BBD-MN3010 and MN3011- is a MN3101 clock necessary?

[René Schmitz]

The outputs of CMOS FFs have less drive than the usuall CMOS gates. At
least thats what I was told. Take a 4011, make an AMV out of two gates
(used as inverters, crosscoupled by Cs, Rs from In to Out of the NAND). One
can steer this in 1/x fashion with a pair of diodes to the inputs of gates.
Now let this follow a RS-FF made out of the other two gates as a driver. 

I used this circuit, its in the Solina, and in the Dr. Böhm. I don't have
problems with noise, because I used steep input and output filters
(10pole)! There is very much clock feedthru at the direct output of the
BBD. It must be filtered away.

To all those who are tempted to use a non overlapping clock:
<IMO flamesuit=on> The clock signal feeds thru capacitively! This is charge
transfer in a series string of MOS switches. Its a differenciated version
of the clock signal. Ideally the rising clock phase would couple as much
signal to the output as the falling one, they tend to cancel. I think it
won't help to have a non overlapping clock. Rather on the contrary: You'll
get two spikes timely separated, instead of simultaneously (which are of
opposite polarity, and partly cancel). The problem with a clock which
overlaps too much is that the odd numbered "buckets" can leak into the even
numbered (and vice versa). This leads to signal attenuation. I think BBDs
are percieved as noisy because often the input filters are not sufficient
for the sampling theorem,
while its not difficult to filter a 50-100kHz signal away from an audio
signal.
</IMO>

Bye,
 René




transconductance | uzs159 at uni-bonn.de
isfutilepreparet | http://www.uni-bonn.de/~uzs159
obeassimilated.. | http://members.xoom.com/Rene_Schmitz