(Part 1)
1- INTRODUCTION
Nowadays viewing conditions needed for Penetrant Testing (PT) and Magnetic Testing
(MT), i.e. UV-A irradiance and visible light illuminance, are well specified and
checked.
This has been somewhat different some decades ago and a flashback may help us to
better understand the current requirements--and the new questions which arise! (Refer
to Chapter 13.4)
2- BACK TO SOME BASIC WORDING
We think it useful to remind readers of some definitions:
VERIFICATION: A verification comprises:
- A check on unit's arrival for discrepancies between a standard and the unit. The
pertaining report is not supplied to the customer unless required.
- An adjustment (setting) if needed to have readings within the tolerance limits
set up by the unit's manufacturer, or required by standards, by the user or by the
verification organism (body)
CALIBRATION (in French: étalonnage): This is a very confusing word. A calibration
comprises only a report of discrepancies between a standard (a luxmeter, for instance)
and the unit under test. No adjustment, no indication by the calibration service
whether the unit meets the user's requirements/needs. It's up the user to:
- If needed, to draw a calibration curve.
- To determine that the calibrated unit meets his needs/requirements;
But a lot of our American and UK friends--and above all, auditors!-- as well as a
lot of Aerospace specifications ask for a calibration...but require a tolerance limits,
require an "Accepted"/"Not accepted" or a "Complies"/"Does not comply" judgement
by the calibration body, after a setting if needed. This "calibration" then in fact
is a VERIFICATION. This word exists in English with EXACTLY the same meaning as in
French....but this word is generally not accepted by auditors. Calibration: yes.
Verification: nonsense!
ACCURACY: In standardisation vocabulary, accuracy is only a QUALITATIVE assessment.
So what about documents/specifications/auditors requiring a "5% accuracy" for example?
Nonsense!
3- UV-A RADIATION AND VISIBLE LIGHT MEASUREMENT
In the '80s it became a topic: Derek STURGER, General Electric, made some conferences
and published reports on this point.
But as incredible as it seems now his remarks went almost unnoticed by the NDT community,
until when MacDonnell Douglas at the end of the '80 went upset when learning a crack
indication on a landing-gear had not been reported by a PT inspector while the supervisor
saw it.
The investigation that followed made it clear that using the inspector's UV-A source
made it impossible to see the indication. The culprit was rapidly identified: the
visible light (erroneously referred to as "white light") coming from the UV-A source.
In fact the visible photons emitted by the UV-A source are mainly blue and violet;
that's why "white light" is inappropriate.
These blue and violet photons went through the Woods filter of the source. Why?
It may be useful first to detail a reaction known in English as "Blue Haze".
This blue haze is due to the UV-A photons which, if not blocked by a filter on the
UV-A source or by UV-A-blocking goggles, enter the eyes. Lens is transparent to the
UV; these photons arrive in the vitreous humour, a kind of "gel" which cools down
and lubricate the eye while bringing in nutritional molecules and taking out residues
due to the living cells of the eye. A lot of these molecules may fluoresce, and do
fluoresce then, leading to visible photons coming from the INSIDE of the eye! This
makes a blurred vision and it is very difficult, tiring to inspect parts for tiny
indications. This is one of the main reasons for our urging PT and MT operators to
wear UV-A-blocking goggles (in addition to the fact that UV-A radiation speeds up
the lens' opacification known as cataract). To improve contrast between indications
and all the visible light in the inspection booth, it is still better if these goggles
also block violet and blue light. All the UV-A sources except LEDs emit a lot of
visible photons, and the violet and blue ones are especially counterproductive while
inspecting parts in the inspection booth.
UV-A sources widely used in these times were vapour-mercury based and similar to
the underneath picture. The top was concave. The outer envelope was made of a thin
Woods filter with a low visible light rejection capability.
The concave hollow led to a low UV-A irradiance on the surface under inspection just
right in the bulb's axis.
Using this kind of UV-A source is forbidden since the mid '90s for PT and MT applications.
4- WHICH MEASUREMENTS ARE REQUIRED?
