Signal-to-noise ratio vs calibration uncertainty vs repeatability in laser power & energy measurement
Thursday, January 17, 2019
It is common to search for the measurement solution with the lowest uncertainty of measurement (i.e. the highest accuracy, if you will) for laser processes involved in demanding applications such as medical device manufacturing or aerospace parts manufacturing, where traceability is key to a production line.
Calibration uncertainty of your laser power meter or energy meter may not be the only specification you need to check. Repeatability, linearity and noise equivalent power or energy are other specifications that must be checked out as well. Let’s make it simple! Here we help you select the right detector for your laser measurement needs, quickly.
First thing to do when you qualify a power meter for a given laser is to know about the laser wavelength, minimum and maximum expected power levels, beam size and type of profile (Gaussian, flat-topped, with hot spots, etc.) If the laser is pulsed, you also need to know about the minimum and maximum expected energy, the repetition rate and pulse width.
Knowing about the environment where the measurement will be done is also important to make sure the chosen detector provides the best performance in the long term. Temperature, vacuum level if applicable, access to cooling water, etc., are common parameters to take into consideration.
The minimum measurement that can reasonably be achieved should be no less than 30 times the specified noise equivalent power (NEP) of a detector. For example, if you’re trying to measure the power out of a 5 mW laser pointer, the 1 mW NEP of the UP19K-15S-H5 thermal power detector would make your measurement quite noisy. In other words, the inherent electrical noise of the device is not sufficiently low compared to the signal generated by the actual laser. However, the 10 pW NEP of a PH100-Si-HA photodiode would give an excellent signal-to-noise ratio (SNR), way above 30.
Then, if the ambient lights are not varying, a proper background subtraction will complete this section and leave you away from unwanted noise. Background subtraction, or zeroing, is a common feature included in all Gentec-EO display devices & PC interfaces.
The typical +/- 2.5% calibration uncertainty of Gentec-EO UP series power detectors comes from using Gold standards from the National Institute of Standards and Technology (NIST) or from PTB as a reference for calibration. It tells you how close you are from what they define as the Watt when it comes to laser beam measurement detectors.
For relative measurements though, an outdated calibration or wavelength-dependent, typical correction applied to your measurement may be sufficient, as long as you do not need to use your data outside of a specific setup with the highest accuracy.
Have a look at this figure, which shows the sources of uncertainty in the calibration process of a laser power detector (ref: How Calibration Works, white paper).
In strict production environments where it is mandatory to trace each parameter involved in the production of one specific batch of goods, it is required to identify the cause of any non-compliant processes and schedule maintenance accordingly. It is required to tell whether a non-conform laser power/energy measurement either comes from a defective laser or from a damaged or out-of-calibration detector.
Let’s run an example where one needs a reliable, high accuracy energy measurement with a pyroelectric detector. The same principles apply for thermal power detectors.
Let’s say you need to measure the energy of single pulses (e.g.: your repetition rate is less than 1 Hz) at 1070 nm with 15 J, 3 mm beam diameter, 1 ms pulse width. You need the best accuracy for your laser process control before manufacturing parts. Your processing machine needs to comply with the highest traceability standards for your customers.
Since we recommend a signal-to-noise ratio (SNR) of at least 30, and considering the typical noise level of the best-fitting Gentec-EO QE series energy detector for this application being 10 uJ (QE12 products), you will have plenty of SNR at 15 J. Noise won’t be an issue here.
With a good SNR and known calibration uncertainty of +/- 3%, the key specification to look at is repeatability which is less than 0.5 % with our QE series: that will tell you about the variability to expect from one measurement to another in the same conditions of measurement.
Let’s assume your laser and detector are both at thermal equilibrium. At 15 J, 1070 nm, minimum 10 mm beam, 1 ms pulse width, single shots, you will see a difference of less than 0.5 % between the energy measurements with the QE25ELP-S-MB pyroelectric detector + its QED-25 attenuator. For differences larger than 0.5 %, that would likely come from the laser and not the detector, but ultimately, one cannot confirm with absolute certainty the origin of this offset between the expected measurements.
Have a look at this article on how to measure laser power. This article is written for power measurement but the steps described for having your laser and detector thermally stable and ready for measurement are relevant for energy measurement as well.
For detectors based on the principles of calorimetry to measure high laser power (unlike smaller detectors, and which is usually is necessary above 2.5 kW), there are other specifications to take into consideration in order to predict how your measurement will behave. That is linearity with power, linearity versus beam diameter and linearity versus beam position. See our HP100A-4KW-HE specifications for example.
Make sure to not mix up specifications! Please contact us if you are not sure where to look at in detectors specifications for your application.
When reading out laser power or energy, we overcome fluctuations coming from noise by using a detector with a sufficiently low NEP (or NEE) and making sure we have a SNR of at least 30. Calibration uncertainty is reduced to a minimum for commercial products using the best calibration standards in the industry. Finally, repeatability is the specification to watch out to assess drift in power or energy over time for your laser.
Keep in mind the typical, additional +/- 1% accuracy on the scale of readout added by your monitoring interface such as MAESTRO.
It is also very important to use your detector away from its damage thresholds in power, energy, power density and energy density (damage thresholds vary with beam profile, power/energy level and wavelength). That will insure the longest lifetime and best performances for your detector that we recommend to have it checked and recalibrated after 18 months for a new unit then every 12 months. That is not a hard requirement because it depends how you will use your detector but it’s a good rule of thumb to follow if you need absolute measurement and the best traceability for your processes.
Please contact us for any service request for your existing Gentec-EO detectors or any questions for finding out the right measurement solution for a new laser application!