Now that you know about NFCR and ASR, you can use them to choose an ADC that's right for the job you want it to do.

Start by asking, "how big is the signal I want to measure?" and "how fast do I expect it to move?"  Those will give you a minimum slew rate.

Then ask, "how often am I willing to accept at least 1LSB of error?" There's no point in paying for an ADC if you expect more than 1LSB of error in most of the readings, but are you willing to live with 37%?  Honestly, that's good enough for a lot of things ADCs are called upon to do.

Once you know how often you're willing to accept errors, decide how big you want your LSB to be. It's much easier to get 1-in-500 million at 8-bit resolution than at 16-bit resolution.

Once you've chosen an LSB, run the numbers on ASR for the minimum slew rate you need. If you need an aperture time in the low picosecond range, talk to the bank and see if a mortgage on your house will cover the cost of that ADC (they exist, but generally cost more than the building around them).

Also run the numbers on NFCR and find out how quiet the circuit will have to be. When you subtract the cost of the ADC from your mortgage, can you still afford equipment capable of measuring at that level?

(bonus points for remembering that you have to add RMS noise and ASR to get the total error, and the sum will be 1.414 times either error by itself, assuming they're equal)

You may find it necessary to relax your initial expectations about resolution and error rate to find parts you can actually buy and circuits you can actually build.

Running the numbers and tradeoffs will give you specific things to look for when you start comparing ADCs:

  • How much does it cost to hit your minimum slew rate at various resolutions?  
  • How much are you willing to trade increased error rate for increased noise tolerance or slew?  
  • Which ADC gives you the best resolution and slew rate in a given price range?  
  • Is the difference in performance between two chips worth the difference in price?

The answers will change from one project to the next, but knowing how to ask the questions will give you a chance of finding an ADC that actually does what you expect it to.

This guide was first published on May 23, 2019. It was last updated on May 23, 2019.
This page (Putting the numbers to work) was last updated on Oct 22, 2020.