Depending on applicable documents there are up to four different requirements for
viewing conditions:
--Illuminance (visible light) for PT and MT carried out with visible products.
--UV-A irradiance.
--Illuminance due to the UV-A sources.
--Illuminance due to the ambient visible light in inspection booths
Specifications then available were quite confusing, incomplete and even untrue.
5- SOME FUNDAMENTAL KNOWLEDGE ABOUT LIGHT, COLOUR, UV-A
One needs to have some knowledge about the electromagnetic spectrum. Measuring visible
light or white light or UV-A radiation is not that easy, especially when low-cost
equipment is used. Human beings, except the few ones which have some non-functioning
cells in the eyes, are able to see a large range of colours. Nevertheless some differences
may occur: one person sees a surface as green when another one sees it as green,
yes, but with a bit of yellow. Purple may be frank violet for someone else. And so
on.
Keep in mind that for PT and MT, maybe more than in the day-to-day life, several
parameters will interact to help get an easy-to-see indication --crack detected--
or an easy-to-miss indication--crack not detected.
Some of these parameters:
- Colour of the "information", be it a road-sign, a PT indication, a cloth, our wife's/husband's
hair.
- The background colour; the background is the surface against which the "information"
must be seen.
- Contrast ratio.
- Viewing conditions.
The colour may come from a "visible "dye or from a fluorescent dye, which may be
almost invisible in plain daylight. The colour may be defined by a wavelength: green
is equivalent to 550 nm, red to 680/700 nm, blue to 405/415 nm. When all the wavelengths
are in the same ratio as the eye's sensitivity curve (our eyes are far more sensitive
to 550 nm than to 680 or 405 nm) we see the colour as "white". White light ...is
white, while "visible light" may be blue, orange yellow...or even white. This tiny
vocabulary adjustment is in fact very important: UV-A sources emit visible light,
mainly blue and violet (when the Woods filter is efficient); so in inspection booths
ambient light is probably more blue/violet than white. Nevertheless luxmeters are
calibrated to measure white light....and still quite a lot of specifications require
inspectors to measure "white light" in the inspection area. Depending on the luxmeters,
figures may be very different when measuring visible light inspection booths! (Refer
to Chapters 9 and 10)
The solar light is yellow; a large part of flora is green. It may not be coincidental
that the human species eyes show their maximum sensitivity in the yellow and green
parts of the electromagnetic spectrum!
Light sources used in offices, homes, workshops, shops, etc are very different and
their emission spectrum is also very different. As examples, the omnipresent cheap
tungsten bulbs give a white light but with orange as a predominant colour. Their
colour is seen as "warm" even if their colour temperature is "cold"! On the other
hand luminescent tubes (sometimes referred to as fluorescent or neon tubes) have
blue as predominant colour, giving a "cold light"...with a high colour temperature!
Colour temperature is linked to the PLANCK black body, and we won't give details
here. Go to Wikipedia to get a thorough explanation.
One begins to understand that sources which give our eyes very close "colour feeling"
have in fact very different emission spectrums. As the colour of a surface as we
see it is highly linked to the wavelengths sent back or absorbed by the surface,
one may easily imagine that the SAME OBJECT will be seen VERY DIFFERENTLY when lighted
with DIFFERENT LIGHT SOURCES.
And now back to our penetrant materials and to the magnetic particles widely used
in NDT. Maybe some of you have already met a red dye penetrant which is no longer
red depending whether it is seen in daylight or under some artificial source!
So we have seen that different sources have different emission spectrums.
Let us have a look now to the surfaces, objects, penetrant indications, etc.
Surfaces which equally reemit all the wavelengths they are submitted to will be seen
as white if the light is white, as green if the light is green, etc. Some surfaces
absorb equally all the wavelengths they receive: they appear as black or as neutral
grey.
In fact many surfaces absorb or reflect some wavelengths. If a "balanced" white light
is used (i.e. if all the visible wavelengths are there), if a surface absorbs all
the wavelengths except the red ones, then this surface will be seen as red: only
the red wavelengths are sent back. But what if the light source does not have all
the wavelengths in its spectrum?
Well, red penetrant indications may become almost invisible!! No chance to find them
if a wrong lighting is used!
Please do not hesitate to contact us if you are interested by learning more on that
topic.
Now, what about the contrast ratio? When the indication looked after shows no contrast
against the background, it is useless to increase illuminance. Imagine a black cat,
lighted with a "balanced" white light, in a room with black walls, ceiling and floor.
Or a green snake in this black environment with a green light. In this latter example
all the surfaces would become green; the snake would reemit a green colour--invisible
against any of the surfaces.
Here is a real example: a large aerospace parts manufacturer. A new Fluorescent Penetrant
Testing installation, far larger, more up-to-date than the previous one, in a roomy
space. Really fine! Very soon operators complain about eye-tiredness; they almost
cannot see the tiny fluorescent indications they look for. «*It was better, easier,
in the previous room!" We check every parameter of the products, and finally go the
plant. Quite immediately, when the UV-A sources are switched "ON" and the white light
is switched "OFF" we see the cause for the problem---literally! In fact walls and
ceiling are covered with a yellow paint....which intensely fluoresces also! That
means the poor guys had to find out few yellow photons coming from the penetrant
indications (fluorescent penetrants were yellow then) when "surrounded" by far more
yellow photons coming from walls and ceiling!
The contrast concept is very important. When working with visible products the best
contrast is black against white or white against black. Black particles are commonly
used in MT as they are the most sensitive (they have very few non-magnetic additives)
especially when a white contrast paint may be applied as a contrasting background.
In PT, though manufacturers know how to manufacture black penetrants, red or orange
or purple colour is specified. Why? In fact black particles, surfaces, parts are
quite common in foundries, ironworks, mechanics, etc. and it is assumed that red
was chosen, many decades ago, because it is an unusual colour in industrial environments.
This is the colour of blood, it is the colour used to draw attention on dangers or
bans (refer to the road-signs, for instance).
Blue penetrants are available on the market; they give a very high contrast against
white developers...but they do not meet --yet-- specifications requirements.... because
they are blue!
Once again have in mind the quality of the "white light" used for inspection. LEDs
give now a VERY CONVENIENT white light and are probably the best choice from now.
Well, we have considered several parameters which, among many others, have some importance
for a reliable inspection. Have we forgotten something?...Hum, yes!!!
What about the sensor able to detect the indication and the signal processor able
to use the information from the sensor and able to decide: accept, reject, repair?
In PT and MT, despite long and costly R and D efforts, nothing better than the old-fashioned,
well-trained system comprising two human eyes and a human brain is functional. This
should be a topic for a future DPC Newsletter or for a technical paper.
(Part 2)
6- NDT SPECIFIC SITUATION
Very simple questions: what is to be measured? How? In which conditions? What about
reliability, reproducibility? And equipment cost? Calibration/verification feasibility
and cost?
Quite easy to ask the questions. Answering them needed decades--and, to be honest,
we are still surprised when reading some requirements, or how the meters are used
or how they are "calibrated/verified".
7- VISIBLE LIGHT MEASUREMENT WHEN USING COLOUR-CONTRAST PRODUCTS
Standards and specifications state minimum illuminance figures of 500 or 1,000 lux
MEASURED ON THE SURFACE UNDER INSPECTION. A 1,000 lux figure is very high--and it
is a minimum! -- and may glare inspector's eyes when reflected on machined surfaces.
We consider this figure as counterproductive for reliability as the operator's eyes
may get tired quite quickly.
A figure between 350 and 600 lux is generally far enough.
Detecting red dyes used in PT needs an adequate lighting. You may get interesting
information through the conference given by Alain MORETTI during the May 2008 Technical
Conferences held by COFREND, the French NDT Association.
Measuring white light requires a luxmeter whose response curve shall be close to
the sensitivity curve of the standard human eye as defined in the CEI.60050-845 document.
