The following methods can be used to reduce noise in a measurement.
Each input channel is averaged over a programmable period of time to produce 1
measured value. This time is referred to as its integration time. This is extremely
helpful at reducing high frequency noise (>1KHz), and can be used to reduce power
line noise as well by integrating for 1 whole power line cycle (e.g. 0.016 or 0.020 seconds). For details,
ATTACH SENSOR GROUND TO INSTRUNET GROUND AND NOT DUT GROUND
If a high frequency (e.g. 100MHz at 1mVrms) resides between the instruNet ground (GND) and the Device Under Test (DUT) ground, as noted here, then attaching the sensor common wire (e.g. cable ground wire or shield drain wire) to both the instruNet GND and the DUT GND might cause a ground loop and interfer with low level (e.g. required accuracy < ±100uV) measurements. One solution is to only attloadach the sensor common wire to the instruNet GND and not attach that common wire to the device under test ground (or chasis). For example; one might attach a strain gage, RTD, thermistor or thermocouple mechancially and thermally to the DUT, yet not electrically.
ADD EXTNERAL RFI FILTER TO i100 AMPLIFIER INPUT
If working with the i100 (not with the i420/i430/i60x
12) and the voltage shown on the computer changes (e.g. < ±200uV) when you physically move the sensor cable (e.g. lift it over your head), then RFI (radio frequency interferance) is probably inducing a small voltage into your signal. An easy solution, in most cases, is to add a capacitor at the point where the sensor attaches to the instruNet; between Vin- and Vin+ if doing differential measurements; and between Vin and GND if doing Single-Ended measurements. This creates a 1 pole low pass filter where a cut-off frequency in Hz equal to [1 / (6.2 * R * C)], where R is the source impedance in ohms of the sensor (e.g. 350 ohms with a 350 strain gage, 120 ohms with a 120 ohm RTD) and C is the capacitance in Farads of the capacitor. For example, Fc = 28KHz = 1 / (350*0.1e-6), with a 350 ohm sensor and a 0.1uF capacitor. The cutoff frequency (filter passes frequencies below this frequency and stops frequecies above) is typically set to be less than the RFI frequency (which is typically >50KHz) and greater than the frequencies that you want to digitize. This technque is NOT recommended for products that directly connect voltage inputs to internal multiplexors (e.g. i420/i430/i60x)
INPUT VOLTAGE DRIFTS WITH TEMPERATURE OR TIME
If the input drifts with time or tempature (e.g. < ±200uV over > 1 hour), please consider software
calibration every several minutes (e.g. set AUTO CAL RATE to 5),
as described here.
The Low Pass popup menu in the Hardware settings area can select a variety of analog low pass
filter options (e.g. i423 options are off (0), 6Hz, or 4KHz).
One can try setting this to see if it helps.
LOCAL EARTH GROUND
Attaching a wire from the instruNet GND screw terminal to a local earth ground will
reduce the chance of the instruNet hardware radiating at a high frequency w.r.t. earth ground
(e.g. 100MHz at 1mVrms). This is sometimes helpful when measuring low level
signals of < 10mV.
Adding optical optical isolation via the i330 or
will reduce the chance of a ground
loop disturbing a low level measurement. This is often necessary when reading a <10mV signal greater than 10meters from the computer. The i330/i601 isolates the
computer ground from the 1st instruNet device signal ground; otherwise, they are
connected, current might flow, and voltage drops might be induced in the measurement
circuitry or the measurement leads. This current/voltage may have several
components: dc, power line frequency, or high frequency spikes (which could have
been induced by millions of simultaneously switching transistors in a computer from
hell with 1ns rise times).
Each instruNet channel includes a programmable digital filter, with a programmable cut
off frequency. This can be used to reduce noise above the specified cutoff frequency.