Let's Design and Build a (mostly) Digital Theremin!

"My main fear is whatever perceived weaknesses the stripped down model might possess will be interpreted as general weaknesses of the digital approach.  A corollary fear is that by not putting my absolute best foot forward to begin with I'll be alienating future professional, more discerning / demanding buyers." - Dewster

I know those fears well.. My bias has been to hold out and go for "top end" first, then, based on the reputation established by this, bring out "lower end" products at lower cost.

But, as you know - it is a lot easier to start will low-end and work up.. I think I made a mistake, because low-end theremins would have been bringing in money while I was developing the high-end.. And I think now that the small size of the market means that any new "high end" theremin would be regarded on its own worth, without too much "bias" because of lower price "starter" product which was not particularly wonderful.

The one thing that could, I think, impact on acceptance of any high-end theremin would be if there were reliability or support issues with the low-end products.. But provided the low-end worked reliably, and performed the specified function/s, I think people would not be put off going for your "flagship".

I am actually looking at putting a moderately priced solid state Lev clone on the market this year - something much less sophisticated than my "flagship" - and am grappling with the issues of price and funding too.. Or I may just offer special one-off builds based on a standard lev-clone board. I could probably do these for about $300 to $500.. But even this is above the price that many would be happy to pay - I think a target price of $250 maximum would appeal to more (go above that, and people are likely to opt for paying the extra to obtain a EW I think) - but that price could only be possible if a batch of 10 or more are made (actually, cost savings really only kick in for larger quantities - say 20+ or a run of 100 to make any reasonable profit) - which means money up - front.

Fred. 

The one thing that could, I think, impact on acceptance of any high-end theremin would be if there were reliability or support issues with the low-end products.. But provided the low-end worked reliably, and performed the specified function/s, I think people would not be put off going for your "flagship".  - FredM

Fred, thanks so much for your thoughts!  You're really helping me come to terms with this.  It seems it's all to easy to slip into "flagship debutante ball" mode and proceed unquestioningly for a long period of time - the lion's share of demand is always on the low end of price.

"the lion's share of demand is always on the low end of price." - Dewster

Yes - it will probably be..

And I think what I said before has some validity.. BUT..

"The lions share" is probably the wrong choice of words, IMO.. Its more like the Hyena's share.. the cheap 'dregs' which the majority of theremin producers are scrabbling and fighting over - the under $250 market where the price is placed so that it is far enough away from the price of a "real meal" to attract those really cannot afford an EW, or those who are just looking for a "snack" or those (few) who really dont like the sound of the EW.

The "Lion" is less bothered by this kind of "market" - In its natural state (before man screwed its kingdom up) its "share" was almost whatever it wanted.. It would go for the kill and the hyenas would fight over what it left behind.

I suppose I have been targeting what I see (or saw) as "The Lions share" - The vacuum in the high end market, the hole left which was once filled by the E-Pro or the E-Vox or even the 91 series, And also filling a "demand" which will only appear when a new instrument with greater capabilities and better sound than anything before it, appears and sets a new standard.

But I think I was probably wrong. I now think that such an instrument probably wouldnt create a demand or alter expectations one iota, for the simple reason that the number of people who are REALLY interested in theremin mastery can probably be counted on my fingers - and only these people would even notice that the "bar" had been raised - and as "raising the bar" would, at least to start with, cause a raising of the price.....

IMO there is no "good" section of the market in which one can make money - the <= $250 market has a lot of competition and possible profit margins are tiny - one needs large volume of sales to make any money in this area, and that is not going to be easy with theb number of competetors in this section..

The Mid price range is competing with the EW - the most established theremin on the planet.. And this theremin is perfectly adequate for most thereminista and wanna-be thereminists..

It is, IMO, only the "high end" where there is a possibility of earning enough money to make the effort worth while - If one can, through clever design, cut the costs without impacting quality, and give features people want but cannot get from a mid-end theremin, then it could be worthwhile ... But only on a "commission" basis - there simply is not enough demand to go into production and even produce 20 of these theremins.

*Back in 2010, George Pavlov (the designer of the T-Vox) told me (and Lydia) this when both of us were trying to persuade him that there was a market - He explained the economics to me - that even though he had a theremin people wanted which had a reputation and was well known, it simply wasnt worth while (economically) to do a production run of another 100 units - that the profit margins made no sense for a run of less than 1000, that there wasnt a market for 1000 so a run of this size would cause a financial loss.. He was even doubtful that a run of 100 could  break even...

This is someone who's theremin Lydia plays (his production was actually for her, and perhaps to explore the market)..

I should REALLY have listened to him - But I didnt want to hear what he had to say!

Fred.

*Please understand that my discussion with George is extremely "paraphrased" - I do not remember exact numbers we were talking about etc.. it had been a long day, we were having a meal (or at least I was.. I think..;-) in the Festival Hall after the show.. So I may well have some details wrong!