The sensor shall be plane and have a cosine sensitivity. Depending on applicable
standards/specifications/documents a verification shall be carry out from every 6
months up to every 24 months. As it is standard with every measuring equipment zero
shall be checked; a figure lower than the lowest figure to be measured is to be checked
as well as one upper than the maximum anticipated figure ( that does not mean the
entire range capability of the meter). A figure comprised in the currently used range
shall also be checked, or better three points in this range.
As an example, a luxmeter with a 1/19,999 lux range, used in the 500/2,000 lux range
may be tested for the following figures:
zero (as for any meter!)
- 400 lux
- 2,500 lux
- and 500, 1,000 and 2,000 lux
Doing so the equipment linearity IN THE EFFECTIVELY USED RANGE is checked.
Some auditors ask for a check up to the maximum displayed figure (19,999 lux
here). This is absurd: many of the calibration/verification installations cannot
reliably do that. Those which may charge a hefty price, a useless expense for something
that will NEVER be used!
Many "luxmeters" are available. Sometimes these "luxmeters" have only a label "luxmeter":
there are supplied without a response curve, or their response curve is ...exotic,
the sensor is deep in a "hole" without any diffusing window--hence, it is impossible
to get a cosine response.
MEASURING UNITS
Illuminance unit is the lux, the consistent SI (Système International) unit. The
"foot-candle" that our American friends like so much is an out-of-the-system unit
and shall not be used!
8- UV-A IRRADIANCE MEASUREMENT
During the '50s a measuring equipment was put on the market by an American company,
UV PRODUCTS. This needle-equipped unit, the J221 (see a picture underneath) was THE
reference meter. In 1967 an American document gave several pieces of information
pointing out some failures in correctly measuring UV-A. Though this meter is technically
obsolete since.. ca 1980, it is surprising to see it in some suppliers catalogues
in 2008!
Let us have a look to the main technical obsolete features.
- A selenium sensor: its surface was very large so as to get a usable electrical
signal.
- From the very first unit there were two measurement ranges: from 0 to 1,200 μW/cm²
and from 0 to 6,000 μW/cm². For some time there was only one trimmable potentiometer
("trim pot") to set BOTH ranges! Then two trim pots were used, one for each range,
nevertheless needing what is called "adjustment by iteration". A mechanical zero
adjustment device was on the front side of the unit. (Such a mechanical zero adjustment
is necessary when using a needle-equipped meter. When our young engineers/readers
read this they cannot even imagine the calibration process: a lot of them, even if
involved in measurement/metrology, have never seen such analogical equipments!).
Needless to say that setting both ranges required some time and patience!
- This unit was very sensitive to infrared radiation ( IR): a 100W or 125W UV-A bulb
uses ca 80% of its electrical consumption to send IR while only 7% are used to emit
UV-A in the 365 nm range!
- This unit comprised a removable filter. UV-A irradiance was considered as the difference
between the reading WITH the filter and the reading WITHOUT the filter.
In the '70s, on the same day we made two demos in the same French major aerospace
supplier:
- Without giving any information to the operators, supposed to know the operating
instructions (the PT/MT installations man-in-charge was very confident), we asked
them to check UV-A irradiance and to give us the figure. They did only one measurement--without
the filter. "So, why this removable hood?" "It's used to gard the sensor." Guess
the man-in-charge's face?
- The second test took place some few minutes later, in the office of this same man-in-charge—an
unlucky day for him! The purpose: to show the sensor's sensitivity to IR without
the filter. The sensor was put at 20 cm (ca 8") from a 75W incandescent bulb: 2500
µW/cm² were displayed. Then we compared all these results with those from a meter
with an interferential filter, as described underneath. Reading was 50 µW/cm². A
very interesting afternoon indeed!
- Conversion factor: depending on the UV-A source (luminescent tubes, mercury-vapour
bulbs, etc) due to their vastly different emission spectrums and to the sensor's
very large response curve in the UV range, a conversion factor had to be used.
- In short this unit was far from meeting users' requirements: to have an easy-to-use,
reliable, not too fragile equipment displaying true figures!
