[sdiy] Static ADC and DAC recommendations?

[cheater cheater]

Hi guys,
I think I've found it. I tried my luck yet again at google, and this
time I stumbled upon the Data Conversion Handbook by Analog Devices on
Google Books: http://tinyurl.com/aoalmp3

(full url: http://books.google.de/books?id=0aeBS6SgtR4C&pg=PA182&lpg=PA182&dq=flash+adc+1975&source=bl&ots=6zRgS6n6Y-&sig=Do-k6ZFLeMV_ild9mu-ePHnQZfs&hl=en&sa=X&ei=gZj4UOmuH47HsgbusIGYCA&ved=0CEIQ6AEwAw#v=onepage&q=flash%20adc%201975&f=false
)

It tells the history of data conversion, starting with
electromechanical flash converters from 1921, going through cathode
ray tubes that used a specially shaped assembly at the end of the tube
to do either binary or gray-coded output (you have to see the
drawing), tube based comparators, tunnel diodes, up until comparator
ICs and the first ADCs. They mention an AMD 6688 4-bit quantizer. The
really nice thing about this is that it has no latches at the output.
It achieves stability by allowing the comparators themselves to
sample-and-hold, and this can be turned off: "If LE is driven high and
LE' is driven low, the quantizer is in the sample mode and operates
like a low-gain, high-bandwidth amplifier. When LE is driven low and
LE' is driven high, the quantizer will hold its existing digital
binary output word". In fact the application note schematic contains
an external latch chip. This looks almost exactly like what I'm
looking for.

Searching further in that book, starting page 203 the
one-comparator-per-bit architecture I mention is described under the
heading "Serial Bit-Per-Stage binary and Gray Coded (Folding) ADCs" as
already existing in 1949. In fact a small modification to this
architecture yielded the 105 Msps, 14-bit AD6645. BTW, that part is
not on Google Books, but the whole book (without the cool hand-drawn
pictures though) can be found at:

http://www.analog.com/library/analogDialogue/archives/39-06/data_conversion_handbook.html

However, the only chips the book mentions with the serial
stage-per-bit architecture are from the 90s, which means they're
fairly complex and integrated and don't allow tricks like the AMD
AM6688. So unless I can find something else, that AMD might be my best
bet.


As an aside, the AD book says that Gray code was invented by two
people named Gray.. Elisha (the original inventor of the telephone)
and then Frank.. Apparently there are no family ties between the two?
That's crazy.

Cheers,
D.



On Fri, Jan 18, 2013 at 1:03 AM, David Moylan <dave at westphila.net> wrote:
> The ADC0820 is dead easy to clock.  I say get a CD4046 and crank it up.  I
> found ADC/DACs a bit daunting until I actually tried working with them.  Way
> easier to work with than I imagined.  I ended up building a digital
> delay...more on that in the next few months.
>
>
> On 1/17/13 3:47 PM, Paul Schreiber wrote:
>>
>> You do realize that if you do not have SOME sort of clock/synchronous
>> latch, the ADC outputs will glitch to high holy hell and back.
>>
>> Look, just get a 65 cent PIC controller, program it to loop forever and
>> bit-bang the proper clocks and be done with it.
>>
>> Paul S.
>> _______________________________________________
>> Synth-diy mailing list
>> Synth-diy at dropmix.xs4all.nl
>> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>>
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy