[sdiy] Moogs last Poly SL8, machine internals.

karl dalen dalenkarl at yahoo.se
Wed Jun 30 19:26:22 CEST 2010

I wonder what the VHM and quasi second osc really was?

There has been scant information about the Moog SL-8 so I am going 
to take this opportunity to tell you what I know about it.

I started with Moog as a Sales Rep in March of '83 taking Bob Wehrmann's place who went to Sequential Circuits. I was probably Moog's worse factory Sales Rep (Central Division) although I had sold a many $100K's of product in Dallas. I was brought to the factory in late May before the Summer NAMM show at McCormick place and was told that I would be performing the "demos" of a new instrument in the hotel room.

Rich Walborn came and took me back to the back development area and I met a long haired gentlemen I believe was Ray Dennison who was the software engineer of the SL-8. The name SL-8 means the SPLIT/LAYER eight voice and the idea behind it was to create a low cost variant of the MemoryMoog that would be more reliable and penetrate the market segmentation that the Prophet 600 and Juno 60 was currently reaping.

The SL-8 was based upon the TI 99/4 16 bit microcomputer and the SW development was done on the TMS 9900 development system. A single precision relaxation oscillator circuit was reset by pulses that were generated by "lookups" within the microcomputer memory and the resultant "DCO" based design was very stable. Standard waveshaping components were derived from the basic sawtooth waveshape and enabled/disabled through CMOS switching. A master output VCA of the oscillator allowed the discrete transistor ladder VCF (also DCO) to be overdriven and be "phattened" in the Moog tradition.

A "Variable Harmonic Multiplier" quasi 2nd oscillator section could be "soft synced" to the main DCO and the VHM reset pulses were generated by the 99/4 controller and this overcame some of the single oscillator stigma since the waveshape of the VHM could be modulated as well by an EG or LFO.

The VCA path was CEM and a stereo chorus programmable per patch with variable spped and depth parameters really fattened up the sound.

When I first saw and heard the instrumant in early June it was barely working. Rich and I worked on HW issues and Ray toiled at writing the code that would make the SL-8 usuable as a hotel room "BUZZ DEMO" synthesizer. After many hard hours of testing and SW debugging, the SL-8 finally started becoming a decent sounding Moog that would split, layer, and edit from the panel scan and store about ten patches. It ran on the development system until two days before NAMM when I suggested that we try running it on a 99/4 in the socket instead of the emulation environment. There were numerous issues, but it actually ran on the IC computer instead of the emulation system before we left for NAMM.

At the hotel suite across the suite from McCormick Center, we had a sales meeting before NAMM and I had been pledged to secrecy not to "spill the beans". Nancy Kewin, Lee Hargrove, and other Sales Reps were booted from the room and Ray, Rich, and yours truly set up a SL-8 with the hidden TMS9900 development environment behind a backdrop. The system was booted and with a pair of SynAmps (for stereo) and started programming the ten demo sounds into the SL-8. I remember hearing Rich say "He really knows how to program this instrument!" as I continued. Nancy, Lee and all the rest of the other Sales Reps had their ear to the door outside until I said, "OK....I'm Ready!!???!"

I performed about a 15 minute demo with the ten sounds and then watched the faces with big smiles thinking ($$$$$$$) commissions.

After a few drinks and everyone discussiong how to market the instrument to the "big dogs" that would be brought over to the hotel from the NAMM floor for a demo.

Next day the major synth sellers in the US were given private demos and the vaporware SL-8 was given thousands of units of token orders.

On the last day of the show I finally got some time to go over to floor and at the Moog booth and I saw an ashen faced Herb Deutsch say to me "Craig, go over to the Yamaha booth". I went over and saw a sleek keyboard with velocity/pressure sensing and put on the headphones.....to my great surprise the six operator DX-7 synthesizer with the new algorithms and feedback added to the older GS-1/TRX FM was astounding. I asked how much...and the $1995 answer explained Herb's crestfallen demeanor. How would Moog compete with this?

Tom Rhea was brought in, there was downsizing, and the rest is known.

The SL-8 blew away the Phophet 600, Juno 60, Poly Six, and the other subtractive synths, but the DX-7 was a megaton range instrument in the sea of synths and went on to decisively command the market.

The DCO and assignment algorithm was capable of polyphonic portamento unlike the other DCO instruments of the time. We discussed emulating the four assignment algorithms in the MemoryMoog although only the NEXT mode was operational for the demos. The integration time of the poly portamento was a linear response curve. This was because that was easiest to do quickly with the time we had. An exponential portamento response curve would have only taken a few more days.

The EG resolution was very high with 7 bit accuracy for all of the envelope values. The Prophet 600 was 4 bit, the Roland 5 bit so the sound design was much easier and predictable. The VCA and VCF would "snap" if you set a 50% to 60% sustain level with no attack, delay, or release just like the Mini or MemoryMoog.

The oscillator had a Schmitt triggered suboctave square wave available that could be switched in at equal output. PWM (ok rectangle wave mod) could be manually set, EG mod with atten, or LFOed with the dual slope integrator LFO triangle/square wave.

The Variable Harmonic Multiplier also allowed some wave shape modulation from the EG or LFO with its own attenuator.

If you carefully set up the DCO with some smooth PWM, and modulated the VHM with some tweaking, you got a pretty good dual oscillator unison sound. The amplitude nulls between beating oscillators on a true 2 or more oscillator synth were not audible, but the waveform animation was phat and very distinctive. With a little syrup on the VCA EG and the internal stereo chorus tweaked it did a great string chorus sound that was very versatile.

I realized great brass, organ, monster unison dual layer basses, great tearing sync, and many other highly usuable patches with relative ease because of the controls' resolution, Moog ladder LPF, and distinctive wave shaping circuits taken from legacy moog designs. The inaccuracy of pitch control through exponential current converters was gone and just enough slop was in the oscillator integration circuits to keep it from being totally sterile with "perfect" interger derived tuning (which is what you really want).

One of the really cool attributes of the TMS 9900 MCU was it had a considerably faster 3.3 MHz clock than the 8051/8031 MCU used in the Roland and the 8085 in the Korg units at the time. It ran rings around a Z80 because of the 16 bit data path and the unique ability of the 9900 to store all user registers in RAM and update the "RAM Registers" in a single instruction cycle. If you called a subroutine or a NMI call was true, you could update the user register directly. This was how all of the parameters were stored/updated once a program or configuration change was made. It also made complex multi tasked threading for the housekeeping, DCO timing, triggering, MIDI, and controller interaction much simpler to code.

Which brings up another point......this baby would have been reliable if built. Almost all of the issues associated with the needed tuning accuracy that the MemoryMoog did not have were mitigated. The MemoryMoogs 12bit DAC only allowed 4096 tuning units. With all of the summed trims there are in the MM, it's a wonder that they even tune at all. You have to get everything right and then it's an evanescent experience. If you are really competent and make it a labor of love, then a MemoryMoog tune may last a few months, but its the Fluke 87 III, mirror, and a crooked neck after that time span.

THE SL-8 would have had none of those issues, but it was still a Quiotic Quest for Moog Music. The Japanese Stentorian Monoliths of Yamaha, Roland, and Korg had the research and investments. They were in it for the long haul as are most Asian companies, and they gathered Traction in the 80's with only Sequential, Emu, and a few others able to eke out some profit.

BTW Roger.....who got the cool Hockey Game in the cafeteria with the great sound effects? Very Happy

More information about the Synth-diy mailing list