DOD 680 MN3205 retrofit

[lars]

I recently purchased a DOD 680 analog delay that had the typical dead SAD4096. Unfortunately these chips are just about unobtainium these days, unless you want to pay about $150 for one. These are my findings for subbing in the much more reasonably priced and available MN3205:
My first goal was to maintain as much as the original circuit as possible. This meant driving the chip with the stock LF351N/4013BPC setup. This was the first challenge. The original setup uses both sides of the 4013 to drive the SAD4096. The solution to get an MN3205 to work is to disable the second set of Q outputs of the 4013. This can be done by socketing the chip and simply pulling the corresponding pins from the socket so they don't get soldered to the PCB.
The other very important step to take is to lower the voltage of the clock circuit. I used a 7808 voltage regulator both to power the MN3205 and provide power to the LF351/4013 clock driver, thus providing clock signals at a voltage the MN3205 can handle. This means you also want to socket the LF351. To provide the power, I simply bent the corresponding pins of the sockets for both the clock driver chips, and ran a bus wire between them. There is also a 47K resistor connected to the LF351 that normally is connected to 15v. I pulled one leg of this resistor and connected it to the 8v bus.
For the connection to the original SAD4096 socket to the daughter board for the MN3205, I pulled pins out of a corresponding socket and soldered jumper wires to them. This whole socket then acts like a molex plug into the original SAD4096 socket.
The daughter board I hand-drew and etched is based off of a circuit from this site: http://www.e-repair.org/restoration3.htm
A couple modifications were made to their circuit. By lowering the entire clock driver circuit to 8v, I didn't need to use any resistors to bias the voltage on pins 2 and 6 on the MN3205. Everything else is basically laid out exactly as their circuit is drawn.
On a side note, with the MN3205 I was able to increase the total delay of the circuit by changing the stock 820pf capacitor on the LF351 to 1n.
Here's how the finished project looks inside:

But who cares how it looks, how does it SOUND?
First of all, I have zero clock noise. Pretty much no noise of any kind really, the thing is dead-quiet, which I've always thought analog BBD-based delays should be expected to have some waterfall noise in the background. One point to note is that this unit does have the extra NE571 compander circuit, which I've noticed some 680's don't have. I would recommend getting a unit that does have this extra circuit.
This is an inspiring pedal to play. It has such a dark, mesmerizing quality to the delays. Especially when you turn up the "Repeat" knob to just before oscillation, you can get these awesome swells and distortions to the repeats.
I hope this write-up can help someone else to bring one of these back to life. It's well worth the effort. I'm sure a newer, PT2399-based delay can sound similar to this, but I don't know if anything can really match the feel of a vintage unit.

[GrindCustoms]

That is beautiful work, congrats getting it back to life!

[lincolnic]

Fantastic work! Looking forward to hearing a clip of this guy.