This module scans the 61 key keyboard and assigns the generated CV and gate signals to the 4 SEM modules according to the adjustments made on the front panel.

Actually, it should do so, upon arrival, no single gate pulse was seen, nor a correct CV.
A defective 723 caused most of the logic to be without supply. But once energized, a hand full of dead CMOS chips paved the way to a working assigner.

[Photo to follow]

A special module of the four voice is its programmer. It allows to set up various parameters and store them into an internal memory with 16 places, separately for each SEM.
It also features two ADS envelope generators for the SEM’s VCA and for modulating the VCF.

The parameters are:

• frequency setting (offset to other sources) for both VCOs and the VCF
• A-D-S for both envelope generators
• Frequency of an internal LFO, individual to each SEM (called “vibrato”)
• Amount of LFO to OSCs (1+2 simultaneously)
• Amount of 2nd envelope to VCF frequency (called “modulation”)

The voltages are converted by a 6 bit A/D converter and then into a bit stream to be either stored into 2 1024 bit SRAMs or bypassed around the memory in manual setting. Due to this, testing and troubleshooting of most parts of the circuits is possible in real time.

Every of the 2 channel boards of the programmer features 4 proprietary envelope generator ICs which are said to be a predecessor of the famous CEM 3310. It is far away from being compatible though, and the custom Oberheim ICs also have an additional VCA integated to the envelope chip.

This photo shows the programmer stacked together, with the channel 3&4 board on top. Some CMOS ICs as well as several LM324 had to be replaced here. The cermet presets are kept in place, while the rarely used carbon track presets are to be replaced.

One potentiometer on the programmer’s front panel unfortunately had a severe crack in the resistive track, so I had to mix up parts of the original potentiometer with the track of a brand new 24mm Alpha brand pot.

A set of new T1 3/4 light bulbs completed the programmer.

The Four Voice has an active output mixer with individual volume and pan controls. Apart from cleaning the potentiometers, two of the used µA741 op amps were defective, one had zero output, the other a large offset.
I decided to use NE5534 with approproate compensation here, whereas all other op amps remain original for sound reasons.

The hearts – the FVS has four of them – is the famous Oberheim SEM synthesizer module. Once intended for enthusiasts as an add-on to their modulars, it has all relevant connections on Molex connectors on its back side. Several of them are internally routed to other modules, but a modification of this FVS is also in preparation, making it into a modular hybrid.

First of all, the SEMs need to be restored to a working state. Potentiometer cleaning is a bit more complicated because of the special pots used for coarse and fine setting of the VCO frequency.
Here’s how this potentiometer looked when disassembled:

After potentiometer and switch maintenance the SEMs will be “re-capped”, tested and repaired where necessary.

I got one working SEM in a break-out case in addition to verify things which was told to sound a lot better than the SEMs of a Two Voice that I have not looked at so far.
So I compared the SEMs from the FVS with the stand-alone SEM and found that the single SEM was lacking a 150nF capacitor in the gate input circuitry that were present in the FVS SEMs.
I wonder which impact this capacitor might have on sound – it clearly slows down the slope of the gate signal to the trigger circuit and envelope generator.
Will it affect the “snappiness” of a short attack time to an audible extent? Hard to say from the circuit, probably from simulation, but to make things clear I will do some testing once the first SEM is up and running.

Here’s the additional capacitor (no comments on my artwork!)