Data Acquisition Device

Your question is FAR too broad for a person using MatLab.

You should know this, or know how to find it.

First, use a search engine.

Type this "data acquisition device compatible with MatLab".

Data Acquisition Software - Data Acquisition Toolbox - MATLAB

https://engineering.tamu.edu/media/610391/13._-_drs._yu_and_han.pdf

http://www.electrochemsci.org/papers/vol10/100100737.pdf

http://web.mit.edu/pweigele/www/SoBEI/Info_files/Logan%202006%20Environ%20Sci%20Technol.pdf

Redfred has already told you: find signals first, and what magnitude they are. Then you can figure out what voltage range you need for data acquisition.

For example, if Vout = 0.237 V, then you would not use a 16 bit 0-10 volt scale data logger.

Have you considered making your own amplifier boards? Then input the signals to an Arduino, and convert them to digital, etc.

Having ever used only one Arduino (an Uno) I cannot say if this is true of Arduinos generally, but the A/D converters on mine are the noisiest I have ever used, bar none.

On the flip side, they make for great (atmospheric/thermal-noise-based) random number generators.

Pair wires, untwisted? Twisted pair? Shielded cable?? Signal level and frequency?

It is possible also to take more frequent data than is needed, and average the readings, and this can (but not always) cancel out some of the noise.

I agree that there are much better data loggers out there on the market, for not a lot of money, that can take negative signals as well as positive while Arduino can only deal with positive going signal, since negative will toast the board. One can try using diode (reverse bias to positive), to protect from negative signal so I have heard, have not tried that.

Yeah, for bipolar signals you would need to do level-shifting externally (and very likely scaling and maybe some other signal conditioning) first, then remove the shift bias in software, a trivial operation. Signal averaging in software to reduce noise; median filters also work well. I use median filters to process the data from two CdS sensors on a simple one-axis optical tracker built using the Uno and a cheap servomotor:

It's not the sensors that are noisy, it's the A/D! It's so bad on mine that it might as well have 7- or 8-bit channels instead of 10. The first few LSBs are all over the map; marginal improvement when connected to relatively low-impedance signal source but never really goes away.

My app is basically just a toy, but for any truly serious work I would not consider using an Arduino. Plus - and as others have already mentioned - the OP didn't say anything about the nature of the signals he/she's trying to measure. AC? What frequency? Frequency range? Amplitude? Unipolar signals? Bipolar? (gasp) Quadrature? DC levels? And this is just for starters.

OP? What say ye?

As per usual they leave the building just when things get interesting.

Yes, you'd think in five months' time they could come up with more info. Hopefully they followed the advice given above and found their answer.

Pencil and paper?

Del

Excellent concept, and with a recording frequency only limited by the user's dexterity.

Paper and pencil is too unreliable.

Stone tablet and chisel is recommended.

Well, it does make a more permanent record. What in hell did Romans do when they goofed up? Did they just make paving material out of it?

They got a new scribe to replace the recently deceased scribe and used the obverse side of the tablet.

I tried to hand you a GA for that one. ROFLMAO!