I am trying to find plans for a simple millivoltmeter

I am trying to find plans for a simple millivoltmeter that can detect minute changes in voltage fluctuation and record them on the computer. The programming I can see this requiring seems fairly straight forward. Not having done that much in electronics however, I'm not sure what to do on the physical end to take the analog, very low current and amplify it so that it can be detected by the arduino (or can the arduino detect millivolts of power?). I have an arduino mega1280. Any guidance or links to write-ups would be greatly appreciated.

I assume you want to use the arduino to measure the voltage then save the data? It is possible to measure in millivolts if you apply a voltage to aref and use the analog reference function. Using the SD library you can save the data to an SD card. For hardware, you would need to know what the voltage range was. If it exceeds 5V or the voltage on aref, your arduino is dead. Otherwise, just connect a probe to the analog input. There is also the problem of resolution - using 1V on aref will give a resolution of ~1mv. Not a lot...

http://arduino.cc/en/Reference/AnalogReference - analog reference
http://arduino.cc/en/Reference/AnalogRead - analog read
SD - Arduino Reference - the SD library

Onions.

I forgot to mention that the voltage will be essentially AC, aka. fluctuating between -/+. This leads me to conclude that I need to add resistors enough to insure that the voltage will always be positive. Is this true or am I misreading the documentation? I guess I'll just look at some ARRL info on amplifiers but if anyone has some Arduino specific plans for signal aplification, that would be great. Thanks.

Diodes would ensure that you are reading dc, depending if you need to know positive or negative, perhaps two separate analog reads off two half bridge rectifier
if it doesn't matter one read off a full bridge rectifier will work
also depending on the voltage scale a resistor voltage divider would be needed, but will decrease accuracy a bit,

Diodes would ensure that you are reading dc

BUT diodes have a voltage drop over their terminals. The actual reading will be different to the input voltage, so there will be no accuracy to it... And another problem: diodes have to have a certain voltage over them before they will conduct. If you are wanting to measure a few millivolts, the diode will stop any current getting through and there will be no (meaninfgul) reading at all.

Problems:
¬The arduino cannot measure AC
¬A diode will loose any accuracy
¬You will not be able to measure anything below a certain voltage.

Time for some serious thinking!
Onions.

So maybe it is time to back off and have you clearly state the signal properties and what information you want to measure from it.

You state it's a low level AC voltage, what is the min and max range of voltages and range of frequencies it can have? What is the source impedance of what ever is driving the signal.

What measurements do you wish, AC RMS voltage value, peak to peak voltage values, frequency of the signal?

All can be performed with suitable external circuitry but we need all your signal specification and measurement objectives to help come up with a complete solution.

Lefty

What measurements do you wish, AC RMS voltage value, peak to peak voltage values, frequency of the signal?

I am trying to graph the voltage change over time, thus measuring the frequency. I'm using this as the steppingstone to a simple EEG/EMG.

You state it's a low level AC voltage, what is the min and max range of voltages and range of frequencies it can have?

The min is basically as low as possible, however realistically, (given the fact that I do not expect to be able to isolate out all extraneous noise) .1-1mV to 5V is what I am shooting for. The frequencies are 1 Hz to 300-400hz (although the this will obviously be a complex frequency)

What is the source impedance of what ever is driving the signal?

The source is human skin.

Google "Galvanic Skin Response measurement".

You do NOT want to measure AC, you want to effectually drop it to ground.
You use a high gain D.C. op Amp circuit to measure the offset voltage between the non-inverting input and the grounded inverting input (and the grounded patient/subject). Takes lots of shielding and stable op amp.

Also, HEED THE ELECTRICAL HAZARD WARNINGS. It could be real easy to electrocute your patient/subject (which will likely be you in the beginning), if you screw up the grounding. Keep it close to the other input and DO NOT LET THE PATH CROSS THE HEART.