[sdiy] Tonewheel relics

[Rutger Vlek]

I thought about that. The gear ratio's are crucial to the tuning of the
tonewheel generator, and I feel that varying belt diameter (tension
related, temperature dependant) could become an issue. Though, I have no
way to quantify this instinct.

As for 3D printing, this is really a nice case where it's not perfect yet,
as far as I can tell. I've asked some experienced people, and what I heard
was basically that the composite nature of 3D printed objects results in a
not-so-homogene material, that has unpredictable wear properties during
applications with friction. Polishing changes the surface a bit, but lack
of internal homogenity is supposedly a problem for gears, especially when
running at higher RPMs (more friction). From memory I recall the drive
shaft of the central AC motor in the Pari runs at around 1200 RPM. It
connects with a belt drive to the first tonewheel drum shaft, which then
connects with gears towards the last (12th) tonewheel shaft, progressively
slowing down each drum such that identical drums given a semi-tone tuning
difference. The last drum connects with a belt to the vibrato scanner.

As for PLL loop: I've never designed one, but it has my interest. What
amount of complexity am I looking at? Can someone refer me to an example
circuit schematic? Each tonewheel drum contains several octaves, the
highest octave as (I believe) 16 dents on a single revolution. So that
would be the preferred source for PLL feedback, I guess. The other octaves
automatically align. The drums are spring-coupled to the gear shafts by the
way, so the PLL loop may be affected by that (bouncing at startup?).

I do imagine this would be a HUGE step forward for the Pari in becoming
tunable and way more silent than it is.

Best,

Rutger





2017-10-24 0:48 GMT+02:00 Sarah Thompson <plodger at gmail.com>:

> Would belt drives work? They have huge advantages over gears if you need
> smooth transfer with no cogging or backlash.
>
> On Mon, Oct 23, 2017 at 4:34 AM, <rburnett at richieburnett.co.uk> wrote:
>
>> @roman and Richie: thanks! From my limited experience with bldc's I
>>> feel they would be costly, especially a 12-way driver system for them.
>>> Would they need hall sensor feedback to get the tuning accurate
>>> enough?
>>>
>>
>> Yes, probably not cheap.  I don't think you'd need hall sensors to detect
>> the field, but you would need to ramp the frequency up at a controlled rate
>> to allow the mechanical bits time to accelerate up to operating speed.
>>
>> The pll sounds nice and classic. I'll look into that. At present I'm
>>> not sure if the pick-up  coils of the generator could handle more
>>> loading, e.g. As required for closing the loop. But a high impedance
>>> buffer would probably be ok.
>>>
>>
>> Take the feedback for the PLL from the highest frequency that you can.
>> The loop filter in the PLL has to filter out the ripple in the compared
>> reference and positional feedback signals.  The higher you can make the
>> reference and feedback signal frequencies, the larger the bandwidth of the
>> loop filter can be, and the quicker the PLL speed controller will settle.
>> i.e.  Don't use a reference signal that is just one pulse per revolution,
>> as this can mean you end up with a very sluggish control loop!
>>
>> -Richie,
>>
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at synth-diy.org
>> http://synth-diy.org/mailman/listinfo/synth-diy
>>
>
>
>
> --
> [s]
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20171024/d6e7d740/attachment.htm>