[sdiy] Moog Memorymoog clone almost completed

[Michael E Caloroso]

The trend I have been seeing is RCA 405x have a high failure rate.
Those 1st generation RCA components did not have a long life.  The
majority of CMOS failures I have seen were RCA parts.

The absolute maximum rating for CMOS inputs is VEE-0.5 < Vin < VDD+0.5.

Some polysynths have a design flaw in that the opamp buffer after the
DAC lacks clipper diodes on their outputs.  DEMUX circuits using 405x
often are on VDD/VEE power rails other than +/-15V.  That opamp buffer
is on +/-15V rails, and on power transitions the state of those opamps
is indeterminate.  In other words, they could generate voltages that
can violate the CMOS absolute maximum rating.  While they are brief,
they can degrade CMOS over time.

Other polysynths use a flawed design for the supply rail to the 405x.
I uncovered this in my Oberheim OBX and OBSX, and it also exists in
the OBXa and Matrix6/1000.

I am not convinced that S&H caps are a contributor to 405x failures.
The 405x are sensitive to input voltages relative to supply voltages;
the S&H caps however are on the OUTPUTS of the 405x.  Even if a 405x
output were to be active during power transition, the S&H cap would
simply charge to the output voltage.  That's not a condition for
damage.

MC

> I have seen lots of 4051 problems on other polys as well.
>
> For years I've been suspecting that on power down the S&H caps hold
> their voltage longer than the PSU. I think CMOS chips don't like signal
> voltage without a supply voltage.