How to build an ADC generating 0-100mV 0.5mV accuracy or better?

Projects General Guidance
21
  1. Arduino Uno is able to do (up to) 16 bit PWM
  2. Charger can measure down to 0.1mV steps - and what? Is it confused if you provide 0.5mV steps or something else goes wrong? If yes how do you know it cannot measure even 0.01mV?
  3. I am not sure if you will be able to output 0.1mV either with Arduino or with the DAC. From my experience digital LOW of Arduino is not 0.000000V but slightly more.
22
  1. Arduino Uno is able to do (up to) 16 bit PWM

Yes, but at the cost of lowering the PWM Fout to 240 hz, making the ripple problem worse. I think that would force the use of a voltage follower between the r/c network and the 50:1 divider to avoid a major ripple problem.

To the OP: if you're going to go the DAC route and accuracy is an issue, ditch the MCP4725 which is a general purpose part, at best. Accuracy is so bad that they avoid the INL/DNL specs for the top and bottom 100 counts and the INL accuracy overall is pretty poor. Also, with the '4725 you cannot provide a separate voltage reference, you have to power the entire device, output and all from the same source. Not so good.

Have a look at something like the AD5691 series parts, they have built-in references and cost less than $5 in single piece pricing from Mouser/DigiKey.

23

Smajdalf:
2) Charger can measure down to 0.1mV steps - and what? Is it confused if you provide 0.5mV steps or something else goes wrong? If yes how do you know it cannot measure even 0.01mV?

Well, because the charger can at least display the 0.1mV steps, I'd like to provide these. However, the integer math provides only 0.25mV steps. I am not one who wants to provide 'invented' data... but also understand that rounding my 0.05mV steps

Smajdalf:
3) I am not sure if you will be able to output 0.1mV either with Arduino or with the DAC. From my experience digital LOW of Arduino is not 0.000000V but slightly more.

Thanks, I will have to figure out what 0 is then :slight_smile:

avr_fred:
Yes, but at the cost of lowering the PWM Fout to 240 hz, making the ripple problem worse. I think that would force the use of a voltage follower between the r/c network and the 50:1 divider to avoid a major ripple problem.

In relation to ripple, a settle time of 1-2 seconds does not pose a problem, given the V out does change slowly; e.g. from 30 to 100% SOC over 3 hours.

avr_fred:
To the OP: if you're going to go the DAC route and accuracy is an issue, ditch the MCP4725 which is a general purpose part, at best.
...
Have a look at something like the AD5691 series parts, they have built-in references and cost less than $5 in single piece pricing from Mouser/DigiKey.

Thanks for that tip... :slight_smile: In the meantime I had a look at the LTC2631, which also has an internal reference... (have a DigiKey account). I do not like the thought of Vcc fluctuations ruining my DAC out signal.

I will get a proper DAC, and report back how I went...