"The lions share" is probably the wrong choice of words, IMO.. Its more like the Hyena's share.."  - FredM

Ha ha!  And on the low low end it's maybe "the dung beetle's share"!

My wife works, we have some (slowly evaporating) savings, and I don't intend to support myself and put kids through college or anything like that by building and selling Theremins, so the financial imperative is kind of relaxed (though I need to make something, even minimal private health insurance in the good ol' US of A is a killer for us semi-oldsters - that and the heinous property taxes in NJ).  I see this as starting off more like guitar building (but hopefully easier / faster / and involve less or ideally no fine work).  Going the series tank / mostly digital route should almost completely eliminate any fine tuning at the end (though I haven't messed with temperature compensation yet).  They might be simple enough to make one a day (not that the demand would likely be that high), and if I could make $100 clear on each one that would be cool.

Got some 34 AWG wire in the mail today, it's not quite as thin and unmanageable as I'd imagined it might be.  56 meters on a 60mm diameter schedule 40 pipe (~300 turns) should give 4 mH.  Got some clear heatshrink too, it looks very thin (it's the type used to hold battery packs together).  If an air core transformer will work in my AFE and it improves temperature stability, that could be a good selling point for any higher end models ("genuine air core inductors give you that organic RCA sound!" - just kidding).

If I put a digital divider on the AFE and used the parallel capacitor for additional phase shift (ala the Smirnov D-sensor, similar to the "BUDGET" reference design in my latest spreadsheet) I could probably use an inexpensive processor rather than an expensive FPGA for brains.  Anyone know of a cheap processor dev board capable of decent audio D/A?

[EDIT] With this Altera board going for ~$20 I think I'll stick with FPGAs.  Now all I need is an inexpensive D/A plug-in board.  There's this for $5 but it's only 12 bits mono.  There's this for $15 which includes headphone amp, but that seems like an awful lot to pay for D/A.

Dewster,

That Altera board does look like great value - What does one need (minimum) to start developing using these? Is there a free (or low cost) development environment which is any good? - Presumably a standard Jtag programmer is all one needs to program the thing, but can one do RT debugging etc over this... Any info you could give would be useful, as I am seriously looking at using FPGA for a project only loosely related to theremins (certainly not thinking about changing my direction when it comes to theremins - The Lev designs and voltage control are my focus at the moment, and I dont intend to deviate from the path again, LOL ;-)

The 12 bit DAC, to me, looks good - What I am about to say will come over as blasphemy to purists - but I do not believe one needs more than a 12 bit waveform from a Mono musical instrument - Sure, 16+ bit is needed for the complexity one gets in composite audio, but for single waveforms, 12 bits is fine.. I have a 12 bit Yamaha TX16W sampler, and it has a wonderful clarity and warmth, particularly when playing solo instruments - When I load this with a solo sound I have extracted (converted) from 16 bit or higher sample, and play it back directly (bypassing the TX16W filters etc) I cannot tell the difference between the 12bit and the higher resolution samples..

The above is NOT true for polyphonic samples - the more "layers" are added the more noticable sampling resolution becomes.. But for a signal having only one fundamental frequency and its harmonics (such as a mono-synth or theremin) or even for signals with a few frequencies and thier harmonics (such as a multi-oscillator mono-synth) I do not believe one needs more than 12 bits -

The above is with the caveat that all dynamic processing comes after the D/A - As in, the signal from the D/A is just a waveform normalized to the full 12 bits - If one wants digital processing to include the dynamics, then one needs 16 bits minimum (I would say more) or the noise floor becomes a pain in the butt..

Fred.

" if I could make $100 clear on each one that would be cool." - Dewster

If you could make a good digital theremin which was retailing at $250, and make $100 clear on each, you would have good reason to feel that this was "cool" ! LOL ;-)

I recon that, as a "manufacturer", one is doing well if, after all expenses etc, one gets 25% ($62) of "clear" profit from what your product sells for.

Its somewhat different as a "cottage industry" and true "clear" is much more difficult to really determine - A major cost is labour - If you do this yourself, and do it for "free", then you may get "$100 clear" - But this is looking at the true situation with rose tinted glasses - If you were to factor in your time, even at minimum wage, the true "clear" will be a lot lower.. (probably in the red, LOL ;-)

Not that it matters (except to the Ferengi) - I would far prefer to spend 10 hours crafting a theremin and getting £1 per hour, than to be serving in some supermarket for 2 hours earning £5 / Hour, or developing some weapon of murder at £100 / hr.

Fred. 

"That Altera board does look like great value - What does one need (minimum) to start developing using these? Is there a free (or low cost) development environment which is any good? - Presumably a standard Jtag programmer is all one needs to program the thing, but can one do RT debugging etc over this..."  - FredM

You'll need Quartus II Web Edition.  They give you everything you need in that.  USB JTAG programmers ("USB Blaster") are available on eBay for ~$10.

