Building a Hammond Novachord

[Radio Wrangler]

I think we've wandered a long way from the issue of the original post. We seem to be trying to design a plain old electronic organ using the top-octave and dividers technology. There are plenty of these still about. They worked, but their sound was not liked. Lack of tuning differences across octaves made them dull. Ok there are far fewer of the beasties using valves, but the limitation lies in the divider architecture rather than the implementation of the boxes in that block diagram.

What sets the novachord apart is the use of variable gain amplifiers to gate each note before they are combined and treated with the tone controls. This allows individually triggered raise and fall time control for the envelope of each note played.

If you can't get a real novachord to restore, then just get an old divider organ to act as a bank of continuous tone generators, and build a replica of the stuff the novachord had that other organs lacked.

If you want to know what a real novachord is like to play, then there are three approaches: Find and buy one, find an owner who'll let you have a go, or build an exact replica.

Without enough data to go on, the replica may not be feasible. Real ones may simply not become available. But doing anything else will leave the uncertainty of what a real one would have been like.

Building a replica would take too much work for just laying a bit of curiosity to rest, but might be worthwhile if you'd used one, fallen in love with it and wanted one for long-term use.

The Novachord faded away, it was not the beginning of a family od derived and competing designs. Maybe it was too complex and costly for its market? or maybe it just wasn't very effective?

David

Another option is to use a few bog standard 6V (they usually come with dual outputs) toroids one per rack, underloaded they give 6.3V quite well. Less bang if it does go tits up, a bit like local regulation on rack systems.

[Lucien Nunes]

The Novachord was perhaps too novel. There was little market for a device to play 'synthetic music', nor was it an emulation of any conventional instrument. There were numerous approaches to simulating a pipe organ at that date, none very convincing but all in a state of vigorous development. Hammond themselves had a good slice of the action with their electromagnetic tonewheel organs, originally conceived as a pipe-organ substitute although soon found to be rather unsuitable for that purpose. As David says the USP of the Novachord was primarily the envelope control and that came at a hefty price. Compton had already shown that this was achievable at modest cost with electrostatic generators and Welte in Germany had achieved yet better results with optical generators albeit at high cost.

If I were tasked with building a Novachord simulator, in which the sounds and controls were to be as authentic as possible without constraint on the implementation, I would probably start with the dividers and DC keyers from a good discrete transistor / MOS LSI instrument of the 60s/70s. 75% of the work would be done already but with considerable tweakability that a later digitally-controlled-analogue equivalent might not offer. I would only resort to valves for the oscillators / dividers / keyers if research uncovered some very compelling reason to do so. If the desire was to include valves for aesthetic value or to make the project more challenging then that is fine, but it is a separate consideration.

[Radio Wrangler]

I did wonder what market they were chasing with synthetic type sounds coming out of an ornamented spinet style case. History tells us that it was a market which wasn't there. Yet there have been all sorts of odd keyboard combinations used (This from a guy who's had one hand on a minimoog, the other on a Willis)

David

[boxdoctor]

Slightly off-topic - Conn got around the "Phase-locked octave" problem in some of their very last ("Starflyte?) all L.S.I. models by using what they termed a "Drop Clock" which put each sub- octave out of sync (phase-wise) at each division. I cannot remember exactly how this worked, but I remember that this clock ran at a lowish frequency, I guess about 20Khz.
The whole aim was to avoid the clinical "electronic" sound. It was most effective as I recall Tony.

[trobbins]

I guess the phase locked octave 'problem' was somewhat a purist view, with sales growth through the 1950's being increasingly to the masses.

For the top octave generators that purposefully allowed B+ vibrato insertion, some modern added in random noise dither or weak slow vibrato may well be a suitable perception mask, and somewhat recreate mechanical blower fluctuations.

[Radio Wrangler]

Early divider organs used a bank of twelve top octave oscillators. Later on there were ICs that divided an RF crystal oscillator down to get close approximations to the top octave frequencies. These were a lot cheaper than twelve carefully adjusted oscillators, but removed even more 'personality' from the sound.

The hammond tonewheel designs still have to have an integer number of teeth on each wheel, and their gearing still has to be an integer to integer ratio, so the tempering is a bit off because the exact ratios required are irrational numbers.

On the other hand, playing a badly off pipe organ can be a gruesome experience (for any audience as well)

David

[Mike Phelan]

Originally Posted by Radio Wrangler

The Hammond tonewheel designs still have to have an integer number of teeth on each wheel, and their gearing still has to be an integer to integer ratio, so the tempering is a bit off because the exact ratios required are irrational numbers.

Plus the fact that the top octave using 256 teeth on the tonewheels isn't really practical so 192 teeth were used and the top few notes left out altogether.

I suppose that purists might say that Hammonds aren't really 'electronic' organs, as the only electronics is in the amplifier!

Good luck with your Novachord - I've never seen one of these, only worked on the Clavioline and Univox.
I'd go with Lucien's idea using discrete divider organ parts so the oscillators weren't locked.
No neon dividers though! I used to work on the infamous Bird.

[Lucien Nunes]

I guess the phase locked octave 'problem' was somewhat a purist view

It was a very significant limitation but it was one to which the market had become accustomed, since commercial electromechanical generators (apart from the excellent Wurlitzer electrostatics) also suffered from it. Only in the latter days did effective means evolve to disguise the thinness of tone, when moving speakers and delay-line scanners gave way to the bucket-brigade audio delay 'presence generators' of the 1970s that offered a new level of richness. Each BBD would have its own fast clock modulated by multiple slow clocks at different depths, the most complex arrrangements being used in the string synths characteristic of the era. A few of the best sample-based generators could get away to an extent without such complications; the Dereux electrostatics of the 1960s used multiple encoded samples with remarkably lifelike timbre and carefully contrived phasing and frequency offsets to give a satisfying ensemble.