Rbe test circuit?

[Martin Czech]

:::So it seems to me that the best circuit to test it is an actual 
:::VCO application circuit, leakage and constancy of the cap are 
:::perhaps of minor importance, since the measurement would set an 
:::ordinary VCO into the kHz range, where the leakage can be neglected, 
:::and you can have many measurements in a small time interval. (i.e. before 
:::the cap undergoes drastical capacitance drift)
:::But I *do* think that the integrator reset time has an impact on the 
:::measurements. But before I can start I'll have to make a little frequency 
:::counter.  
:::

As I posted earlier you can get a good real 5 digit DMM for reasonable
money today.  So the direct measurement would be possible.  OTOH you are
right, the DMM does sophisticated double slope integration anyway, so we
can measure nA with single slope, or in other words sawtooth oscillator.
At least @ 27C the CA3140 should have low bias current, and the frequency
problem could be solved by using a good cap in the uF range. This could be
checked simply with a removeable bridge. Will the integrator drift away?
Board problems can be avoided via air insulation, bend susceptible
legs out of the socket: air wiring. See B. Pease "What's all this fA
stuff anyhow..."

There are big (bulky) 10uF poly-something caps for loudspeaker
purposes on the market, but I don't know about leakage and tan delta.

I think it is very important to discriminate between expo error,
comparator error and opamp error.

By the way: I would recommend the ELV frequency measurement kit, because
it's cheap and can compute F after measuring T, which is a good feature
for our "dc" audio stuff (well, people that use to deal with >5GHz say
that I'm doing dc electronics...).


m.c.