Beginning '80s the English company Levy West supplied a so-called luxmeter (sketch
and principle shown underneath) meeting the now obsolete BS 4489:1984 "Method for
measurement of UV-A radiation (black light) used in non-destructive testing" standard.
This meter, though graduated in lux, was not a luxmeter and was supposed to indirectly
assess the "black light level", as UV-A irradiance was then called. A fluorescent
panel lighted by the UV-A source emitted a green light then measured by the meter.
The panel comprised a dispersal of fluorescent mineral particles (probably: selenium
sulphide, cadmium selenide or a mix of cadmium and zinc selenides) in a polystyrene
binder inserted between two transparent polyester sheets.
A kind of correlation was supposed between the UV-A irradiance on the panel surface
and the visible radiation sent back by the panel. As an example a 1750 µW/cm² UV-A
irradiance gave a 250 lux reading.
Further these panels aged quite quickly.
This meter was also described in the ISO 3059 1974 issue (obsolete), and was described
in the French NF A 09-520 April 1989 issue Annex H (also obsolete) as able to check
the fluorescence intensity of fluorescent penetrants.
In this same '80s decade Spectronics Corporation put its DM 365X and DSE 100X, respectively
a radiometer and a combined radiometer/luxmeter, on the market.
A digital radiometer is used to measure UV-A irradiance in a small bandwidth from
340 up to 375 nm. The 1 cm² sensor is flat. The meter displays readings from0 up
to 19,990 µW/cm². Any other unit (such as the photometric lux, phots, foot-candles
or any other related to visible light) cannot be used to measure UV-A irradiance.
The W/m² (or its subunit the µW/cm²) is a RADIOMETRIC unit: no conversion factor
exists between a radiometric unit and a photometric one.
Depending on the applicable specifications verification, calibration shall be done
every 6 months up to every two years. Flanking the used range during verification/calibration
is similar to the method used for luxmeters.
The DM 365X (pictured underneath) as guessed from its name is a radiometer aimed
to measure a small bandwidth centered on the 365 nm UV-A radiation. Its 1 cm² sensor
does not respond to IR and red radiations.
This meter was right in line with the NF A 09-599 October 1988 issue standard titled
‘’Non Destructive Testing. Surface Control methods (penetrant testing, magnetic particle
testing) Characterization of ultraviolet sources.’’
No requirement was written then about visible light (illuminance) measurement in
the UV-A booths.
Again in the beginning of the '80s decade this same Spectronics Corporation company
put on the market the DSE 100X (pictured underneath), a combined radiometer/luxmeter
using the foot-candle as illuminance unit and the µW/cm² as irradiance unit. An-L
version came later using the lux as illuminance unit.
When the aerospace industry required measurement of both the UV-A irradiance and
the illuminance in inspection booths, these units gave reliable results.
In the very beginning of the 2000 years Levy West, now named APPLIED SCINTILLATION,
also put on the market a combined radiometer/luxmeter named LEVY HILL Mk VI (pictured
underneath) very similar to the DSE 100X and -X/L.
More recently the French-designed and -manufactured POLLUX (refer to the underneath
picture) displayed innovative features:
- Small and light
- Nice ergonomics
- Only one non-interchangeable sensor for both UV-A and visible light measurement
directly connected to the display unit
- Simultaneous figures for UV-A irradiance and visible light illuminance
(part 3)
9- UV-A SOURCES ILLUMINANCE MEASUREMENT (VISIBLE LIGHT)
We can date requirements to measure visible light in inspection booths to the end
of the '80s decade. The most common UV-A sources used for NDT then were luminescent
tubes, mercury-vapour bulbs and, more recently, sources based on metallic halides
or on ionised Xenon. These sources have a wide emission spectrum ranging from the
extreme short UV up to the microwaves area. For different reasons a lot of these
radiations are filtered for NDT applications and only the 340 to 415 nm is used.
UV-A sources generally give rise to visible blue light.
