Some notes on the operation of the OB-1 programmer circuit. R Grieb 2/25/2016 (Based on unit with Serial #201, which did not have the CEM3310 envelope generators) This is based on my understanding of the circuit, based on the schematic and some measurements. It may contain errors. The pots and switches are always digitized, converted into a serial stream, and then applied as controls to the synthesizer circuits, even in manual mode. U1 osc freq in my unit was 505 KHz. C4 was at 126.66 KHz, or Osc/4. This is fed to counter U21. The S7 output of U21 has a period of approx 89 C4 clock cycles, or 707.6 uSec The T1 signal that resets U21 is also used to clock U17, so U17 increments once for every full cycle of U21. Each U17 value selects a particular pot to convert, or a group of six switches. U17 counts from 0 to 21, and resets to 0 on 22. So its period is 22 * 707.6 uSec, or 15.57 mSec. That gives a rate of 64.68 Hz for scanning all of the pots and the switches. During each count of U17, U21 counts from 0 to about 89. The S1-6 count value is converted to an analog voltage by U16 and its resistor ladder. This voltage is fed to one input of an LM301, the other input of which is the output of a multiplexer circuit composed of three CD4051 muxes. These muxes select one of 18 potentiometer to be converted to a 6-bit digital value. When the S1-6 DAC value matches the voltage of the selected pot, the LM301 changes state. This feeds into the clock input of U4, to latch the current value of S1-S6, which is the pot digital value. At this time, U11 and U12 are driving S1-S6 onto the B1-6 control values bus. U4 is a shift register, so after the pot value is loaded, it can be shifted out, either into U22 SRAM or fed into shift register U45, where it is converted back to parallel form, fed to a DAC, and then stored on one of the S&H CV capacitors. U25 is wired as a 2:1 mux to select either the serial output of U4 or the 1-bit serial output of the 1Kx1 SRAM chip U22 to feed to the synth circuits. This mux is controlled by the MAN and MAN* complementary mode signals. U22 stores the 8 programs. Since there are 1024 bits available, each program can use up to 128 bits. These bits are written to the SRAM and read out serially, one at a time. There are 18 scanned potentiometers, which are converted to 6-bit values. That makes 108 bits for the pots, which leaves 20 bits for the 20 switches that are scanned. U17 counts 0-21. During counts 0-7, P1* is low, which enables mux A1 on the pot board. This mux is used for the eight envelope control pots. During counts 8-15, P2* is low, which enables Mux A3. This mux is fed by the VCO and all but one of the VCF control pots. During counts 16-21, P3* is low, which enables mux A2. This mux is fed by the two remaining scanned potentiometers. There are four other potentiometers that are not scanned, and are not saved in the program. Since the pots are converted to 6-bit values, it's convenient to read the 20 scanned switches in groups of 6 as well. P3* is low for six counts of U17. Only two counts are needed to convert the two pots connected to the A2 mux. The select inputs for A2 come from U17 low-order bits. Inputs 0-3 are tied to gnd, and the two pots connect to inputs 4 and 5. So probably the four rows of switches are scanned during the first 4 counts of P3* low time, and then the two remaining pots are converted during the last two counts. Chip A5 on the pot board drives one end of the scanned switches low through isolating diodes. The other ends of the switches connect to the B1-B6 control values data bus. Only one output of A5 is low at a time, and at that time, its switches will pull B1-6 low if they are closed. This bus is pulled high with resistors, so any bit that is not pulled low by a closed switch will read as a "1". So the B1-6 control values bus has the switch row values at some times, and the pot digital values at other times. All of these values are latched into U4, and shifted out serially. While contact is made with one of the eight touch pads, a signal is coupled into the clock and data inputs of the two 74C174 chips on the touch board. This will cause touched pad's output to go high at the output of the 74C174. The eight outputs are encoded into three binary outputs plus one output for touch pad #1. These outputs are valid while the pad is being touched. All four of the output signals are OR'd using diodes, to create the signal at U31-13. The output of this gate creates a pulse to set flip flop U32, which turns off manual mode. Pressing the "manual" touch pad creates a pulse which resets U32, which enables manual mode. The binary-encoded version of which touch pad as touched is latched into U14 and then used as the upper three address bits of U22. So these bits select one of the eight programs, either for writing or for reading. Touching the WRITE touch pad causes D10 to go low. This enables U30 pin 13 to charge to 10V. On my unit it takes about four seconds for the capacitor to charge to 10V. Once U30 pin 13 reaches a logic 1 state, U30-12 will go low. This enables U18 13,12,11,10 to oscillate with U15, at a fairly low rate, which causes the MANUAL LED to flash. The MAN signal is still high at this time, and MAN* is low. Assuming the WRITE pad is still touched, if we touch one of the program pads, flip flop U32-1 will be set, driving MAN* high. This feeds into U31-9 which drives U15-13 low. This stops the LED flashing. If the write pad is touched, and we are in manual mode, MAN* at U31-9 will be low, and U31-8 will go low from write being touched. So TSW will be high and TSW* low. TSS* will be low as well. So TWAS will be high at U32-9. This will enable the WR signal on the next T1 clock pulse, causing data to be written into SRAM U22. It seems that when the RC pulse at U30-6 times out, writing will stop. We are only writing to the selected program area , so if we write the data more than once it shouldn't be a problem. (I may be wrong about what stops the writing) Chips U23 and U22 operate at 5V, unlike the rest of the programmer, so 4050 level shifters must be used to interface to them. Since the SRAM is only 1 bit wide, six addresses must be written to or read for every pot or switch row. So address counter U23 needs to count through six addresses each time U17 increments by one. U33 provides the same DAC reference to both the DAC used to convert the pots, and the DAC which is used to drive the synth CV's. The latched M7-9 program select bits are decoded and used to drive one of the front panel program LEDs. The LEDs are disabled in manual mode, except during writing. The F and E program board connectors feed to connectors on the rear panel. E is for a cassette interface, and F is for an external device which could select a particular program. I don't know if either of these interfaces was every used by Oberheim.