[sdiy] VCO reference voltages

[rob at emulatorarchive.com]

Hi
The implementation of precision voltage regulation for the "critical" 
on-board voltages of a VCO, as per Tony's recommendations, does make a 
worthwhile improvement to VCO stability. I have reworked a number of 
legacy designs from ARP, E-mu Systems and Roland with 10V and 5V 
precision regulation - with positive and measurable outcomes. I have 
also experimented with potting the VCO core and found this to have 
advantages as well, especiallty for the ARP design.

In terms of reduced voltage swing for Frequency, it is easy anough to 
compensate for this in the frequency voltage summer. Or in the case of 
the E-mu Systems VCO, who needs 60kHz anyway...I also like dedicated LFO 
designs where the waveform is more accurate at low frequencies.

There are a number of re-worked VCO design on my new web site at 
www.amsynths.co.uk
I plan to load the Eagle CAD files  so anyone can download the 
schematics and boards and make them.
I usually use REF01 type voltage regulators with line regulation of 0.006%/V
Line regulation on a 78L12 is 18mV.

Regards
Rob
www.amsynths.co.uk

David G. Dixon wrote:
> Thanks for all the responses.  I think I've formulated a design philosophy.
>
> Here's what I'm thinking:
>
> 1. Use voltage regulators for "critical" on-board voltage references such as
> thresholds, servo currents, and for critical adjustments in shapers, such as
> triangle or saw alignment.
>
> 2. Use rail voltages for coarse and fine tuning pots.
>
> My rationale is that the tuning knobs are going to be twiddled constantly,
> so if their end points move slightly as a result of PSU loading, this is not
> important, and will go unnoticed.  After all, the PSU load is largely fixed
> unless modules are added or removed from the cabinet, and this doesn't
> happen during operation.  Hence, the rail voltages should remain fixed to
> within 1 or 2 mV during actual operation.  It is only when modules are added
> that the rail voltages may change measurably, and only then when the PSU is
> at or near its drive capacity.  However, all this will do is change the
> range of the tuning pots very slightly.
>
> If the rail voltages change measurably as a result of just twiddling knobs,
> then the PSU is probably at or near its drive capacity, and I don't think
> that module design should have to compensate for this.  Hence, it is really
> only inexactitude in the rail voltages themselves that should be compensated
> for, and this is accomplished by using references for those voltages which
> should take "exact" values on the board.
>
> The other downside of regulating rail voltages for the tuning pots is that,
> because regulators only regulate to within 1 or 2 volts of the rails, using
> regulated voltages for the rails will limit the range of the pots,
> particularly in a +/-12V system.  This is not necessarily a problem on the
> upper end, but it could be on the lower end unless the VCO has an LFO range
> switch.