All about the 140 ADSR

Overview Back in 1978 Roland introduced the System 100M which included a dual envelope generator (and LFO) module – the 140. Its famous for its snappy fast envelopes but it can also do longer timings as well. Its been cloned a few times but for 2023 and the AM8141 I wanted to get an accurate replica. The original timing is:

• 1.5ms to 7.5 s
• 4ms to 15 s
• Release 4ms to 15 s

Achieving these timings with the correct stage curves and “feel” turns out to be a lot harder than it looks, especially in Eurorack with its lower voltage power rails! Let’s look at the issues and how to solve them.

Slide Potentiometers Roland made use of 2M log slide potentiometers in the System 100M (and 700 and SH-5), which are not available commercially. The options are to either have some manufactured (Behringer) or to use 1M log and double the size of timing capacitor to 6.6uF (6.8uF is available). Eurorack replicas usually go down the 1M log route and sometimes change the timing capacitor. Various values have been used; 4.7uF, 10uF, but no one seems to go for the correct 6.8uF value, which means the timing is out.

The original 140 has a fixed minimum attack time created by a diode and a resistor to ground, the Decay and Release stages use a 330R resistor to set the minimum time to 4ms. When the stage timing potentiometers are halved this fixed timing remains but with twice the time due to the higher timing capacitor value. The Decay and Release resistors can be reduced but the Attack stage cannot. So a Eurorack replica usually ends up with the wrong minimum timings, and the timing capacitor is then reduced down to 4.7uF to get the minimum times.

When designing the AM8141 we wanted to be 100% accurate and have used custom manufactured 2MA slide potentiometers and the same 3.3uF timing capacitor, to ensure the timing of the stages is the same as the original.

The D Taper Roland used D taper potentiometers for the Attack, Decay and Release stages. This taper is still manufactured but called A05, it is a steeper curve than the more usual A15. The AM8141 uses custom slide potentiometers that have been manufactured with the A05 taper, to ensure its easier to set low value stage timings from the minimum to center of the slide potentiometer.

Attack Curve This is an important part of the design of an ADSR as its very audible to users, so we must get it right! The attack stage of most envelope generators aims to charge the timing capacitor to the +power rail. As a true exponential curve it can never reach this value, so a threshold voltage is set to ensure the stage /can end. Most designs use a value of 2/3rds of the power rail, in the case of the  Roland 140 this is +10V.

When replicating the circuit in Eurorack it is easy to make the mistake of changing resistor values so that the threshold (and therefore the output) remains at +10V. However this changes the shape  of the attack curve and the timing, as its 4/5ths of the power rail. I fell into this trap initially, as have Behringer with their 140 replica. The outcome is a minimum attack time increase and the top of the curve starts to flatten out. The extra duration can be compensated by reducing the timing capacitor but the curve remains the same, and is audible.

Some Eurorack replicas simply keep the resistors the same as the original and put up with the fact that the output is now +8V rather than +10V.  The AM8141 module takes an accurate approach by retaining the same resistor values and 2/3rds threshold but with Op Amp amplifiers to boost the output back up to +10.0V. This ensures the correct attack curve and envelope timing.

Outcome The AM8141 module is a replica of the prototype Roland 141 module, and it contains a 140 style ADSR, Gate Delay and Inverter Amp. The ADSR design follows the above rules; 2M log slide potentiometers with D tapers, the retained 2/3rd threshold and +10.0V outputs. It has the following envelope timings, just like the original;

• Attack = 1.5 ms to 7.5 seconds
• Deacy = 4ms to 15 seconds
• Release = 4ms to 15 seconds