Keithley 182 Sensitive Digital Voltmeter
The speed and noise of the Keithley 182 were improved in comparison with the older Keithley 181. The graph on the left shows the noise density derived from 36 rounds of data acquisition with the input shorted, from respectively the Keithley 181 and 182. In both cases GPIB Talk only was used to get continuous data using a while loop in LabVIEW. The integration time was one mains power cycle, and all filtering was disabled.
The upper limits of the spectra are defined by the acquisition interval, which in average is about 287 ms for the Keithley 181, and about 73 ms for the Keithley 182. This is slightly slower than the 4 Hz and 16 Hz rates stated in the specifications, but could at least partly be explained by the fact that I live in a country with 50 Hz mains rather than 60 Hz. Over a 0.01 to 1 Hz span the noise density of the Keithley 181 is 20.0 nV/√Hz, calculated as an RMS average over the FFT bins. For the Keithley 182 this figure is improved to 7.96 nV/√Hz. The spike in the noise density close to 0.7 Hz for the Keithley 181 could be reproduced, but the reason why it appears is not identified. |
The Keithley 182 includes settings for analog and digital filtering. When enabling any of these the acquisition interval increases from about 73 ms to about 250 ms. Also, if you want to avoid the non-linear speed-up feature that overrides the filter when the change of the input signal surpasses a certain threshold, you should use either the "slow" digital filter and/or the analog filter, or no filtering at all. The other two digital filters invoke the speed-up feature, which is not ideal if you intend to use the 182 as a low-level digitizer. The graph on the left shows the "slow" digital filter (setting P3), and the -3dB cutoff is close to 16 mHz.
|