Let’s Get Real
With the drawing updated and complete, it was time to start making the drawing a reality. I set about working on the mechanical design. One of the first things that needed to be done once we had a basic mechanical design was to translate the panel layout into a dimensioned drawing so that hardware engineer Tony Karavidas could lay out the panel PCBs. Tony would also lay out the main board and then I would get a dimensioned drawing from him so that I could add the PCB mounting hardware to the bottom panel metal and add back panel cutouts for things like phone jacks, MIDI jacks, and the USB port.
One major difference between the Prophet-6 and every Sequential Circuits and DSI instrument that had come before it, is that it has discrete VCOs and VCFs. (Some would quibble over whether they are truly discrete as there are some silicon packages involved, but I’m going with discrete. No sane person likes to match transistors.) Historically, every Sequential Circuits instrument used either Solid State Music or Curtis Electro-Music VCO and/or VCF chips, not to mention envelope and VCA chips. When Dave started DSI in 2001, the only analog synthesizer chip still in (limited) production was Doug Curtis’s CEM 3397 and it was used in every instrument leading up to the Prophet-6, with the exception of the Pro 2, which had DSP oscillators and discrete (there I go again) SSM 2040-inspired and SEM filters. As I write this in 2025, there are multiple VCO-on-a-chip and VCF-on-a-chip options. In 2014, there was one (technically DCO, not VCO) and it wasn’t suited to the instrument we had in mind.
Tony was responsible for the design of the individual voice cards, each sporting two VCOs, the triangle suboscillator, a SSM 2040-inspired, 4-pole, discrete OTA, low-pass filter, and a 2-pole, resonant, high-pass filter.
A note about the filter. There were a couple of 2040 filter clone schematics floating around online at the time, the most well-known being the late, great Jurgen Haible’s and a variation produced by Rene Schmidt. However, neither of those designs were used. We did what Jurgen likely did when designing his. We went back to the original 2040 design documents and the data sheet. I also hesitate to call the Prophet-6 filter a SSM 2040 clone, because we never actually made direct comparisons with a genuine 2040. We just knew it sounded good.
All of this happened relatively quickly, as software design couldn’t really get going until there was hardware to work with. Software engineer Chris Hector was responsible for much of the operating software of the synth, with Dave focusing on the low-level programming, like CV-related tasks and other things related to making the voice work. That was what he liked to do. He had employees to do the things he didn’t want to do. Dave loved “bringing up” a synth. That is, getting it to make noise—any noise–for the first time.
One of the great things about the Dave Smith Instruments development team was that everyone knew what they needed to do and how to do it and, given the distributed origins of the company (that is, everyone working on their own at home), everyone was capable of working independently. That said, there still needed to be someone steering development and making sure we all knew what everyone else was doing, keeping us focused on the job at hand and stepping in as an editor or referee as needed. Dave was a great project manager. If that sounds like faint praise, look at it this way: Not one of the instruments that came out of Sequential Circuits or DSI would have happened without Dave. He was the key to everything.
The one aspect of the Prophet-6 that got Dave excited was the prospect of building an analog polysynth without the limitations imposed by the microprocessors of the ’70s and ’80s. Dave had a knack for problem solving. One could argue that the Prophet-5 was inspired by his desire to solve problems. The same could be said for MIDI. With the Prophet-6, he wanted to eliminate the Tune button.
The tuning stability of voltage-controlled oscillators (VCOs) and filters (VCFs) is affected by temperature. There are various things that can be done in hardware to improve the stability, but nothing completely eliminates the issue. If it’s a monophonic analog synthesizer, you can turn it on well in advance of when you need to play it, and hope that its internal temperature and the ambient temperature reach a point where they’re stable enough that you can tune the synth and it will stay in tune while you perform. However, if it’s a polyphonic analog synthesizer, even small differences in oscillator and filter tuning between voices become apparent. Air conditioning ducts and direct, unfiltered sunlight are not your friends.
Enter the Tune button. The Prophet-5—and most of the programmable analog polysynths that followed it—had an autocalibration routine that could be run to get the oscillators and filters back in line. Hit the Tune button, hope the synth reawakens before the downbeat, and—voilà!—back in tune.
Dave’s solution for the Prophet-6 was to use the autocalibrate routine to build a table of oscillator and filter tuning offsets tied to the internal temperature of the synth at the time of calibration and then interpolate between them for a range of operating temperatures. The calibration still must be run manually, but once the tuning table is populated, calibration should only need to be run when the operating temperature is beyond the extremes of what’s in the table, eliminating the need for a dedicated Tune button. Brilliant!
(People sometimes ask me if Dave had any patents. I don’t think he really believed in patents. He definitely didn’t want to have to deal with [or pay] attorneys, unless it was absolutely necessary. So, to recap. Marketing.Patents. Lawyers. Furthermore, when it came to synth design, he didn’t like slide potentiometers, white synthesizers, synthesizer models designated with combinations of letters and numbers like “XL-5,” or light-hued wood like maple or ash.)
The Effects
Bob Coover’s DSP contributions first appeared in the Prophet 12 in 2013, but the Prophet-6 was the first DSI instrument to have a dedicated multi-effects processor. We made a list of what we thought were “must have” effects. I no longer have that list, so I’m relying on my memory here, which may be faulty. I believe most of the effects on the list appeared in the initial offering: BBD delay emulation (which had first appeared in the Pro 2), “clean” digital delay, chorus, phase shifter, and hall, room, plate, and spring reverb emulations. Flanger and tape delay were also on the list. I don’t recall when the flanger made it in. Maybe when the effects were updated for the OB-6? (The OB-6’s ring modulator, a tribute to Tom Oberheim’s vintage effect, was also rolled back into the Prophet-6.) The tape delay never did make it in and didn’t appear in an instrument until Take 5. Why did it take so long? No idea. Lack of DSP horsepower, maybe? I was pushing for it. The model already existed. In fact, “Drive” in the Character section of the Prophet 12 is based on the Echoplex model’s tape saturation.
The “Easter Egg”
If you’ve ever had a guitar amp or an old mixer with a spring reverb tank, you probably know that abrupt physical shocks can cause the reverb springs to make a thunderous, clattering sound. During one of the weekly meetings, someone suggested as a joke that it would be funny if the digital spring reverb emulation offered the ability to make that happen. I don’t remember who made the joke. It could even have been me. Then someone—Dave, I think—suggested adding an accelerometer and some samples of a reverb tank being bashed for the sole purpose of making the joke a reality.
To hear it, the spring reverb effect must be active and the mix level turned up. Then just slap one of the wood end panels hard with your open hand.
As far as I am aware, the effect was rarely triggered unknowingly. I think Technical Support did receive a report from a customer who encountered it while playing in a very loud, bass-heavy environment.
Note: The accelerometer used in the Prophet-6 is now obsolete. If stock of the part becomes depleted before production ceases, it will no longer be present on the PC board. If you’re whacking your synth and nothing is happening, that might be why.