Modeling Hammond Tonewheel Organs inside of Reaktor Primary or Core?

cs5947

https://community.native-instruments.com/discussion/comment/9171#Comment_9171

Thanks :)

Okay, fine. I’ll link just three tones instead. I checked inside Frederik Kilander’s setBFree, and it appears to have 3 tonewheels linked in each direction, -3db, -5db and -10db the farther the generator is from the pickup. It’s in tonegen.c

I’m a little afraid to mess around with the Core Cell sine waves because they seem really complex, so I’ll just stick to Parabolic waves for now. These are similar to sines but they have slightly more harmonics. But they’re also kind of linear, so I guess I am just better off messing around with core cell sines for nonlinearity.

The service manual is very good and detailed indeed. It offered info for the order in what tonewheels are arranged, the transformers (they make the sound a little warmer but i really don’t know how you can model these inside of Reaktor) and a bunch of other stuff I’ll have to keep in mind.

Thanks again! Your input helps a lot

colB

https://community.native-instruments.com/discussion/comment/9224#Comment_9224

Yes, the gear ratios are part of what determines the tuning, also the frequency of the main drive (the gearing is always proportional to that), but also, the tone wheels themselves have more or fewer 'petals', those also define the frequency.

Looks to me like most pairs of tonewheels on the same gear are... maybe one with very few 'petals' more like a octagon or pantagon or whatever, and another with many many little half circle like petals... the one with few will give a low frequency and the one with many will give a high frequency.

Both of these are also likely share a small amplitude amplitude modulation at the gearing frequency. There are 12(iirc) different gearing frequencies, so 12 different amplitude modulations... might not be sinusoidal, because they are mechanical due to tolerance of component manufacture - disks not being completely round and flat, tolerance in the fit over the shaft etc. So the parasitic AM will be at one of 12 frequencies, but the might all have different phases and/or different waveshapes.

To understand what might be happening to generate waveshapes, think about some polygon with straight sides... a pentagon say. rotating away...

there is a pickup in a fixed position so that its close to the rotating edge. the closest point is when a corner of the polygon passes the pickup, the furthest point is when its half way along an edge. as the disc rotates, if you imagine a plot of the distance between the edge of the disc and the pick up, it might look something like a row of bowls lined up, a bit like a parabolic wave, but full wave rectified and inverted, then with a positive offset applied...

then you have to factor in the fact that magnetic field strength is inversely proportional to the square (some say cube?) of the distance... so if you took your rectified inverted wave, and took its square root, making sure to remove DC and scale it appropriately, you get something back that's a bit sinusoidal

There's a bit of a snag though the pickup is not a point its a physical thing, and its magnetic field is three dimensional, and it will also have some sort of coil to generate current... so you get effects like hysteresis and have to factor in that its an area so there will be some integration effect happening in the intersection of the field and the disc edge (that's the pic in one of the papers you posted).

Why are some wheels polygonal, while others have petals?

At lower frequency notes the ratio of the field size of the pickup to the physical length of the wave cycle is smaller and maybe can be ignored. At higher frequencies, there are at least two possible problems?, the sides would be so short that the difference between the longest and shortest distance to the pickup would be tiny and maybe noise becomes too big a problem when enough gain is applied. The other is that the wave cycle length as a physical part of the tonewheel is similar or smaller than the size of the effective part of the magnetic field of the pickup. So the integration effect of the pickup becomes more significant?...

I think the petal shapes combined with the pickup integration effect at higher frequencies probably give a similar waveshape to the polygonal tonewheel at lower frequencies (on the same gearing) and the adjustment of the shape is to account for the more dominant effect of the pickup magnetic field. Although it would be surprising if the designers didn't then also leverage this to create different harmonics too.

I'm thinking maybe stuff like this

or

or maybe

some slightly skewed, slightly wonky sinusoid :)

(that third one uses @ANDREW221231's potato wave as a starting point)

first thing I guess would be to do some math to work out the function for the distance to the pickup as a polygon turns near it :)

Also get some good pics of tonewheels, or even better, some scope plots of their output.

ANDREW221231

https://community.native-instruments.com/discussion/comment/9233#Comment_9233

hey, that's one pretty good use for a bad sine wave! the potato could even be desirable over alternatives for trying to emulation behavior or of electronics where even order harmonics are present, or desired to be

your description of the tone wheels reminds me a bit of the process for power supply ripple modulation in emulating electrical circuits, which is is generally worthwhile. once you pay attention, everything that's electronic and makes sound has ringing at that frequency somewhere inside...do we try to follow the best power supply ripple smoothing circuit designs? surely some of the nicer kit had this effect well below perceptible limits. how many radio shack specials owe some beloved quirk in their sound to one or two design choices where parts where parasitically pulled our of parameter?

anyway you just mortared this poor fellow with an information bomb i suspect he will need some time digesting, its got me thinking, but no longer about organs, so i'll cut this short here, but there is a high likelyhood i will send you something of interest privately later

ANDREW221231

i don't know jack about physical modelling, but i do understand compulsive interests pretty well having had a few of my own in my time. its almost beyond belief how much learning a thing like this will carry you through. and if you continue to bash your head against the wall and always look for new angles of attacking the problem, eventually it will yield. even if you are limited in your technical knowledge

im reminded of an anecdote about opening a walnut: you can crack it deliberately with a hard metal tool to obtain the meat. this is maybe like Chet Singer deliberately applying his very particular set of skills (which you would do very well to study, btw following signal path with a scope will teach you everything about how it works without touching a formula)

the alternative is to let it soak in some water in your hand for a while. if you continue to ponder like this eventually the shell will soften and fall away on its own. for me, who started this hobby knowing nothing, this has never failed to wlrk

anyway, i understand you've sort of hit on physical modelling as the way to go, but i would still recommend trying a few others. for instance im confident the sine bank could do a perfect hammond pretty easily (while being generally less interesting), its nice to be able to compare. you'll get to that critical mass point of understanding much faster, where knowledge starts to compound

nuromantix

Elecric Druid has lots of interesting info.

https://electricdruid.net/technical-aspects-of-the-hammond-organ/

cs5947

https://community.native-instruments.com/discussion/comment/9233#Comment_9233

Not sure how to emulate the wonky/skewed effect but I think it looks usable!

cs5947

https://community.native-instruments.com/discussion/comment/9398#Comment_9398

“anyways you just mortared this poor fella with an information bomb” LOL

colB

https://community.native-instruments.com/discussion/comment/9736#Comment_9736

Not sure how to emulate the wonky/skewed effect but I think it looks usable!

Problem is that you need to have data to know if the real thing is anything like that at all - otherwise it's just total guesswork, and probably wrong. No point in trying to emulate something when you don't have access to the thing you are trying to emulate.

I guess if you build the rest of the module up accurately, you can to A/B tests with different waveforms and pick the one you like

Studiowaves

Dood, the tonewheel needs to constantly spin and you need drawbars that adjust the volumes of the tones and when you play a note you have to fade in the already running oscillators that simulate a tone wheel. The oscillators need to by syncronized like a tonewheel too. If you don't do this your notes will rarely play in phase. Example, play a low c then the c an octave above. If the notes are coming from the running tone wheel you get different volumes depending on when you play the notes. The harmonics like the second is the same as the c an octave up. That harmonic needs to be the same as the tone wheel, if you free start any notes it will sound horrible. Also each harmonic is crossfaded when they get to high, it's called foldback. Once you have that setup you get to simulate a leslie cab which is where most replicates flunk out. The spinning horn has ever changing standing waves in the cabinet. This is nearly impossible to replicate so more power to you if you can pull that off.