In these "ancient" times it was sometimes thought that a better filtering would lead
to a 340/380 nm bandwidth in the inspection booths: no visible light could be seen
in the darkness of the booths. Price of such filters would be so high that this way
was dropped.
10- BEGINNING '90S SITUATION
Well....a bit fuzzy! A first document by GENERAL ELECTRIC described two methods.
The first one dealt only with the 125W bulbs.
First switch on the bulb, wait for stabilisation. Then while wearing UV-A blocking
goggles compare the pattern given by the bulb to a series of photographs classified
from "acceptable" to "not- acceptable".
Not really the best idea to "accept" UV-A bulbs!
The second method came in addition to this test: a check was to be made with a luxmeter
whose sensor is put at 15" - 38 cm- from the source or the filter. Three luxmeters
were approved, but the figures they gave for the same bulb were very different! Bulb
stabilisation was questioned but it was easy to understand why these differences.
This point is to be explained later in this paper.
- ARDROX DLM 1000 (Photo, below): 55 foot-candles, i.e. 591.8 lux.
- SPECTRONICS DSE 100 X: 15 foot-candles, i.e. 160.5 lux.
- MINOLTA T-1: (photo below) 2 foot-candles, i.e. 21.52 lux.
ARDROX DLM 1000
MINOLTA T-1
Models T-1 and T-1M1: Measurement range: 0.01- 99,990 lux
Model T-1H: Measurement range: 0.1 -999,900 lux.
This led to documents forbidding using built-in filter 125W bulbs, thinking that
this problem would not occur with the unfiltered 100W bulbs needing a separate Woods
filter.
PRATT AND WHITNEY did not require the visual comparison but came also to the visible
light measurement on the same erroneous way as GE.
Then ROLLS-ROYCE and SNECMA wrote their own specifications. They decided first to
correlate the UV-A irradiance and the visible light illuminance: this seemed logical
in these times.
The ROLLS-ROYCE Bulletin "standardised" a 1200 µW/cm² UV-A irradiance and, after
different test, decided the following figures for different luxmeters:
-.ARDROX DLM 1000: 8 foot-candles i.e. 115.78 lux.
- SPECTRONICS DSE 100: 2 foot-candles i.e. 21.52 lux.
- MINOLTA T-1: 3 foot-candles i.e. 3.22 lux.
- LEVY HILL Mk V (predecessor of LEVY HIL Mk VI): 0.9 foot-candle 9.68 lux.
Some time later Rolls-Royce forbade use of the Minolta luxmeter.
Very soon it was shown that many bulbs had to be put so far from the sensor for a
1,200 µW/cm² reading that the "spot" was far larger than the sensor.
A formula for the irradiance was issued to counterbalance this problem:
in which Y comes for the irradiance and Z as the maximum allowed illuminance according
to each approved luxmeter.
A very complex system to measure viewing conditions!
The British Civil Aviation Authority(CAA) in the issue Nbr 95 of the Airworthiness
Note published in 1990 published a requirement almost duplicating the Rolls-Royce
document.
One may wonder why different luxmeters may give so different figures when measuring
the same source, in this case, when measuring the illuminance from a UV-A source.
This is related to the average human eye response curve: luxmeters whose sensors
almost match this response curve are the right ones! Those giving a higher figure
are too sensitive to the blue/violet part of the spectrum--exactly where the visible
photons from UV-A sources are -- when those giving low figures have a response curve
shifted to the red part of the spectrum.
Human eyes are FAR MORE SENSITIVE to blue and violet photons when put in scotopic
or mesopic conditions (i.e. when illuminance is low) than when in photopic conditions
("normal lighting conditions", when the cones, the retina cells which allow us to
"see" colours may correctly work). Inspection booths make the eyes work in mesopic
conditions: the inspectors' eyes are then more sensitive to the blue/violet photons
coming from UV-A sources than they would be out of the booth. Hence a "luxmeter"
whose response curve shifts to the red end of the visible spectrum does not give
useful informations on the "real working conditions".
11- STANDARDISATION CAME TO RESCUE
These hectic times full with doubts and questions gave way to more serene times.
