the left half, +9V on the right. On this proto-board, the busses are broken in the middle. I have jumpers on all of them joining the two halves except the bottom red bus.
The next higher red bus is Vref. The one above that is +9V. The top red bus is not used.
The signal flows pretty much from right to left. U1 is at the far right. U2 is near the middle. U4, the 5V reg, is left and below U2.
U1A is a unity-gain input buffer. PT2399s can be a bit noisy, so I added treble pre-emphasis ahead of the brick and de-emphasis after. Spring reverbs typically have a bass-cut ahead of the reverb pan, which itself is a form of treble pre-emphasis. C2 performs that function here. The BODY pot (VR1) and C3 provide variable bypass around C2, adding some of all of the bass back in. The BODY control is more-or-less copied from the Levitation's TONE control, however the BODY control in this circuit has a much wider control range. U2A works with C2 to provide treble pre-emphasis, it mixes in the AMBIENCE feedback and provides extra gain to improve the overall signal/noise ratio. The two DECAY controls are copied from the Springboard. The LO-DECAY varies the decay of channel 1, the darker of the two delay channels. The HI-DECAY varies the decay of channel 2, the brighter of the two delay channels. The difference is subtle. The AMBIENCE control extends the decay of the treble content, adding some extra shimmer to the reverb. It was copied from the Levitation, however in this circuit it goes right up to the edge of self-oscillation. U2B amplifies the brick's output and performs the treble de-emphasis. The Springboard has Baxandall BASS & TREBLE controls after the brick, which I felt was a bit much. Instead, I opted for a BMP-style TONE control. The TONE control is flat at noon. It can be used to restore some of the bass lost before the brick or add brightness. The REVERB control varies the amount of wet signal fed to the mixer, U1B. The wet and dry mix is about equal around 3 or 4 (10:00 to 11:00). With REVERB at 10, the wet signal almost completely overwhelms the dry signal. U1B inverts the dry signal, which in my mind is no big deal since there is no reason to wrap an external blend control around this circuit. U4 provides regulated +5V to power the brick. Because the gains are low and the TL072 has good DC precision, I was able to DC-couple everything except the AMBIENCE control, saving quite a few 1uF coupling caps compared to the Levitation. The brick is AC-coupled internally at its input and output. Of course, the pedal's input and output have to be AC-coupled since the opamps are all biased to Vref.
I'm satisfied that this circuit does what I intended it to do. There might be some small component value tweaks. Does it sound as good as a real spring reverb? Nope. Is it useful as a guitar effect? I think so.
There are designs out there for digital reverbs using 2, 3 or 4 discrete PT2399s. The T60 by Brian Thorn is a good example. It corrects some of the Belton brick's shortcomings, particularly the delay modulation. There's no free lunch; doing it that way takes up a lot more board space than a Belton brick.
Control (L-R):
REVERB - TONE - AMBIENCE - HI-DECAY - L0-DECAY - BODY
All blue busses are ground except the very top one which is unused.
The bottom red is +5V on the left half, +9V on the right. On this proto-board, the busses are broken in the middle. I have jumpers on all of them joining the two halves except the bottom red bus.
The next higher red bus is Vref. The one above that is +9V. The top red bus is not used.
The signal flows pretty much from right to left. U1 is at the far right. U2 is near the middle. U4, the 5V reg, is left and below U2.
