Hello Kai,

If the cost of your help is only my supplying the details of my project then I have found a bargain! Here are some that pertain to your comments:

First, the need for 16-bit data is based on my customer's marketing approach, not the end user's actual needs. Ten bits would be enough for that, but having more resolution will allow me to do some range selection in software.

The serial communication will consist of a PC listening to the serial port. I will be taking ten readings per second so I will be able to send individual readings between measurements. The PC will not be talking to me except to abort a test.

I totally agree with your concern for proper ground management and I was planning to use a 4 layer PCB. Every hour spent on grounding at the beginning of a project saves three at the end.

There will be very little external analog electrical circuitry. The output ramp is scaled to 1 to 5 volts and sent to a pneumatic servovalve. the servovalve uses it to regulate a pressure input to a pneumatic to hydraulic pressure amplifier. This output is sent to a pressure transducer only for local servovalve feedback and also to a hydraulic flow amplifier. That output is applied to the specimen and also to the pressure transducer that I read and send to the PC. When I detect that the specimen has burst I end the test. Both the servovalve and the transducers are proprietary units powered by 24 VDC. The transducer that I read has a 0 to 5 volt output.

Back when I was designing instrumentation in the eighties the A/D converters were either Successive Approximation (fast but very noise sensitive) or Integrating Dual Slope (slow but relatively noise tolerant) This will be my first use of a Sigma/Delta converter. How is their noise succptibility?

Thank you very much for your interest in my problem.

Don