Large Primes and Bodies went to safer and stricter rules for PT and MT, stating for
instance:
- Minimum illuminance figure for products used under white light.
- Minimum (and sometimes also maximum) irradiance figure for inspection under UV-A
radiation.
- Maximum illuminance from visible light in areas of inspection under UV-A radiation.
- Approval of combined radiometers/luxmeters
Finally the ISO 3059 standard, titled ""Non-destructive testing - Penetrant testing
and magnetic particle testing - Viewing conditions" was issued in 2001, though some
important users did not approve it.
This document states that the measurements shall be done in "the real working conditions",
hence no distance specified between, say, the UV-A source and the radiometer's sensor.
Some clues:
- Inspection in white light: 500 lux minimum ON THE SURFACE under inspection. The
Americans require at least 1,000 lux.
- Inspection under UV-A radiation:
- An irradiance of 1,000 µW/cm² minimum, and a MAXIMUM of 5,000 µW/cm² for PT applications.
- An illuminance (visible light) of 20 lux maximum on the surface under inspection,
all the UV-A sources switched "ON", at the working distance. The "real working conditions"
mean that measurement shall be carried out with the same surface/UV-A source(s) distance/location
as those used for inspection, that curtains, if any, shall be as they are when parts
are inspected, etc. The visible light may come from the UV-A sources or from other
sources ("light leaks" in the roof/curtains, etc). Measurement done in the real working
conditions duplicate the REAL inspection situation! This is simpler, more useful,
more reliable than any of the other complex procedures in which the user tries to
measure the visible light from UV-A sources in artificial conditions, then measures
the visible light coming from any other point.
- UV-A irradiance on the parts in the washing station shall be 300 µW/cm² minimum
while the visible light on the parts shall be kept at 150 lux maximum.
ROLLS-ROYCE, quite often a "freelance prime", has specific requirements for UV-A
inspection:
- No UV-A LEDs allowed.
- Surface-UV-A source set at 38 cm (15") for fixed UV-A sources, and at 5 cm (2")
when using battery-operated sources or flexible light-guides.
- A visible light illuminance of 5 lux maximum in the inspection booth in the non-UV-A
lighted area (in the darkened area surrounding inspection area) and 20 lux maximum
in the lighted area (in fact, the inspection area).
-A UV-A irradiance of 1,200 µW/cm² minimum and 5,000 µW/cm² maximum.
-Maximum visible light illuminance value is specified according to the approved combined
radiometer/luxmeter which is used and the (UV-A) ultraviolet irradiance value.
Depending on the primes they are working for, subcontractors may have very different
checks to carry out, many documents to fill out even before beginning the inspection
process!
12- THE END (?)
As you may guess these times spanning over several decades, unheard of by many of
the current users, Level IIIs, auditors, etc was an important and exciting step.
13- ADD-ONS AND COMMENTS
13.1- COMMENTS ABOUT ISO 3059 STANDARD
The ISO 3059 standard points out that measurements shall be carried out in the "real
working conditions". Our American friends have immediately fired this standard due
to "no distance between the sensor and the UV-A source is specified".
The "real working conditions" requirement allows for preventing a - real- case as
this one:
In these old times a prime's specification stated that when measuring the UV-A irradiance
in the inspection booth the UV-A source must be at a distance of 38 cm (15") from
the sensor. With this condition the minimum figure was 1,500 µW/cm².
One day we saw an inspector who did this check, then who pulled up the UV-A source
at about 60 cm (24"). Irradiance on the parts was then well below 1,500. When we
draw his attention to the low irradiance figure, he gave us the prime's specification.
He showed us that the UV-A source's check was to be done at a distance of 38 cm,
but nowhere it was written that INSPECTION should be done with at least 1,500 µW/cm²
or at 38 cm. We had to say he was right. When we reported that to the spec's writer,
first he was disbelieving. But he had also to admit this was true!!! The following
issue of this specification specified that INSPECTION also was to be carried out
under at least 1,500 µW/cm²!!!
