I would like to use the LTC2485 for directly measuring a high side shunt voltage for measuring a highly dynamic current.
The shunt resistance will be 30 ohms and the voltage after the shunt will be regulated (to compensate for the shunt variable drop).
So the VIN+ and VIN- signals will not correspond to a fully differential signal, because VIN- will be kept at a relative constant voltage whereas VIN+ can go up to 1.6V from VIN-. Therefore the middle voltage will be changing, possibly at a high frequency.
So my questions are:
The LTC2485 has a very slow conversion rate. To measure the current of a rapidly changing signal, a SAR converter such as the LTC2512-24 might be more appropriate. The LTC2512-24 has a wide common mode range, so it is not necessary to worry about the relation of VIN+ to VIN- as long as both inputs stay between ground - 0.3V and Vref + 0.3V. The recommended input filter is 10ohms and 6800pF. This allows the inputs time to settle after each sampling glitch. If not run at the maximum sampling rate of 2Msps, the filter can be increased.