Science Communication: Discussion on measurement errors in oscilloscope testing.

minidso

一、Definition of Measurement Error

Measurement error is difference between a measured value of quantity and its true value. In statistics, an error is not a “mistake”. Variability is an inherent part of things being measured and of the measurement process.
This post intends to discuss the measurement errors from device testing. The errors usually come from either amplitude(voltage) or time. The more you know about where they come from, the better you can control them.

Frequent questions and problems:



1. Large amplitude in triggering
2. Sharp measurement errors of voltage
3. Huge measurement errors of time span
4. Weird signals
5. Separation of low noise
6. Unusual waveforms out of unknown reason
7. FFT transform spectral errors value

二、Where do the amplitude errors come from? (errors between measured value and ideal value)





When the analog signal transforms into the digital signal, the digit capacity of the ADC sampling chip decides the error value because the capacity decides how many equal shares can the signals be cut into.
Digitizer with N bits vertical accuracy transforms analog voltage into N bits digits.
＃Bits resolution
8 255:1
N (2n-1):1
Digital output is in a binary system with symbols.
In binary system with symbols, the code on the screen top is +127, the middle 0, and the bottom -128.

ENOB decides the ultimate quantization error. And the ground noise also contributes to the amplitude error.


The frequency response contributes to amplitude errors：Frequency response curve




Conclusion: The amplitude result is different when the same-amplitude sine-wave in different frequency is tested by the same oscilloscope.

The probe could also give rise to the amplitude errors.


Conclusion:
1. Don’t use the probe when you can use the cable.
2. Use shorter earth wire if possible
3. Long earth wire and signal wire will form a ring circuit. The larger size the circuit covers, the more disturbance of space radiation it gets.

Time direction errors

In terms of sampling clock error, check if the sampling position is ideal.

Sampling chip overlapping sampling errors and interleaving errors

Trigger jitter refers to the deviation between the measured triggering point and the true one. Most triggers in the market are analog. Even if you modify them with software, the trigger jitter is severe. RTO digital trigger jitter is at 500fs. The ground noise of the oscilloscope will turn into uncertainty of jitter measurement.

How to avoid or minimize errors?
The error value is not decided by ADC’s design digit capacity but the ENOB.
Average algorithm:
For steady signals, it is advisable to dwindle errors by removing noise with average algorithm.

Fine cable and probe
Fine cable means good frequency response, high compatibility with the oscilloscope, short earthing conductor and high standard of cmrr of differential probes.

Statistical approach
Statistical approach serves to secure maximum, minimum, average, and variance value of certain confidence interval.

Sound filter

Increasing sampling bits with software is in fact equal to making the measurement more accurate through the sound filter.

--------------------------------------------------------------------