Generalmusic (established in 1987 as an Italian musical instrument manufacturing company), launched in 1990, the GEM WS2 initially proved to create some very mixed feelings with musicians (and music shops), due to the innovative (even futuristic), features, along with 128 programmable and excellent voices, including the multi-sampled Grand Piano (with 3 dynamic levels).
Branching out into the professional synthesisers in 1992 with the GEM S2 and S3. The S2 was a 61-note keyboard and the S3 was 76-note keyboard. They did release the S4 88 note version but this was a special.
Both the S2 and S3 models had an excellent semi-weighted keyboard complete with touch/release velocity, polyphonic aftertouch and metal contacts. The front panel included seven programmable sliders, seven programmable buttons and Pitch and Modulation wheels. All functions were logically and clearly displayed on a large graphical LCD with light blue neon back-lighting.
Whilst these synthesisers gained an enormous Worldwide following (even to this day), in part due to the excellent depth of sound quality, performances, DSP effects, resonant filters (Lowpass, Highpass, Bandpass, Parameter Cut, or Parameter Boost), and multi-samples, two oscillators per voice, they did lack polyphony (16-note polyphonic), and some other nice to have options.
Generalmusic listened and shortly after, an upgraded version, aptly named S2 Turbo, S3 Turbo, was released in 1994. This added many new features and improvements such as:
- Tone generation: variable oscillator assignment options, two new operating modes (dual oscillator and single oscillator)
- Polyphony: max. 32 voices
- Sounds: 500 sounds and 100 ROM performances
- SRAM (2MB = 4 x 4 Mbit (512K x 8bit) -Battery Backed)
- Sequencer: the sounds can be edited while the sequence is running, LFO can be synchronized with MIDI clock
- Internal memory: A battery-backed RAM disk can be set up (with additional memory modules)
- Sample features: Sample Translator integrated into the operating system, import of samples in many formats, Akai multisamples are read (no sound data)
- MIDI: Compatible with the GM and Roland GS standards
The S Series gained such an incredible foot hold in the music industry, from home players, music studios, bands and the music trade support - Worldwide. When we think back, armed with the knowledge of what can be achieved in todays technology, GEM were definitely at the forefront of musical instrument innovation.
This did leave the original S2 and S3 owners, whom just bought their new keyboard synthesiser a year earlier, in frustrated limbo. So, without further ado, GEM released a retro-fit Turbokit - eveyone was then happy!
GEM continued making a large range of home, semi-professional keyboards, digital pianos and a wealth of pro audio equipment, through the sister LEM branding. Unfortunately, due to various reasons, some of which we will never know, sadly they ceased business in 2009, before becoming bankrupt in 2011.
However, the legacy of these certainly keyboards lives on!
What is the Turbokit?The Turbokit is essentially an add-on daughter board (expansion), that is hard wired into the main motherboard. it contains larger memory EPROMs, SRAM and an array of logic. It also contained a potential SCSI interface, which never saw any light. Personally speaking, the lack of hard drive was a big mistake but hey ho.
In addition some components are changed/removed/added to the main Motherboard such as Floppy controller and PCB mounted battery. This is a time consuming process and needed ideally an electronic engineering minded person to do it.
The kit includes a custom metal bracket to support this board on top of the motherboard. It's positioning is such that it can connect to two existing EPROM DIL sockets, 9 way header and separate wires.
Fitting the Turbokit is not for the faint hearted as PCB tracks on the Motherboard have to be cut and components removed and added.
Turbokit FeaturesThe Turbokit offers the Musician many features and is well worth the trouble and expense. However, I have seen motherboards damaged but repairable, and at least a couple destroyed in the distant past, by people upgrading their S Series.
Historical Issues from both Turbo and non-Turbo Keyboards.
- Leaky Ni–Cd Battery (otherwise known as a NiCad Battery)! This has caused so many issues, and continues to as well! The damage caused by the leaking battery is often found too late as the acid eats into the PCB, tracks and surrounding components - often making the motherboard irreparable.
- DIL IC Sockets on both the Motherboard and TurboKit (Cheap Tin Plated). The low cost DIL IC sockets are often found to tarnish, where the tin contacts fail to make connection to the inserted IC.
- DIL (Dual inline) ICs`Failing. This follows on from the above. Sometimes removing the IC from the socket and replacing again can fix. Other times it is best to remove, clean the contacts with something like a fibreglass contact cleaning pen and reinsert. Heat also makes the IC work loose, sometimes even making the IC fall out!.
- Ribbon Cable with IDC and DIL PCB Header. This header contains two ribbon cables with DIL IDC header plugs. These plug into the original EPROM sockets on the motherboard.
