oberheim     matrix-1000

 

general info

midi spec

firmware notes

hardware notes

 

 

 

 

 

 

 

 

 

.............

There's a great debate going on between owners of Matrix-1000's and Matrix-6's about whether there is a recognizable difference in sound between those two boards. The difference (if it exists) can be caused either by different manufacturing process of CEM3396 voice chips ('wide-body' version in M-6 vs. 'narrow-body' version in M-1000) or different circuits generating DCO frequencies. In case of M-1000 all 12 frequences are derived from the 6809 CPU E signal (2 MHz, derived from an 8MHz crystal) by dividing it down with 82C54 timers, resulting in very stable and consistent pitch. M-6 has three free running High Frequency Oscillators, one for all six voices of DCO1, the rest for the voices 1-2 and 3-6 of DCO2, respectively. Besides, they are not quartz based but LC oscillators naturally resulting in some drift between their frequencies (there's an HFO calibration procedure in M-6 service manual to take care of this).

.............

Matrix-1000 uses much more physical memory than the 64K memory space of 6809 would permit. It's got 32K of ROM for firmware, 64K of ROM for 800 factory patches and 32K of RAM for User Patches, Global Parameters and voice data. 8K is reserved for various IO addresses. To comply with this M-1000 employs memory paging, fitting 64K of factory patches into 8K memory space by splitting it into 8 8K pages. Curiously, there's an empty 8K slot in the memory map which on the board is wired to an empty space for a 28-pin socket. While the schematics read '27512, FOR FUTURE EXPANSION' the wiring is actually for a 27256 (32K ROM). Now, 32K ROM would mean 400 more factory patches... there were rumours about Matrix-1400 (with 400 extra patches naturally) that never saw the light of day... could this be it?

.............

One of the few things done in hardware is Oscillator Sync. There is no information whatsoever in M-1000 manual about DCO Sync modes. Matrix-6 manual lists three different DCO Sync modes but no explanation what they actually mean. Two sync signals, SYNC1 and SYNC2 are present in M-1000 (and M-6) hardware. When SYNC1 line is active the clock signal driving DCO1 is reset each time when DCO2 reaches a whole phase, a feature traditionally known as 'hard sync'. When SYNC2 is active, the clock signal for DCO1 is the result of OR'ing DCO1 and DCO2  clock signals (sometimes called 'soft sync'). Relationship between Sync signals and value of Patch Parameter 2 (DCO Sync Mode) is straightforward: bit 1 becomes SYNC1, bit 0 becomes SYNC2. Therefore DCO Sync Mode = 1 (SOFT Sync) activates SYNC2, DCO Sync Mode = 2 (MEDIUM Sync) activates SYNC1. DCO Sync Mode = 3 (HARD sync in Matrix slang) means both SYNC lines are active. Finally, DCO Sync Mode 0 means no hardware sync. There is an interesting thread explaining this in detail in [music-dsp] forum.   

.............

CEM3396 is an all-analog chip. The only 'digital' signal it ever sees is the clock pulse that determines frequency of the generated sound. The DAC chip (AM6012) on the board has no 'direct' contact with the sound. The role it has is providing all the analog control voltages for all six of the CEM3396 IC's (one per voice). Each chip has 10 CV inputs, adding up 60 control voltages altogether, which requires a lot of multiplexing.    

.............

Matrix-1000 runs internally on a 50Hz clock. This means that once every 20 milliseconds one of the 82C54 Timers kicks M-1000 out of the idle loop to check whether something has been transmitted over MIDI, whether a voice is gated or whether a panel button has been touched. Hopefully, nobody has sent anything, pressed anything or gated anything so it can go back to idleness, otherwise it has to do a lot of processing. According to today's standards 20 ms is a looong wait, so considering all this talk about the latency of DAW's and whether MIDI lags...  

.............

Data sheets for the most interesting parts inside Matrix-1000 can be downloaded here:

CEM3396 Microprocessor Controlled Synthesizer-On-A-Chip
MC68B09 8-bit Microprocessor
MC68B50 Asynchronous Communications Interface Adapter (ACIA)
82C54 Programmable Interval Timer
DAC312 12-Bit High-Speed D/A Converter
TC4051/4053 CMOS Multiplexer/Demultiplexer

Any comments/suggestions/complaints/corrections? Pls forward them to me.