Regulator (NT6)

Updating my initial notes, the NT6 board takes a supply of around 18V from the transformer (depending how you wire it and what you plug it into) and according to the schematic should output −12V, +12V, and two +5V supplies to the other boards, along with a −20V to ground supply for the power LED, and a 7·15V reference.

The exact position of Si1/2, C1/2, and D1/2/3/4 is indeterminable from the schematic; they are numbered here arbitrarily.

An anomaly is that while in the schematic, the A109D is shown with a capacitor/resistor pair connected across pins 3 and 12, in the order pin 3–resistor–capacitor–pin 12 (3–R–C–12), on the board, the order is reversed to pin 3–capacitor–resistor–pin 12 (3–C–R–12).

(Update 2019-09-10: I’ve corrected the numbering for the Tesla MAA723 on my diagram. The schematic uses the numbering of the board holes (and MAA723CN version of the chip) bracketed within the chip outline, while the exterior numbers are the pin numbers as in the MAA723/MAA723H datasheet. It’s not clear from the schematic which of the exterior numbers are 6&8, but this can be checked on the datasheet. So the diagram shows DIP holes 1, 8, 9, 14 empty, but has pins 10,1–5 in holes 2–7, and 6–9 in holes 10–13. The diagram is also updated, as in the earlier version I missed an almost invisible trace between pins 7&8, which does however (sort of) appear on the schematic.)

VCO

The VCO board handles various tuning and keyboard CV/gate functions, and has two oscillators which each generate a square-wave signal at a pitch determined by key-press (an octave above the highest setting on the ‘octave’ switches), which is fed out to the two octave boards.

The board includes some hand-soldered ‘floating’ components (four for each VCO); these are not directly indicated on the A109 schematics, but they are on the B81 schematic, though it’s not apparent whether they were part of the B81 PCB. I’ve listed these as ‘FT’ for transistors and ‘FR’ for resistors, and used the numbers from the B81 schematic.

This VCO board has two anomalies which may be unique:

• A trace has been cut with a soldering iron between R6 and R14/TL5; this seems deliberate, and is presumably a later mod (it leaves R6 as a dead end; which could have been achieved by leaving R6 out during manufacture — or desoldering one end instead of cutting a trace, for that matter . . .  ); the effect is to prevent direct contact between the tuning circuit and the impedance transformer circuit. Why, or what that implies for function, I couldn’t say.

  

• A solder connection in the VCO2 circuit between TL21 and C26, which also connects to IS6 pin 10; this seems accidental and probably a manufacturing flaw resulting from the spindly and overlong tails of the polystyrene caps. ^[5] This connection is not made in VCO1.

  

Once everything else is sorted I’ll try reversing these and see what happens. So these are marked on this version of the VCO diagram, but there should be an amended version in future.

One further anomaly, which I assume is normal for these, is that the pin 3–12 capacitor/resistor pair order for IS1 is inversed from the schematic here too (3–R–C–12, not 3–C–R–12).

VCA

The VCA board provides the main output and headphones amplifier, and also contains the LFO, the Vibrato circuit, and the Noise Source. (I think the actual noise source is T6; which is only connected on two legs, intentionally.)

There are also some mismatches between the VCA schematic and the actual VCA board:

• In the schematic, IS1 pin 4 connects to R8 and R10, while pin 5 connects to R9 and R11. On the board, pin 4 connects to R9 and R10, while pin 5 connects to R8 and R11. Or put simply, R8 and R9 are swapped at one end; and which end you consider as swapped determines the number you assign to the resistors. (I’m assuming they’re swapped at the IS1 ends, so the numbering here matches the schematic at the T2/T3 side.) In effect this inverts the input signals to the op amp compared to the schematic.
• The pin 3/12 capacitor/resistor pair order on the A109Ds is also anomalous here: In the schematic, IS3 and IS6 are shown with 3–C–R–12; but on the board, the order on all A109Ds which use pairs (rather than a single capacitor) is 3–R–C–12.
• In the schematic, R40 is shown as connected between TL22 and the junction of R39, C22 and R34; on the board it is connected between TL22 and the junction of R39, IS3 pin 10 and C19 — i.e. at the other end of R39.
• There is no −12V decoupling capacitor shown on the schematic, but there is one on the board, between C30 and IS6. I’ve called this CX1.

On this board (but perhaps not all):

• R4 is 2·5K on the schema but 2·2K on the board.
• R21 is 250K on the schema, but its case marking is just ‘220’. It measures (in-circuit) 4·36KΩ, so I expect it’s a 220K.
• C28 is a 100µF 10V cap on the board. It’s a Frolyt of a similar age to the others, so is probably original, though below spec.

Octave (FW1, FW2)

The Octave boards divide-down the signal from the VCO board, allowing transposition over five octaves, and use the divisions for waveshaping, producing a sawtooth waveform and a 25% pulse-width. Although the boards are shown and numbered separately on the schematic as FW1 & FW2, they are identical circuits, so this diagram serves for both. n.b. the D100D chips on these boards are Quad NAND gates, and are represented as logic gates on the schematic.

The missing stripe should include R57; its position is as in the B81 version.

EG

This board receives a gate signal and outputs a CV envelope signal in response. It has five presets on switches 1–5, and a manual setting on switch 6, controlled by external pots.

Some issues are:

• There are two R1s on the schematic, first and second in line from the signal input at TL1; I’ve made the second R2.
• There are also two pairs of components, D1 and R7, and C2 and R10, which appear in reversed order on the schematic compared to the board.
• The switch-pin numbering on the schematic does not match a simple ordering of the actual pins on the switches, and the group of contacts for sustain level (lower right on the schematic) are not numbered — they’re actually the 5a/6a group of pin-pairs. So using the schematic numbering, the on-board pin order for each switch ends up as:

  

  I assume the schematic predates the attempted implementation.
• The missing stripe should include R17; its position is also as in the B81 schematic.

VCF

This board contains a ladder filter, with presets/manual/envelope/LFO cutoff control, and presets/manual resonance control.

• The order of C9 and R36 is reversed on the board compared to the schematic.
• There are two R27s on the schematic, at either end of the filter ladder; I think the upper one, connecting between R22/T7/T12 and the +12V rail, should be R21.
• One of the polystyrene caps in the filter ladder has no number in the schematic; I make it C2.
• In the schematic, the connection through the ladder from the audio signal input runs up the left side (T3–T6), and connects to the non-inverting input (pin 3) of op amp IS1 through C8 and R32; while the feedback input connects to the right side (T8–T11) and ultimately to the inverting input (pin 2) through C7 and R31. (Obviously this is a simplification; the sides connect repeatedly to each other.) However, on the board, the audio-in chain connects to the IS1 inverting input, and the feedback chain connects to the non-inverting input. As the two chains are identical between the filter inputs and the final resistors before the op amp, these component numbers are indeterminable from the schematics; the swapover could occur at any of several points in the signal path. I’ve gone with the last possible — in the drawing, the connections between the final resistors and the op amp inputs are R31→pin 3 and R32→pin 2. This is also about the simplest way of drawing a correction on the schematic.

Full Set

All diagrams as multilayer PDF here. (483KB)

Other Components

There are a few components in the synth which are not on boards — keyboard, control pots, the actual waveform and noise selection switches, modwheels, and some power supply components. I’ll go into these in more detail in the next installation, which will also go over the rest of the refurbishment. This had to be left half-done a few months ago pending accommodation/workspace availability, but should be possible to complete by Summer.