- Metal Bracket (Expensive Fabrication). The complete kit should have one of these brackets included. However, I do recall sometimes it was missed out at the factory leaving the 'installer' with some head scratching of how to mount the Turbokit.
- Physical Layout and Fitting. With space at a premium, coupled with the semi-weighted keyboard hovering over the motherboard, mounting space is tight.
- Reduce weight. The original Tubokit board is quite weighty due to the large 32 pin EPROMs and SRAM, together with interfacing components.
- Move from old PTH DIL to SMT
- Buss IC Connectors. Lots of data and control lines.
- Control Header.
- Battery.
- Main Motherboard updates to both sides.
- Metalwork. As the new deign was to be surface mount (SMT), the board itself was much lighter in weight, so the inclusion of a metal bracket was not ideally a requirement.
- Ribbon Cable. A way to connect the 9 way went through many thoughts and testing.
- The connection to the two old EPROM sockets were the main the data bus interface. This proved very problematic for a variety of reasons.
- Battery. Least said the better!
- Difficult Installation. Some changes was made to the original Turbikit schematic design, in order to remove some of the motherboard modifications
- Keyboard Rear Bar. This presented space issues between the bar and the motherboard component heights. Some time was spent on potential solutions and this lead to the early prototype and final design.
- Old PTH Components. From the start I wanted to remove ALL PTH components on the Turbokit. Some are still required on the motherboard.
- Ribbon Cable with IDC and DIL Header. The concept of the design was to completely remove the need for this ribbon cable. It has always been an issue and an alternative direct PCB header was the preferred route.
- Data Bus Timing and Noise. The motherboard is very busy in data signals and components. This coupled into the Turbokit led to work in ensuring timing signals were not displaced and noise kept to a minimum.
- An element of reverse engineering was made in order maintain data bus integrity and signal quality v unwanted crosswalk of the switching lines.
- Ni-MH Charger. As I wanted to completely get away from the leaking battery situation, it was decided to 'remote' the battery away from the motherboard and Turbokit. As such, a Ni-MH (Nickel Metal Hydride), battery was preferred, with a connector on the Turbokit and a way to charge this battery was then included on the new Turbokit design.
- Programming. None required.
- Code / Binaries. Use existing EPROM code but in smaller OTP packages.
- Mechanical Challenges. Space, interconnections and fixtiures.
- Reliability. A must.
- Battery. Ability to mount this away from the main motherboard - other end of the keyboard enclosure.
To take the proven electrical design of the board and reposition of components to make improvements to data bus routing and board balancing.
- Shape altered to accommodate additional components
- Re-added the SCSI interface connector (future options were being investigated)
- Altered the 9W connector to ease end-user alignment and installation. This proved on the surface a simple item but in practice had many mechanical alignment issues.
- Change some components to ease production of SMT hand placement (as these boards were all hand placed and then put into a SMT oven)
- Added links to facilitate future options
- Added hooks to facilitate future options
- Dedicated S Series motherboard. without investing in a full test fixture, a complete S Series is required to enable the Turbokit to be fully tested.
- Basic Jig. A basic jig was made which brings up the Turbokit but this can only go so far in testability.
- Wear on IC DIL Sockets. When testing the Turbokit on an S Series motherboard means the two DIL header sockets quickly become unusable as the contacts wear. A number of tests were done by replacing the sockets with a Turned pin but this caused issues with the Turbokit not fitting correctly.
- Reliability. With the new Turbokit correctly fitted to a S Motherboard, it proves to work reliably, even in light transit. Without bolting the Turbokit down, gigging for example may lead to other failures.
The finished Turbokit?... Maybe
This project was on the surface reasonably straightforward but so many obstacles and unwanted 'spanners' crept in during the concept and design iterations. Many hundreds of hours was spent on the project, ideas, concepts, designing, reverse-engineering, component selections, component availability, technical interfacing, testing, mechanical issues, electrical, redesigning, charging options, EPROMs, SRAM memory, board layout, Beta PCBs etc. Thinking back, it was certainly a challenge.
The new Turbokit design works well and certainly fills the gap left by the non-availability of the original. However, although a few had been made and sold, the investment required, in time and money, to make a test fixture unfortunately means it is not really cost effective. Maybe it has life but we shall see.... I still have most parts and could / maybe pursued to revisit this project, the board itself is solid - the daunting part is testing.
The Turbokit can be found at DitroniX.net
Special thanks to a lifelong friend, Don Turner, whom put up with my R&D OCD during the design processes and helped with various mechanical ideas on-route!
Comments