Is It Still Too Soon After Bob Moog's Death (2005!) To Openly Discuss EtherWave Pro Design Details?

Is It Still Too Soon After Bob Moog's Death (2005!) To Openly Discuss EtherWave Pro Design Details?

Asking for a friend. ;-)

I believe I've got pictures of the pitch and volume boards (link).  Can't remember where I got them, maybe from Element14 a while back? Photo info says they were taken with an Olympus C5050Z, and they are from March 2012 or earlier (when I downloaded them?).  They look like the boards in a picture taken by Thierry of a disassembled EWPro (from Pintrest) and from this review: http://www.1377731.com/etherpro/.

Given enough time, some enterprising soul could probably reverse engineer them and have tales to tell.  But are we "allowed" to at this point or not?  And who decides?

For the controls, "Timbre" seems to select one of 6 preset tones, with the 6th being variable via the "Wave, Brightness, Filter" knobs, with "Volume Response" and "Pitch Response" as front panel trimmers.  Could someone clearly explain exactly what all of these do (in electronic music or electrical engineering) terms?  What do the trimmers trim, sensitivity (hand gesture octave width)?  If so, what is the range of the adjustment?  If not, is it more of a linearity thing, or what?

Also, are there any manuals out there?  Moog Inc. doesn't seem to have any on their site, which is kind of strange IMO.

My interest in this is actually rather limited.  I'd be satisfied with a peek at a schematic, or a general description of how the timbres are generated.  And I hate secrets, particularly in the Theremin field, as they almost always work for the man and against the user. People should really know what this highly-touted instrument is before they go plunking down serious cash on the used market. FredM took what he knew of it to the grave, but I'm sure Thierry, and perhaps Peter Pringle, could spill some beans if they desired to do so.

Bob Moog did never publish any circuit details, thus it has to be considered as protected content until 70 years after his death (similar to the Ondes Martenot, where the family persecutes publishing or using unauthorized details). That’s why I don’t tend towards sharing what I did in reverse engineering for servicing and repair purposes, at least for the moment. 

The day will for sure come when I’ll change my mind. I’ll not live forever, thus this knowledge has to be transmitted. But at least up to now, I didn’t find an appropriate candidate who was a skilled engineer and a gifted musician at the same time, so that he/she could really understand how things work, and why.

I’ve moved away from the idea that it is sufficient to be a brilliant engineer to understand the full musical interaction between a thereminist and their instrument. Many EEs have blind spots on the musical side and judge many things under completely different criteria than a musician. Reading many of the technical discussions here in these forums (among others) seem, at least in my eyes, to confirm this point of view - which made that over the years, I often lost the motivation to partake in such conversations.

"Bob Moog did never publish any circuit details, thus it has to be considered as protected content until 70 years after his death."  - Thierry

Horse shit.  If it's patented the patents only last for so long and they're published for god's sake.  If it's not patented it's at best a trade secret and fair game.  Either way it's entirely legal to reverse engineer things and talk about it openly.

This is an instrument they only made a couple hundred of, and practically gave most of those away.  Bob's dead and Moog Inc. by all appearances has had no interest in it whatsoever since then.  There are no pro-quality instruments to buy new, and many of those who might design and build them won't let the nuclear secrets out of their cold dead hands long enough to actually move the field forward.  The real hurdle here is manufacturing, not intellectual property, so why all the tight lips?  It's extremely counter-productive and against the spirit of how Bob published schematics and how-to articles in the past.

And people wonder why the Theremin seems back on its way out.

For example, if you have a schematic of an electronic device, the owner has a copyright on the drawing. You are not allowed to make a photography and publish it. But It's allowed to redraw it and then sharing the new schematic.

 Horse shit.  If it's patented the patents only last for so long and they're published for god's sake.  If it's not patented it's at best a trade secret and fair game.  Either way it's entirely legal to reverse engineer things and talk about it openly.

Yes, it would seem obvious that in order to patent something, some sort of novel innovation would have to be disclosed in the patent application (ie circuit details). This is just basic patent logic, children understand this.

Pitch Side Pondering

I scoured the web and found a few more pix.  I zipped up all my pix with datasheets for the ICs I could identify on the pitch and volume boards (link).

In it we have views of the front and back of both the volume and pitch PWBs.  The volume board has many traces on the top and bottom, it might be only a double layer and therefore traceable given a certain amount of effort.  The pitch board unfortunately looks to have more layers inside, with virtually no traces visible on the outer layers.  Ah well, we can still wildly guess what might be going on.  Here is a quickie annotated view of the component side of the pitch board:

Oscillators: Instead of the usual IF variable inductor can, Bob used a fixed RF choke and a fixed cap + variable cap.  The TO3 transistors look quite a bit beefier than the small signal ones used in the EtherWave Standard and Plus.  Interesting the way the ground plane under the chokes is "combed" on both sides, presumably so they don't look like shorted windings for any leakage fields (which would hurt Q).  I believe Bob stuck something like his standard voltage to capacitance pitch null tuning circuit on the fixed oscillator side, and I assume the oscillator topology is his standard differential pair.

