I've had a similar problem with the Mega and Uno. I found that I can use a nylon screw with part of the head cut off with a knife. Certainly not an ideal solution. You could file or cut metal screws but it's a bit more work. If you're using a metal standoff you also need to watch out for the contacts on the bottom of the board.
Designer of these boards is clearly not professional in electronics.
It's not just the useless mechanical hole on the UNO (reset of the Atmega 16U2 and from the SDA and SCL pins no room for the screw head).
There is also the absurdity of the pinout of the UNO board with Aref that is not as near as possible from A0-A5 as it is in Atmel pinout.
On the microcontroller pinout Atmel has placed side by side Aref, GND, A0-A5 and if Atmel did so it is not by hazard.
Atmel also said in the datasheet : an inductor is mandatory between Vcc and AVcc.
Look there on a scheme Arduino if there is one.
Because all of these errors, some seem to find normal to be forced to make an averaging with the analog measurements.
No it is not normal : Aref track is noisy INSIDE the board.
For an electronician Nano layout (Gravitech) is better than Arduino layout.
Metric : Europe and Asia (several billion inhabitants) use International System, sorry for North america.
The inductor between Vcc and AVcc is recommended if using the ADC, not required. I don't think I've ever seen a '328p design that actually included it though, and we also don't see herds of people complaining that the ADC is too noisy. Might matter more with a noisy power supply, or in situations with more interferance, or for people who really need every bit of adc accuracy they can get (but most Arduino users aren't doing anything where that's necessary) - but of all the design flaws on the Arduino board, that one just doesn't rate.
What about the pin header not being aligned correctly o nthe shield pinout so you can't use generic perfboard to make a shield at home? Supposedly this was a mistake on the first arduino board, that got carried through because people had to make shields that matched it.
Or the fact that they went with a 16u2, despite that they never gave people the tools (like, oh, maybe an arduino core so we could write a sketch for it?) to use it's advanced functionality - leaving us with a serial adapter on the official boards that's easier to brick than the ones on the clones with extra potential functionality which nobody uses.
Or the fact that the capacitors used for the regulator are inconsistent with the parameters in the datasheet?
Or the fact that they used such a lousy regulator - if they'd gotten one with under a volt of dropout, we could have used 4 AA batteries, instead of needing 7+V, which means 5+ AA's, but you can't get 5-cell holders, so that's likely going to be 6S, so 9v, meaning more heat to dump from the reg, while inspiring people to use 9v smoke detector batteries, which do not have enough capacity or current handling capacity for peoples' projects.
Or the fact that they used a resonator instead of a crystal?
I don't think I've ever seen a '328p design that actually included it though
First : UNO from arduino.org has an this inductor .
Second : try like this with a 100 nF 0805 capacitor beetween Aref et Gnd --> micro pin not board pins :
Other point : for what reason 5V regulator was with a DPACK package on DUEmilanove and UNO R1 and since UNO R2 in SOT223 package which can not dissipate as much heat?