Too bad there isn't a bit of user Flash on that Altera board, I suppose I could add some via I2C.  May need a base board to solder everything to anyway and put supply filtering, I/O, etc. on it.  I'm thinking of having separate tiny drive/sense boards at the antenna ends connected via small wiring harnesses.

I agree with your 12 bits comments completely.  I think I'll go with the second 16 bit converter - it has integrated headphone drive, and perhaps more importantly digital filtering before the DAC, so I don't think I'll have to do any complex steep filtering in the FPGA.

It might not be too expensive or complex to add another axis to the volume side (another antenna + coil + phase detector) - something that would significantly differentiate it from other offerings out there.  (FEATURE CREEP!)

"I would far prefer to spend 10 hours crafting a theremin and getting £1 per hour, than to be serving in some supermarket for 2 hours earning £5 / Hour, or developing some weapon of murder at £100 / hr."

Amen to that.  It's amazing how many in our profession have blood on their hands.  If you count the indirect blood (your company does "defense" work but you don't directly, etc.) it's probably > 99%.  They really ought to make more of this in engineering ethics courses you have to take in college (we just read Frankenstein, and the shuttle O-ring fiasco was presented as more of a whistle blowing cautionary tale).

 (FEATURE CREEP!)

LOL ;-)  - Stop yourself now!  ...... But its so damn difficult to do that, isnt it ? ... You have everything there to cope with the extra signals, all you need is to add a few cheap bits and therebye gain all these extra things FREE ;-)

LOL ;-)  ... And before you know it, you have something which looks like a Moog Modular, and costs more than an E-Pro!

;-)

"They really ought to make more of this in engineering ethics courses you have to take in college" _ Dewster

LOL ;-) Why would they do that? - The last thing "they" want is for any student to ACTUALLY start thinking about things like "ethics" - The main demand for engineers is driven by demand for "defense" hardware - Industry sponsors would back away rapidly from any college which ACTUALLY did "ethics" .. and 'worse'..  IF such exposure to ethics did influence the students, it would soon show up on the statistics.. "50% of  this ethical universities graduates are unable to find employment because they were screwed-up by the ethics cirriculum - and a high percentage of these graduates cannot be regarded as responsible individuals or True Americans*(Britons,Europeans,Russians..insert any nationality / race / culture / religeon here to taste) as they exhibit disrespect for our establishments and are unpatriotic"

Spent the day slaving over a hot breadboard, trying to get an analog pitch-only Theremin with series tank oscillators to work.  Am getting similar linearity compared to the digital prototype but can't seem to change the sensitivity much.  You analog guys got it tough!  But you've got it easier when it comes to drift - both the pitch and reference oscillators seemed to drift well together - the null distance didn't change much even though the room temperature was changing a fair bit.

The digital approach has a lot of rigamarole up front, but after that it's so much easier.  Nulling is a snap, sensitivity and operating point are abstracted somewhat away from the vagaries of the analog front end, and linearity can (hopefully) be improved using simple math.

So anyway, UNCLE!  And back to the digital realm for me!  This exercise has even managed to turn me off to the notion of using an analog oscillator in the AFE.

[Product idea: pitch only Theremin with built-in headphone amplifier.  4 pots on the front: pitch (null), waveform (odd / all harmonics), coupling, volume.  Headphone out can also do pitch preview.  Is there any Theremin out there that has a coupling control?]

" Is there any Theremin out there that has a coupling control? " - Dewster

LOL ;-) .. Not "out there" - But I have mentioned it and used it- it just seems so damn obvious... A little audio mixer for headphones which can feed preview and main "to taste" (and an audio input socket to allow one to feed an MP3 player into, so you can play along with a backing).

I think I posted this recently on Levnet, before I resigned from it.

Fred.

Yep.. heres the posting:

-----Original Message-----
From: Sent: 02 February 201306:36
To: Subject: RE: So..... I got my Leftie E-pro yesterday.....

“! The headphone jack is useless, unless you need to practice in silence.” - Amey

As in, the output level is too low?  If this is the case, you could run this headphone output into a small mixer which has a headphone amplifier built in.. (as in, boost the preview output level)

This is something I have put into my prototypes, and have on the one I use.. A small in-built mixer which allows preview and theremin output signals to be mixed ‘to taste’ to the headphones.. (I also have an audio input which can be mixed in) – A lot of the time, one wants to practice using headphones – or certainly do if those you live with aren’t  tolerant – the ability to mix the preview and main outputs is, to me, an obvious requirement – Even for performance, I would imagine it would be useful as the PA speakers may be far enough away that there is a time-lag between the preview and the returned sound.

Fred.