This past Friday, I assembled an instrumentation amplifier using LT1028 op-amps for Johannes's laser current noise measurements. The schematic for the instrumentation amplifier is attached. As Johannes mentioned in a previous post, because of the large voltage drop across the laser diode, a DC coupled instrumentation amplifier cannot have gain of more than ~4, which means low resistances have to be used to stay at the input noise level of the LT1028. Johannes and I thus selected resistances of 100 ohms for the resistors labeled R1_1, R1_2, and Rg in the schematic, and 200 ohms for the resistors labeled R2_1, R2_2, R3_1, and R3_2, such that our gain would not exceed 4.
The initial LISO simulation of the circuit (whose .fil file can be found on gaston at the path ~/izabella/instr_amp/instr_amp_lt1028.fil, and which did not include the 10nF overcompensation capacitors shown in the schematic) yielded the results in the attached plots. From these plots, we learned that the expected noise of the circuit was ~2.6 nV/rtHz, while the gain was ~3.
We didn't get to fully measure the noise of the realized circuit, but we ran some initial tests: setting it up in the lab, we powered the LT1028's using the battery of the SR560, and we saw the output noise of the circuit to be ~15 nV/rtHz. Dividing this by the expected gain of ~3, this resulted in about 4 to 5 nV/rtHz of input referred noise with a grounded input. This is about 2 times our expected noise, which is not what we were hoping to see. Perhaps this descrepancy is due to the fact that the actual circuit I assembled used the LT1028 op-amps, which are low-noise, while the resistors used in the circuit were not low-noise.
A thing to do in future: actually measure the transfer function of the circuit.