Timeline for Unexpectedly poor DNL, INL and ENOB on an ADC
Current License: CC BY-SA 3.0
18 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 12, 2016 at 23:28 | vote | accept | avg | ||
| Dec 9, 2016 at 10:47 | answer | added | metacollin | timeline score: 13 | |
| Sep 6, 2016 at 17:54 | history | tweeted | twitter.com/StackElectronix/status/773217795014352898 | ||
| Apr 14, 2016 at 23:46 | comment | added | avg | @student No problem - as I've said, it's good to know that something like this may be normal. To be honest, I didn't even think of implications of having ADC on the same die with MCU - it seems like it can be a major factor in degrading the performance then? | |
| Apr 14, 2016 at 22:26 | comment | added | student | @argus We used to do this as experiment with undergrad students: whole point was to show them how bad these ADCs can be when on the same die with MCU. Sorry that I cannot provide more details, it was really long time ago. | |
| Apr 14, 2016 at 22:22 | comment | added | avg | @student Thank you for the sanity check - the results still look very suspicious to me. Have you been able to pinpoint a cause of or even find a solution to your problem? | |
| Apr 14, 2016 at 21:20 | comment | added | student | Few years back we observed something similar on ATmega's from Atmel. I believe it was 10-bit ADC with ENOB going below 7 bits. | |
| Apr 6, 2016 at 20:17 | comment | added | avg | @user105891 I don't think that the datasheet says 6 LSB anywhere - perhaps for one of the smaller packaging options; I have the other one - hence one would expect a smaller LSB. However, this indeed seems like this is just how it is - a poor ADC. Thank you for the comment! | |
| Apr 6, 2016 at 20:15 | comment | added | avg | @BrianDrummond Thank you for the suggestion regardless - I've learned quite a lot about impedance in the process. This is all a process of eliminating possible root causes, after all. | |
| Apr 6, 2016 at 20:03 | comment | added | user16324 | That looks low enough to me, to be an unlikely source of problems. | |
| Apr 6, 2016 at 18:51 | comment | added | avg | @BrianDrummond. The source impedance is less than 100 Ohms - quite a small value, which is a good thing AFAIK. Is it so? | |
| Apr 5, 2016 at 23:32 | comment | added | avg | @BrianDrummond I will make sure to measure that tomorrow morning when I get to my setup and post it here. Thank you for the suggestion! | |
| Apr 5, 2016 at 22:53 | comment | added | user16324 | What source impedance is feeding the ADC? You may need a buffer amplifier. | |
| Apr 5, 2016 at 21:01 | history | edited | avg | CC BY-SA 3.0 |
Updated with some new test conclusions
|
| Apr 5, 2016 at 20:27 | comment | added | avg | @ClaudioAviChami I believe mine is a linear power supply. It's make is Rigol DG2041A. Could you hint me as to where I can start with checking the ADC clock - a quick Google search doesn't look promising :) | |
| Apr 5, 2016 at 20:15 | comment | added | user105891 | These errors are like expected. Datasheet says in worst case 6 LSB. You could try some calibration or better use a separate ADC for precise measurements. | |
| Apr 5, 2016 at 19:42 | comment | added | Claudio Avi Chami | Did you check the P.S.? Is it switched mode? If yes, can you put an LDO between P.S. and ADC? Did you check the quality of the ADC clock? | |
| Apr 5, 2016 at 19:33 | history | asked | avg | CC BY-SA 3.0 |