By R.W. Stowe, director of applications engineering, Heraeus Noblelight America LLC
Q. Why all the Data? Radiometers as Diagnostic Tools
A. We often are asked to troubleshoot a UV process that has changed in some way, or where radiometer readings are not as expected. Could the cause of a process change be in the chemistry or in the UV exposure? In the measurement conditions? Or in something else?
Radiometry is the tool of communication when it comes to assessing UV exposure. Radiometers can provide information on peak irradiance in W/cm2 or mW/cm2 (some can give the complete irradiance profile) and exposure in J/cm2 or mJ/cm2 in specific wavelength bands. That’s usually not enough. We seem to want all the data. Why?
The information usually desired for process or equipment troubleshooting is as follows:
- Model number and manufacturer of radiometer used
- Type of UV system, irradiator and bulb (i.e., MP mercury or additive type or LED and nominal wavelength)
- Style of reflector and reflector type (elliptical or parabolic, broadband or dichroic)
- Distance from the lamp face to the work surface (to the back of the radiometer or to the face, if measuring LEDs)
- Type of power supply (ballast, transformer, solid state – or model number)
- Peak irradiance in UVC, UVB and UVA bands (or UVA2 for LEDs)
- Exposure in UVC, UVB and UVA bands (or UVA2 for LEDs)
- Speed of conveyance, accurately measured (or precise timing cycle, if static exposure)
This information allows comparison to properly working lamps or to recorded baseline data. Then, we can assess the difference, change or error. The need for baseline data, taken when systems are new or refurbished, can’t be overemphasized.
Very rough rules of thumb include the following:
- If the peak and the exposure are lower, it may be the result of a change (intentional or not) in lamp power or in lamp position.
- If the peak remains and the exposure is lower, the cause may be belt speed or timing error.
- If the peak is lower, but the exposure remains, the cause may be focus or bulb position.
- If the peak reading is erratic, but exposure remains steady, the cause may be “sample-rate” error.
- If the ratio of UVC to UVA has changed, contamination or age may be affecting the reflector or bulb.
- A change in the spectral distribution ratios can indicate a change (intentional or not) in the selection of the bulb or reflector type.
What are some of the sources of errors?
A commonly asked question is, “Why does the peak measurement vary so widely?” Remember: irradiance does not change with speed, and exposure is strictly inversely proportional to speed. So, if the peak irradiance (Ip) appears lower at higher speeds, something must be wrong. There will be normal variations in peak measurements, but this demonstrates the error caused by sampling rate of the radiometer.
“Sample Rate” Error
Example: The PowerPuck® takes samples at 25 samples per second – and, at a speed of 200 fpm, its samples are 1.6 inches apart! At this sample distance, there is a random probability that the radiometer will not be under the focus of the lamp when it takes its measurement. This is the cause of the difference in peak irradiance (Ip) measurements. Sometimes, it will be correct. (For this reason, we recommend that reference measurements be made at 20 fpm.) If the conveyor has only a 50-200 fpm range, then reference measurements should be at 50 fpm. At 50 fpm, the sample distance will be a little less than ½ inch. If a peak error is suspected, take several readings of peak – don’t average them. The highest will be the correct one. This is the best number to be used for any speed.
Fast Sampling Rate
Some radiometers measure at faster rates (some, much faster). When the model of radiometer has been identified, so has the associated sampling rate. Some can sample so fast that they actually track the line frequency pulsing of the UV-emitting plasma in the bulb! This can result in a different kind of error, and usually in an erroneously high peak reading. Different types of power supplies drive the bulbs with different amounts of ripple in the electric current. That’s why knowing the type of radiometer and the type of power supply is important.
Exposure (Energy) and Speed
Rather than send a radiometer under a lamp at lightning speed, it’s easier and more accurate to measure the exposure at a lower speed and to calculate for higher speeds. This saves work, not to mention eliminating the ballistic problems of “catching” a radiometer traveling at high speeds.
Exposure varies strictly and inversely with speed. An unknown exposure at a speed, Vx, can be accurately and quickly calculated from the energy at any other speed, Vo:
Ex = Eo • Vo/Vx
The actual speed of measurement is critical. Always measure the speed of a lab conveyor, for example, with a surface speed tachometer. Generally, don’t trust a panel speed indicator: Measure the actual speed.
Many lamps can produce peak irradiance levels up to 20 watt/cm2. If the radiometer does not have the range to record the peak produced by the lamp, then not only will the peak reading be in error, but the exposure reading will incorporate that error. The instrument dynamic range can be determined from knowing the model.
Various manufacturers of radiometers build different responsivity into them. Most of these instruments are detector-filter types, and the combination of components in each determines the wavelength band to which it responds. Just stating “UVA” or “UVC” is not sufficient. Knowing the manufacturer and the model of the radiometer will allow the upper and lower limits of its wavelength bands to be determined.
Many variables effect the overall performance of a UV-curing system, and understanding the measurement of those variables is critical to troubleshooting its performance. The lamp supplier, integrator, formulator and any others you consult may ask for loads of data to get the system restored or optimized. Bear with us when we seem to ask for more data than you may have.
R.W. Stowe is director of applications engineering at Heraeus Noblelight America LLC. He can be reached at email@example.com.