By the same token, concerning the visible light in the inspection booth: checking
first the" visible light coming from the UV-A sources", then the ambient visible
light. Wouldn't it be easier, more convenient to check the REAL conditions used for
inspection: all the UV-A sources switched "ON", the distance surface/sources which
is used for inspection?
We even had asked in the EN meetings that the visible light illuminance be measured
at the operator's eyes level (20 lux maximum). Why? Just to prevent using the "outrageous"
PT installations of a manufacturer: parts to be inspected were manually held by the
inspectors in small transparent boxes in which small luminescent UV-A tubes gave
a poor irradiance while the inspectors' eyes were in the plain, abundant white light
of the workshop!
13.2- WARNING
Keep in mind that you may find on the market digital luxmeters which, though satisfactorily
for use under white light (in the absence of UV-A radiation) may give false readings
when used to measure illuminance in UV-A booths as the sensor's window fluoresces
under UV-A!
13.3- AMERICAN STAND ABOUT ILLUMINANCE
Though the ISO 3059:2001 standard is an international standard the American stay
stubborn with the 1,000lux minimum for inspection under white light (natural or artificial).
This requirement for us is not substantiated enough.
13.4- UV-A LEDS (LIGHT-EMITTING DIODES)
After extensive research LEDs manufacturers are able to supply powerful UV-A LEDs.
These UV-A sources are of the utmost usefulness when using fluorescent products,
for instance in NDT (PT and MT) but also for Leak Testing (LT)--in addition to criminal
investigations.
One must be aware of their numerous advantages and of their few drawbacks to use
them the best way. LEDs have a very long life expectancy--except for harsh mechanical
action! We say as a motto that someone beginning today in the PT/MT world and supplied
with a UV-A LED will retire with this same LED still running--even after 42/43 years!!
These sources have a very high efficiency: that means they draw very few energy.
Using rechargeable or non-rechargeable batteries is very convenient. No need for
heavy batteries.
They release a very low amount of heat--once again, due to their efficiency. Hence
a more comfortable condition for the users. They switch "ON" almost immediately.
Their bandwidth is narrow, often near 370 nm, sometimes 385nm. Some now are even
very close to 365 nm. Marked advantage: no WOODS filter required! Hence a low weight.
This narrow bandwidth nevertheless has a drawback: all the UV-A sources used till
now emit a bit of visible light--and other radiations--and the WOODS filter let some
visible photons go through. This is considered as a drawback...but it has also the
advantage that the inspector may "see" the beam, even in the absence of any fluorescent
product.
The user is then sure that he UV-A source is "ON" and that if no PT/MT indication
is seen, it is not because the UV-A source is "OFF". UV-A LEDs do not emit visible
photons. So it is a good idea to have a very weak pilot light to confirm the LED
is "ON" or a very weak visible light along the UV-A beam running only if the LED
is running.
You may have noticed that the wavelength may be different from 365 nm. The above
mentioned wavelengths do not prevent the fluorescent dyes detection: these dyes are
efficient for a wide range of wavelengths. But MEASURING irradiance will be the problem:
radiometers sensors are more or less centered on 365 nm (the top of their sensitivity
curve shall be in the vicinity) and the bandwidth shall be +/- 10 nm at halfway the
sensitivity curve. They incorporate other wavelengths than the 365 nm that are in
all the other UV-A sources. But due to their design LEDs have a narrow bandwidth.
So if the LED emits at 385 nm it will be on the side of the sensor's response curve,
out of the maximum sensitivity area. Even when the LED emits at 365 nm it does not
have any other wavelengths photons which would have increased the figure.
In short when using an LED it is likely that the irradiance reading be lower than
the real figure! We have to consider these measuring/calibration points as of the
utmost importance.
Nevertheless LEDs are probably the future source of choice to make fluorescent dyes
visible: 365 nm may still be used but 405 nm will probably be preferred for safety
reason. This point will be explained in a soon-to-be-published DPC Newsletter.
In any case LEDs will give the best service.
THE END
