vc wave select

[tomg]

> The LM3914 outputs go low when activated. So, we add a 150K resistor from
> the -15 volt supply to the summing node of the op-amp. Now the with the
> control
> input at 0 volts, the summing amp output is at 10 volts and all the
LM3914
> outputs
> are low and the CD4532 outputs 000 binary.

Everything is fine up to here. 
 
> As the control input goes from 0-10 volts, the summing amp output goes
from
> 10v to 0v and the outputs go high in order. The CD4532 then outputs
binary
> code in order 000 to 111. You can cascade two such sections to get four
> bits.

In order for this scheme to work the 3914 would have to be in bar mode
and the 4532 would have to detect the first line going high then while it 
remains high detect the next line as it goes high.

The 4532 won't do this, it can only see one high line at a time, all others

can change state all they want and the 4532 will stay right where it is 
until the initial line goes low. Then it can see the next data line that is
high
and change state.

> The design originally appeared in the quantizer module for the Sadin
Image
> Processor,
> an analog video synthesizer. Using LM319 comparators instead of a LM3914,
> the three
> outputs of the CD4532 are fed to the RGB inputs of a RGB to NTSC encoder.

I've never seen this but I'm pretty sure the 319s were wired as window 
compatators and applied a single high data line at a time to the 4532s.

I tried this in a half-ass fashion, just using parts of stuff. The cascaded
4532s
work fine but you need an additional 4071 to gate the outputs. I was able
to
get the 3914 working with the 4532s by using 4049 inverters with a 100k 
pull-up resistor on the 3914 (in dot mode) side.

This works great! But there is a but, this would require 2-3914s,
2-4532s,1-4071,
3-4049s, plus pull-up resistors. Not quite as elegant as 2-3914s and
3-74LS30s
and it can't be used as the wave switcher because you can't change the bcd
output.

Thanks Grant it was worth a try.

-tg