Martin Latecka wrote:

 >Anyone have an idea on how I can measure capacitance?
 >I wanted to charge a cap with a constant current and measure
 >the time it takes to reach Vref. I have some _serious_ doubts
 >about this though.

  Why bother about constant current?  You are measuring time constant,
which is expressed adequately AND LINEARLY in terms of the charging
resistance, the capacitance and the (logarithm of the) ratio of
initial to terminal voltage.  If therefore you measure the time it
takes for the voltage to decay to a fixed proportion (resistor divider;
comparator) of the charging voltage (regulated and stable, but not
precise) you only require precision of the three resistors and timing
reference, and short-term stability of the charging voltage.

  The only design concern is how you implement ranging, as you need to
minimise stray resistance and capacitance, the former when measuring
large capacitances, and the latter, small.  If you are only concerned
with a small range, this is no problem.  A version of manual ranging
selection is to built alternate probes for different ranges.

  I have considered (but not built) a probe for picofarad and
fractional-picofarad measurements only, to be built using SMD into the
probe itself.  The circuit and components for values upward of 100uF
into the multiple mF (millifarad) range would be quite different, such
as VFETs for switching instead of CMOS.

  Cheers,
        Paul B.