#avr Logs

Apr 29 2021

07:18 AM specing_ is now known as specing
10:19 AM Sponge5: Hi, quick question.. I have an analog signal coming from a scale through a divider (10V->5V) and then into ADC. Is it possible that the ADC will convert nonlinearly? In other words: The scale is confirmed as linear, but the values on ADC aren't. Small voltages get even smaller values.
10:20 AM Sponge5: 1.3kg is 1.3V gets converted to 123. 0.2kg is 0.2v gets converted to 17. 0.05kg is 0.05V gets converted to 2...
10:21 AM zgu: yes the ADC can have linearity errors
10:22 AM Sponge5: I see, but they're meant to be +-2 LSB, no?
10:22 AM Sponge5: not sure where I read that, but I just remembered thanks to you
10:22 AM zgu: switchable gain stages might help if you need better accuracy at low voltages
10:23 AM Sponge5: yeah, the scale has switchable gain, does AVR have it too though?
10:23 AM zgu: on some of them you have to read a factory calibration value and load that into the ADC to get the quoted accuracy
10:23 AM zgu: which chip?
10:25 AM Sponge5: mega2560
10:26 AM Sponge5: I'll look into the datasheet myself if you have to look it up as well
10:29 AM exp: iirc it's differential only
10:29 AM zgu: yeah looks like the gain stage is only in differential mode
10:30 AM exp: Sponge5: there's a lot of complexity in doing an ADC
10:30 AM exp: and in keeping online, apparently
10:35 AM exp: anyway checking the specs, single ended conversion at > 200khz is ~3 bits
10:35 AM exp: and 0.05v becoming 2 is within that range
10:36 AM exp: sorry 3 LSBs
03:22 PM JanC is now known as Guest66672
03:22 PM JanC_ is now known as JanC
07:17 PM specing_ is now known as specing