The Roland MKS-20 brought the sound engine of the RD-1000 stage piano into a MIDI-controlled rack module. The RD-1000, launched in 1986, marked the beginning of Roland’s RD stage-piano series and became closely associated with Elton John’s live performances. Its module counterpart offered access to the same piano technology from an external MIDI keyboard, controller or sequencer, without the keyboard instrument’s physical scale. The Roland MKS-20 consequently belonged to a particular strand of Roland’s digital-piano development: a focused instrument built around piano and related keyboard sounds, with its own programmable tonal shaping and performance functions.
At its centre was Roland’s Structured Adaptive system, usually abbreviated to SA. This digital sound-generation approach reproduced changes in tone associated with the note played and the strength of the player’s touch. Rather than providing a large catalogue of unrelated instruments, the Roland MKS-20 concentrated on eight basic voices: Piano 1, Piano 2, Piano 3, Harpsichord, Clavi, Vibraphone, E. Piano 1 and E. Piano 2. Its 16-note polyphony supported chordal playing and sustained passages within that voice limit, while velocity-sensitive MIDI reception allowed an external keyboard to supply the player’s dynamic control.
The memory structure combined eight protected preset settings in Bank 1 with 56 programmable settings across Banks 2 to 8. Each basic voice therefore had seven additional variations, giving 64 selectable settings in total. These were variations of the eight underlying voices, rather than 64 different sound engines. The presets could be edited temporarily, but their original settings returned after switching the unit off and on. Edits in the seven user banks were retained by the battery-backed memory, allowing individual combinations of level, equalisation and effects to remain available between sessions.
Operation centred on eight voice-selection buttons, a Bank button and Roland’s α dial. Function mode selected the parameter to be examined, while Edit mode allowed its value to be changed. The backlit LCD displayed the selected voice, bank or parameter, with an adjustable contrast control to suit the viewing position. Separate chorus and tremolo buttons provided direct effect switching, and the MIDI message indicator confirmed incoming MIDI activity. This arrangement brought sound selection and the principal editing functions onto the module itself.
Each setting could have its own voice level and three-band equalisation. The bass and treble sections used shelving filters at 100Hz and 10kHz, while the middle section provided eight selectable centre frequencies from 400Hz to 4kHz and an adjustable bandwidth setting. All three bands offered boost or cut in 1.5dB steps across a ±10.5dB range. Voice level was separately adjustable from −13dB to +2dB in 1dB steps, referenced to the corresponding Bank 1 preset. These controls allowed variations to differ in both tonal balance and output level while retaining the identity of the underlying voice.
Chorus and tremolo extended the editing beyond equalisation. Each effect offered independently adjustable rate and depth, with 15 settings for each parameter, as well as an on/off setting. Their configurations formed part of the editable tone settings, so recalling a variation could also recall its effect treatment. Together, the equaliser and modulation effects provided the means to prepare contrasting versions of a piano, electric piano or other keyboard voice without relying entirely on external processing.
MIDI reception could be assigned to any of the 16 channels, with Omni operation also available. Program changes selected the 64 stored settings, providing remote access to the same bank-and-voice structure used on the front panel. Reception of program changes and MIDI chorus/tremolo switching could be disabled when those messages were not wanted. The MIDI input accepted performance data, while MIDI Thru passed an unchanged copy of the incoming MIDI stream to another device; it was not a MIDI output for transmitting newly generated performance data.
Voice Preserve addressed the practical problem of changing sounds during a performance. With this function enabled, a MIDI program change could select the next sound while the current notes continued. The new setting took effect once all notes had been released and the damper was off. This deferred change applied to MIDI program changes rather than selections made with the front-panel voice buttons. Disabling Voice Preserve allowed a received program change to take effect immediately.
Tuning and register adjustment were also available independently of the main tone-editing controls. Master tuning covered A = 438Hz to 446Hz in 0.1Hz steps, and octave shift moved received notes by up to two octaves in either direction. Chorus and tremolo were temporarily disabled in Tune mode and restored on leaving it. The octave-shift setting was retained after power-off, allowing the module’s register to remain matched to its controller or musical role.
An optional M-16C memory cartridge provided external storage for the edited settings of all eight banks. Save and load operations transferred the complete 64-setting bank arrangement, making the cartridge useful for backing up user programming or restoring an alternative set of variations. The internal memory’s battery-backed retention and the cartridge’s separate storage served different purposes: one preserved everyday edits inside the module, while the other provided a removable copy.
The audio connections accommodated both unbalanced and balanced systems. Left and right output jacks were accompanied by a three-position output-level switch, while separate left and right balanced XLR connections provided another route to a mixer or recording system. The level switch affected the unbalanced outputs rather than the XLR outputs. A front-panel stereo headphone socket allowed direct monitoring. The enclosure measured 90mm high, 480mm wide and 400mm deep, with a weight of 8kg.
The Roland MKS-20’s continuing identity was tied to this particular combination of SA piano voices, programmable equalisation and onboard effects. Roland later included recreations of both the Roland RD-1000 and Roland MKS-20 in the RD-2000, maintaining a connection between these earlier instruments and its later stage-piano range. Its historical interest therefore extended beyond being a keyboardless companion to the RD-1000: it preserved a distinct Roland approach to expressive digital piano and electric-piano sounds in a separately controlled module.
Features
- Structured Adaptive sound generation with touch-responsive keyboard voices.
- Programmable equalisation, chorus and tremolo for tonal variations.
- Eight protected preset settings and 56 programmable user settings.
- Storage of all 64 tone settings on an optional M-16C memory cartridge.
- Backlit LCD with adjustable contrast for operation in low light.
- Balanced XLR outputs for connection to compatible audio equipment.