Hi to all!

I try to use diode temperature sensor for precise measurements (+/-0,05 K) (fig.1) and have faced with a problem of electromagnetic noise pick-up which lead to displacement of the characteristic to greater temperatures.

For noise avoid I use capacitor (fig.2) as it was adviced in the article g767.pdf, but sometimes errors are still felt. Wires are stranded and their shields are grounded only on devices.
Url:
http://www.gmt.com.tw/datasheet/g767.pdf

For some reasons the capacity should reduce entrance resistance of the voltmeter on the high frequency, but the experiment results with it much better then without. At the same time, when I use capacity both near the voltmeter and near the current source the results are very bad.

How do you think, what basic requirements should be shown to a current source and voltmeter? Whether should they possess a high-resistance output/input accordingly?

Attached Images

dts.jpg (28.6 KB, 134 views)