Yamaha CS80

> JL and I were discussing the reason why Yamaha decided to run the
> supply line to the cages.. anyone have a clue why they did this other
> then perhaps trying to battle ground loops?

Or getting a very low GND impedance from board to board.

I have a pet theory about why the CS-synths sound so increadibly good (in 
addition to their unique and brilliant performance-type interface), which 
has to do with GND impedance and linear operation of the VCOs.

Generally, linear VCOs have an advantage over expo VCOs at tracking and high 
frequency stability.
And they have a disadvantage at low frequencies because offset voltages will 
be so much larger relative to the actual CV, when your CV goes down 
exponentially across the keyboard.
(Linear VCOs -> Exponential CV.
Upper limit of reasonable CV range -> rather low voltage at the low end -> 
sensitive to offset voltages.)

Now it's well-known that a *slight* increase of detuning (in terms of 
"cents") in the low range creates a more powerful sound, as it partly 
compensates the natural increase of beat rate that comes with fixed detuning 
(in cents).
The CS synths, as well as the Korg Trident and the early Moog Modular VCOs 
are examples for this; and I verified the positive effect of such a 
paramater by artificially introducing it in an exponential, 
Oberheim-SEM-style VCO some years ago. Switch in a little "linear detuning" 
(i.e., the kind of detuning that would be introduced by *offset* in a linear 
VCO), and a dull sound suddenly starts to "bloom".
But as soon as you overdo it with the offset, suddenly you're lost in ugly 
dissonance.
A fast movement in at low frequencies sounds "wrong" - you can easily verify 
this with pulse width modulation, where you create something similar to 
beating, but at a fixed rate. (Unless you have a keyboard tracking of the 
PWM LFO, like the Oberheim OB8 - but I disgress.)

The CS-80 *has* linear VCOs. I'm convinced it *does* benefit from ever so 
slight offset effects. And it's immensely critical to keep these offsets at 
bay, in order not to leave the pleasure zone and slip into ugly dissonnace.

The PSU rails, like some oversized Star Point, certainly help.

My heart missed a beat when I first checked the PSU voltage on my new CS-80, 
found it about 100mV too high, and thought I'd have to do a lot more 
calibration than I intended.
Then I learned I had to measure it across the PSU rail at the card cage, and 
it was spot on, +15V.
(Perfectly adjusted by Oldcrow, and happily not come loose during the 
transport.)

But that means, we're having a *lot* of offset voltage between the actual 
PSU and these rails. (The PSU has "sense" wires that allow the supply 
voltage to be stabilized to +15V *at* a remote point, which is at the big 
rails of the card cage.
Just think of what a 50mV offset would mean for the VCO's control voltage!
(Asumming - and over-simplifying - that the 100mV are equaly distributed to 
GND and +15V.)
For comparison, we adjust the offsets in the KAS and S&H and M boards to 
something in the 100uV range !!

I *think* (though its more an educated guess and I've never really 
calculated this), that we might be able to compensate even something like 
50mV, but that the *fluctuation*, depending on which M board is adressed, 
how many boards are currently active (current consumption ...), may be just 
in the tolerable range of a couple hundred microvolts *with* this massive 
"distributed starpoint" that is formed by heavy GND rail.
If it were perfect, it wouldn't be a CS-80.
If it were less perfect tthan it is, you could never, ever tune it.
I really admire the GND concept of these ingenious Yamaha engineers, who 
were probably just striving for making the tracking as good as they possibly 
could, with the parts available at the time.
Personally, I wouldn't dare to setup such a concept of percise CV 
distribution in a 180Watt-drawing system without using instrumentation amps 
and balanced CV distribution.

Got to stop writing now. :)

JH.