Digital: The front panel octave switch is a big hint that there is almost certainly something digital (gasps, grasps pearls, faints) going on. One way to lower the pitch an octave or two is to divide the heterodyned and filtered wave by 2 or 4, but this gives you a square wave.  A better way is to divide both the fixed and variable oscillators by 2 or 4 before heterodyning, and I think that's probably what's going on here.  There is a dual 4-bit binary counter super close to the quad XOR gate, and they are located smack dab in the middle of the board, so I imagine the fixed oscillator is fed to one counter, and the variable oscillator to the other.  Then counter tap 2 or 4 is picked off both counters and mixed (heterodyned) with one of the XOR gates.  If I were doing this I might use three of the XOR gates to do three heterodynes at once (one for each octave) and pick the one I want, and that way the octave switching is simpler.  And maybe you could mix in the higher ones to get different waveforms or something.

Analog: It doesn't surprise me to see three dual op-amps (TL072) on the board.  And the 1-of-8 analog switches are probably there for voice timbre selection, but what's up with the three dual transconductance amplifiers?  They can be used to make VCAs, VCFs, VCOs, and general voltage control.  So my guess is voltage controlled mixing and filtering, but could there be something more nefarious (VCO) going on here?  Youth wants to know...

CV: The pitch and volume control voltage outputs have to be generated somewhere.  If there is true heterodyning going on with the pitch generation (I assume so due to the presence of the XORs, but maybe not), then any voltage derived from that will almost certainly be exponential.  Synth modules want a linear response here (1V/octave) so there should be a logarithmic converter somewhere.  Over on the volume board (connected via ribbon cable to the pitch board) there is a quad matched transistor array, and I assume that is being used as the heart of a dual CV log converter.

The Crux: If someone who knows would simply tell me whether the EWPro waveforms we hear are either:

1) the direct result of heterodyning

  - OR -

2) generated second-hand by a VCO

I'd drop this investigation right now because that's all that really interests me.  Keep all other of your other secrets until someone more deserving comes along, or take them to your grave for all I care.  But please, spare me the lectures on how it's impossible / wrong for musical engineers (like Bob M.) to engineer musical things (like the EWPro).  I get poked in the eye around here enough as it is.

No VCO. 

Oscillator signals go to LM 393 comparators to be transformed into square waves. Then directly mixed in XOR for highest register. At the same time divided by 2 and by 4 in a quad D-Flipflop. The /2 and /4 signals mixed separately in different XORs. The /1 mixer output signal is then used for the tuner output and the pitch CV generation. But all three pitches go to the register switch which will decide which of the three goes to the wave shaping.

BTW: this had already been written in these forums...

Reasons

FredM was a great analog engineer and a great person, and he was deeply shocked and offended by something going on with the EWPro pitch side voice generation.  I should have asked him what that was and now can't (RIP) but to hear him talk around it, it was something antithetical to analog.  

The main reason I care about all this is: IF the Theremin voice we hear emanating from the EWPro is VCO generated, then (just as with the Theremini) there could very easily be latency issues associated with pitch to voltage conversion, as well the filtering of that voltage (FredM and I both fought with this in our designs).  Latency issues can trivially arise in digital or indirect analog approaches, so analog isn't necessarily immune.  And IF the EWPro is using a VCO, then many reservations regarding various digital approaches go out the window.  What's good for the analog goose should be fine for the digital gander (if properly implemented).

"No VCO."  - Thierry

Thank you.

"Oscillator signals go to LM 393 comparators to be transformed into square waves."

And perhaps level shifting?

"Then directly mixed in XOR for highest register. At the same time divided by 2 and by 4 in a quad D-Flipflop. The /2 and /4 signals mixed separately in different XORs."

Just as I guessed.

"The /1 mixer output signal is then used for the tuner output and the pitch CV generation. But all three pitches go to the register switch which will decide which of the three goes to the wave shaping."

Makes sense.

"BTW: this had already been written in these forums..."

Why didn't you say so in the first place?

FredM was shocked about the (in his eyes) destruction of the beautiful oscillator waves by the comparators which transform them into simple +5V square pulses and about the further processing in cheapest CD40xx digital cmos ICs.

BTW: Indirect sound generation via VCO as done by Moog in the Series91 and Ethervox Theremins is not forcibly inducing latency as these both excellent instruments prove. As I could observe, not only in the Theremin domain, latency comes mostly from lousy coding in digital designs, be it by over-filtering things to mask design flaws or by proceeding samples blockwise instead of real time, the hardware requirements are much cheaper then...