Can I get a 16-bit value from paired 8-bit A-D's?

[abradarba]

Since many 8-bit A-D converters are available, what would I need, generally, to combine 2 8-bit parts to obtain a 16-bit value from an analog signal? Would I divide the analog value on the front end, or divide it after conversion, assuming my circuit split the signal in the middle? General answer will be fine - thanks!

 

[ve7brz]

The circuit you would need is more complex than you imagine. Your 'first' A/D coverter finds the value to 1 part in 256, and your second must find the error or difference between the 8 bit conversion and the actual input to 1 part in 256. Consequently you need two A/D converters, one D/A converter, an op-amp or two and some precision resistors. A 10 or 12 bit A/D is simpler and cheaper.

 

[Lewish]

Texas Instruments has some pretty reasonably priced 16-bit A/Ds. Have you considered them? What sampling rate do you need?

 

[electricuwe]

ve7brz,

you missed the Sample-and-Hold circuit in your description.

 

[ve7brz]

Sorry electricuwe,

I forgot to mention that.

 

[cbarn24050]

Hi, 16 bit is allmost impossible to get for someone of your level of expertise, 12 bit is a more realistic goal.

 

[abradarba]

Yeah, "level of expertise" is putting it pretty generously. Someone else asked me this question, and he was interested in model rocket telemetry using some 8-bit components he knew of which had high bandwidth - he though 1 gigahertz might have been the number. So both of us are out of our league. Thanks for the responses - sounds like with all the requisite circuitry, (sample and hold, for heck's sake...), there's a bit of a bottleneck outside of the high speed of the A/D's.