All measuring devices distort the applied measurement signals.
Some more, others less.
For which measuring tasks you should rely on the measuring instruments with the best analog input stage on the market, read here: LTT24 and LTTsmart.
Signal quality is always important. But how much of this is absolutely necessary in order to be able to make a desired measurement statement?
Even if we all have a certain basic experience, the first gut feeling is usually dramatically wrong.
The rarely considered cause of this is: signal distortion.
The connections are very easy to describe: to be able to meet?
If a signal is applied to a measuring device, it is (unfortunately) always slightly distorted in the input stage of the measuring device. Here, the signal appears as if there were additional higher-frequency interference signals (= harmonics). These are not present in reality, but appear in the measuring device and thus in the measurement results calculated from it.
In many measurement tasks, these "virtual interference signals" are exactly at such frequencies at which the actual measuring signal also shows important information, so that this important information is falsified or, in extreme cases, even completely superimposed.
If a measurement statement is based on these superimposed signal components, then false statements are to be expected.
Summary:
If there is not only one interesting frequency in a signal to be analyzed, then it must be taken into account that the distortions of slower signal components lead to virtual interference that "collides" with higher-frequency signal components.
A spectacular example is blasting tests: a powerful boom generates a shock wave that makes the material of an object to be measured sweat. In some places, the load capacity of the material is soon exceeded and microcracks occur, which quickly expand until the entire object tears.
These microcracks occur in places where many high-frequency shock components positively overlap and add up.
If you want to record this metrological and recalculate the crack positions from the measurement data, you will be surprisingly wrong if you have not previously relied on the correct measuring device with a low-distortion input stage.
The mentioned high frequency components have very small amplitudes --- compared to the enormous amplitude of the large boom.
Now the distortion strikes: the boom generates faster signal disturbances, also called harmonics. A large boom thus brings large distortion-interference amplitudes at higher frequencies, which superimpose the measurement signals actually present there.
Again, whenever there are [SLOW LARGE] and [fast small] signals at the same time, then an extremely low-distortion meter must be used.
The LTTsmart and LTT24 have by far the best analog input stages on the market!
With LTT meters, the distortion harmonics are less than 3 ppm of the applied signal. The first harmonic is even smaller than 1 millionth of the input signal.
At the same time, the devices impress with an incredibly flat noise floor, which is below -140dB over the entire bandwidth from DC to 2 MHz – without any internal interference frequencies that would be noticeable as spikes. This is absolutely unique!
As a rule, one's own measurement operation has little in common with blasting tests – but nevertheless, the distortion also destroys important percentage points of the accuracy statement in all everyday tasks.
An (almost) arbitrary example of a regular standard measurement is the power measurement on battery-powered multiphase electric motors.
Such electric motors are driven by so-called PWM signals, i.e. pulsed signal sequences. The physics that defines the losses occurs in the steep flanks of these pulse sequences. Fast measurement technology is a must have here.
And here, too, the distortion of the large voltage pulses superimposes the much smaller and higher-frequency signal components that define the losses.
If the input stage of the measuring device used distorts these pulse sequences too much, a fast A/D converter no longer helps to prevent the "virtual interference power losses".
The LTTsmart input amplifiers are better than all other systems on the market by a factor of 100. An advantage that brings between 5% and 7% more accuracy in direct comparison measurements.
Mathematically, it can be shown that with a higher sampling rate, the resolution necessarily decreases. Physical laws describe this dependence.
If one could build an absolutely perfect preamplifier, the optimum of the mathematical power calculation of PWM signals would be achieved at a sampling rate of 2 MHz to 4 MHz.
And Labortechnik Tasler GmbH has been developing and using such almost perfect preamplifiers for 25 years. In combination with the high-precision 4 MHz A/D converters, the galvanic isolation and the high synchronicity of the channels, this is the perfect combination.
But not only this fact is market leading.
While other power analyzer compact systems calculate the digitized data internally according to the usual performance parameters and only transmit these slow calculation results to the connected PC, the LTT measuring systems continuously send all measured raw data to the PC using a patented process.
Once there, the data is calculated with highly optimized algorithms with only a few percent processor utilization to all conceivable performance parameters.
At the same time, the raw data remains available in a ring buffer in the PC, so that higher analysis packages can access the exact underlying raw data at any time in the event of conspicuous performance parameters. Also retroactively.
Thus, the LTTsmart systems when used as power analyzer open up the comprehensive analysis of parameter fluctuations of the electric motor test specimens.
In industry, it is important not to operate such data analyses in an additional software island solution. For this reason, Labortechnik Tasler GmbH offers suitable interfaces to many test bench solutions.
These include the test bench software from Gantner Instruments (GI.bench), Stiegele (MLab), MeasX (DasyLab) and National Instruments (LabView).
Of course, there are also interfaces to Python and Matlab, as well as command libraries to customer-specific programs that can be used under both Windows and Linux.
