Hey everyone, i'm semi new to the whole arduino thing. I'm currently working on a project that requires an arduino use an accelerometer to take readings. However, my readings are severely varied in any position. (Even at complete rest) The only way to get stable readings is if I plug my laptop up to my arduino. The reason this is an issue is because I want to use a 12 volt source in order to power my entire set up. I have tried to power my arduino before and after I drop voltage (using a buck sensor) down to 5 volts to power my other device. Is this a noise issue or a grounding type issue? I can get stable readings if I take rolling averages of my values however this makes for slow response times which are not acceptable. Could voltage to arduino readings that very from 8.56 to 8.59 be the issue?
Should I try to power the accelerometer separately from the board using a regulator? (Thus the accelerometer only interfaces with the board through an analog pin?) 
Try using a linear regulator rather than a buck convertor.
For the purposes of my setup I am using the buck converter because I have a load that will draw up to an amp sometimes, Thus the regulator will be getting very hot ,even with heatsink. (Ask me how I know) However, Do you think it would be a good idea to use a linear regulator in line with the power source BEFORE it gets to the buck? So for instance I drop the voltage from 14-12 volts (Think car battery) to say 7-9 volts for the arduino (If the accelerometer is still using it as a power source) then further down the original power source wire I use the buck converter for the big load that needs 5 volts?
Yes, without an oscilloscope its difficult to tell but i suspect the output of the buck is dirty.
Im surprised it gets hot though umless you are overloading it or possibly faulty.
Its a good idea to keep the electronic supply separate from the load supply, if load is inductive its pretty much essential.
well the buck doesn't get very hot right now. I haven't even attached a heatsink to it yet. The linear regular was getting quite hot though. I will try to pull the supply the arduino before the buck to see if it gets better. If not I will then try to supply the arduino and the accelerometer separately but both before the buck.
im not sure i understand all that correctly.
My suggestion is power arduino from the battery and on board regulator, and the accelerometer, via a separate linear regulator if necessary.
Connect the buck directly to the battery and load.
Connecting the buck to a linear regulated supply i would expect to give heat problems given the load.
Let the buck take care of its own regulation, it will be much more efficient that way anyway.
Edit
A la battery can have a terminal voltage of 12.9 v that can power an arduino directly provided there is not significant current drain from a shield in which case the on board regulator can struggle . Have you measured current draw ?
dragono:
. Could voltage to arduino readings that very from 8.56 to 8.59 be the issue?
Not if you a reading that at the arduino power input plug
I have another suggestion.
Power connection problems are very common here, this sounds like one of them.
The electronics and project guidance forums are much more visited than this one.
Draw a schematic and post a picture in a new thread
You will likely get much more attention there.
Or ask the mods to move this one.
Well I found my issue. Its the Load. (LEDs) everytime I plug the power in the live sensor readings fly all over the place. Atleast I know for certain where my issue lies now.