high resolution AD converters

For a project I am working on there is a need for greater AD accuracy than is provided with stock Arduinos.

I have observed several people attempting to work with higher bit resolution AD converters, but nothing is definitive.

Specifically I am looking at using the maxim11040 initially. It is a 24 bit AD converter this will serve as a reference point for what is the lowest bit resolution feasible AD converter we can use to finish the Project.

The project in question is a multiplexed temperature sensor where we are using differential scanning calorimetry to observe multiple sites within organs in parallel. We need to directly observe the freezing isotherms and traditional DSC are insufficient for non-homogeneous materials

Has anyone any experience with the above AD converter or other similar ones? It seems like it would fit nicely with the arduino as the cooling controller and the cascade multiplexing option is quite nice.

If noone has directly used this device what about experience with other AD converters (>14 bits)? Example SPI interfacing code? Circuit boards and trace? Runtime implications (table lookup for correction?).

Thanks in advance

I have used 18bit with I2C with MCP3421 and it works well. A guy at work has used 24 bit with Arduino and here is the link he sent me http://interface.khm.de/index.php/lab/experiments/connect-a-ltc2400-high-precision-24-bit-analog-to-digital-converter/ I work for an analytical instrument manufacturer and have used a DSC many times. As Richard states be careful with low level signals, you can pick up hum from lights, RF from various devices. Creating a 20-24 bit device for measuring low microvolts is more difficult than it appears. :(

we are using matched 100ohm platimum RDTs in a wheatstone and compensated for non-linearity in a separate circuit. Amplification is done through a high CMMR instrumentation amp. the rtds are pulsed using a precision current device to keep self heating below a couple 0.01mA.

I understand that 24 bits of dynamic range is a lot, however for an initial experiment it is well justified because then we have direct proof of the circuits ability to operate at a specified bit resolution. Shoot for the moon, end up in Chicago... i guess.

and i forgot to mention the whole board is going to be put into a faraday cage and we have considered board placement to prevent rf interference on as many lines as possible.

I am aware the difficult is in the low signal level. We are using high quality instrumentation amps. Noise is typically nv/sqrt(Hz). This gives us at least ppt accuracy when we need better than 10 ppm.

the issue always involves trade-offs, going from an rtd in Wheatstone with differences in resistance of <1 ohm and amplifying that amplifys resistor and current noise. We are keeping current low and using low noise amps to gain maximum accuracy. If this fails we move onto a Kelvin bridge or something more specialised for measuring even smaller differences and dynamically switch when we move between levels using reed relays.

The point behind the post is that if there is a library for AD interfacing or someone has done a boatload of work that my team is unaware of, there is no need to duplicate the effort so a little fact finding is in order.

Normally, normally, you would go with a Lakeshore temperature sensor or controller or equivalent nano-voltmeter that's cut out for this type of job. I've never used a DSC, must be a heck of headache to build. My research advisor used to do that stuff and no more. There's plenty of commerical models.

Is your research advisor out of money or your project manager never done his job in the field? ;)Arduino may have been to the outer space but it's not meant for this. Any good-quality nano-voltmeter has a standardized interface to transmit data. If data transmission is an issue, someone has done infrared transmission through a glass.