Glossary of terms
We have an active glossary feature within our website which means all technical terms relating to signal conditioning, filtering and signal processing are underlined, if you hover you mouse over the word it will display a description about the term, we hope this is useful. However we have also listed a range of terms below with details about their meanings, of course if you would like further details and technical support please let us know
A circuit for removing the constant DC voltage from an input signal so that just the AC portion can be processed.
+10V – +2V sine signal, with a +6V DC offset. Making the dynamic signal all positive.
DC Coupled Input
Same as input.
AC Coupled Input
Input AC coupled, the offset (DC component) is removed, the signal is now centered about 0V.
The generation of unwanted additional frequencies in a wanted signal, caused by sampling the signal at too low a frequency.
To judge how well a particular low pass filer response will act as an anti aliasing filter, you need to know its stopband attenuation and its stop band ratio.
To calculate the number of bits of protection that the filter will provide against aliasing, divide the stopband attenuation of the the filter (in dB) by six. This number does not have to be equal to the number of bits of resolution of your sampling system but its a good starting point.
To calculate the lowest sampling ratio which you can safely use before aliases occur in your data, just add one to the stopband ratio of the filter. This criterion does not provide unconditional protection against all aliasing but is sufficient in most cases.
For instance, a filter giving 72dB rejection at two times the cutoff will provide 12 bits of protection and can be used with sample rates of three times cutoff (or higher).
The ratio between the output and input signal levels of a signal handling device. Usually expressed in dB.
dB = 20 log Ratio of signals
A filter in which a frequency region with a low attenuation lies between frequency regions in which attenuation is high. Simple bandpass filters can be made by connecting a highpass filter and low pass filter in series.
Sometimes called a band-elimination filters. A filter in which a frequency region with a high attenuation lies between two frequency regions in which attenuation is low. Simple bandstop filters can be made by connecting a highpass filter and a lowpass filter in parallel in the appropriate way.
A widely known class of filter responses with excellent overshoot properties and linear phase
Note – in many applications where a linear phase filter is required the Kemo response 41 may be a better choice.
A widely known class of general purpose filter responses.
Note – in many applications Kemo responses 41 General purpose and 01 anti-aliasing could be better choices.
Cutoff frequency span
The ratio betweeen the lowest and highest cutoff frequencies that a particular model of a filter product can be set to.
A log ratio between signals. Typically the ratio between input and outputs in an electronic system.
The deciBell is 1/10th of a Bell, and used in the early Bell telephone systems as a measurement of power.
dB = 20 log (ratio of signals)
For a gain of 10:1 in a system
dB = 20 log(10/1), = 20 log 10
= 20 x 1 [log of 10 is 1]
= 20 dB
Therfore working the other way
40dB = 102
Some common values
-80dB = 1080/20 = 104 = 10 000:1
-6db = 10 -6/20 = 0.50118 ≈ 0.5
+6dB = 10 6/20 = 1.99526 ≈ 2.0
+1dB = 10 1/20 ≈ 1.12
+0.1dB ≈ 1.01 OR 1%
The ratio of the frequencies (or frequency differences) which specify the extremes of the passband and stopband of a filter.
The ratio between an input signal and all the unwanted extra signal present at the same time.
In simple terms the ratio between the smallest signal, or noise floor, and the maximum signal in a system.
A system with 1mV noise floor and 10V maximum signal, gives 10 000:1 or 80dB (104 becomes 20×4=80dB).
Measured in dB per octave. The slope of a straight line connecting the two points which specify the extremes of the passband and the stopband and stopband of a lowpass or highpass filter. The higher the effective slope, the sharper is the filter.
A type of mathematical function which lends its name to a class of filters (sometimes called Cauer filters) which provide maximum effective slope for a given amount of ripple.
A type of filter with quite good overshoot performance but more stopband attenuation than a Bessel filter.
ICP supply or IEPE Supply
A source of operating power for a popular type of signal transducer.
IEPE stands for Integrated Electronics Piezo-Electric.
An IEPE is a constant current source, typically 4mA, up to 24V. The transducer effectivly acts as a variable resistor in the circuit, generating a voltage, typically a few 100 mV.
IEPE transducers are simply and easy to use. An IEPE source eliminates the requirement for a dedicated charge amplifier. It reduces the need for special cables, it allows longer cables. It transmits the signal at a higher level and lower impedence, reducing cable and environmental noise in a measurement system.
NOTE ‘ICP®’ is a regsitered trademark of PCB Corporation.
Low noise cable
Low noise cable is a particular reference to cables that must be used with charge output sensors such as piezoelectric accelerometers and which contain a special semi-conductive layer to reduce triboelectric noise.
The amount by which the output of a filter exceeds the correct final value when fed with a sudden step change in the input.
The order of a filter is equal to the number of significant poles in the mathematical function describing its frequency response.
What you send to Kemo when you want some filters.
A filter feature which allows the amplitude of a filter response of a filter to be altered to reduce or eliminate overshoot.
The ratio between the gain of a filter in its stopband to the gain in its passband. Usually expressed in dB.
The variation of the filter gain across the passband. Usually expressed in dB.
A measurement of the input signal at an instant in time.
Noise in signal wires caused by the generation of static electrical charge from the layers within the cable rubbing together. This noise is added to the charge within the signal wire from the charge sensor, causing an error in the data.
Also called total error. The amount by which a filtered signal differs from a pure version of that signal simply delayed in time.