I built this 284 for myself.
Here is the rear of PCB1. Since this design has low currents, I chose to not use any sockets so I could more thoroughly clean under the ICs.
PCB2 has a jumper for internal/external +24V and 8 trimmers.
|TR1||Channel 1 Delay/Duration|
|TR2||Channel 1 Attack/Decay|
|TR3||Channel 2 Delay/Duration|
|TR4||Channel 2 Attack/Decay|
|TR5||Channel 3 Delay/Duration|
|TR6||Channel 3 Attack/Decay|
|TR7||Channel 4 Delay/Duration|
|TR8||Channel 4 Attack/Decay|
I set delay, sustain, and decay to 0 and attack to 10S. I could then trigger a channel and easily measure the length of the attack phase and calibrate it at 10 seconds. Then I set attack, sustain to 0, delay to 10S and decay to something short. I could then trigger a channel and measure the delay and calibrate it to 10 seconds.
With all four channels cycling and set to internal 24V, +15V was 160 mA and -15V was 75 mA.
Here are some scope images of various waveforms. This scope image shows Attack-Decay with no Delay, Duration, or Sustain
This scope image adds Duration to show Attack-Duration-Decay with no Delay or Sustain
This scope image adds Sustain to show Attack-Sustain-Decay with no Delay
This scope image shows Delay-Attack-Decay with no Duration or Sustain.
This scope image adds Duration to show Delay-Attack-Duration-Decay with no Sustain.
This scope image adds Sustain to show Delay-Attack-Sustain-Decay.
I spent extra care in building this 284. With the CV modification of the maximum times, I knew there would be very low currents and any leakage on the PCB would cause issues. I used cut tape transistors for matching, organic solder with frequent cleaning, and no sockets so I could clean around the legs of the ICs (which you can't do with sockets). The results are a very impressive maximum attack duration of over 45 minutes! I set delay, duration, and decay to minimum, attack to maximum, and used a +10V CV. I connected all four channels together with a single trigger and monitored the results. To me this is impressive for an analog delay and I don't know why anyone would want such a long envelope. Assuming the other three segments have the same duration, this would result in a 3 hour envelope!
The 284 is an interesting design. It is comprised of two AR envelope generators operated in a loop fashion. The first envelope generator is for the delay and duration and the second envelope generator is for the CV output.
At the end of attack of the first envelope generator it triggers the second envelope generator and remains in a hold. At the end of attack of the second envelope generator it releases the hold of the first envelope generator (to start the decay) and remains in a hold. At the end of decay of the first envelope generator it releases the hold of the second envelope generator (to start the decay).
Since the second envelope generator has not started the attack, the attack of the first envelope generator creates the delay. Since the second envelope generator holds at the end of the attack, the decay of the first envelope generator creates the duration.
This scope image shows the two envelope generators. Cyan is the output of the first envelope generator at R35 (+ input of the op-amp). Magenta is the second envelope generator at R56 (+ input of the op-amp).
The sustain switch sets the decay of the first envelope generator to 0. This scope image shows sustain on but with a narrow trigger so the envelope consists of Delay-Attack-Decay
This scope image shows sustain on with a wide trigger so the envelope consists of Delay-Attack-Sustain-Decay
The CV input simply biases the dual pairs so more or less current is steered to the integrators.