AMBIENT TEMPERATURE TESTING ISSUES FOR

PROCESS ANALYZERS

Bob Farmer, Udo Gellert, Ty Campbell

Siemens Industry, Inc.

408 W. Highway 60

Bartlesville, OK 74006 USA

KEY WORDS

ambient temperature, temperature testing, chromatograph, process analyzer, environmental test

chamber, electronics life, Arrhenius Equation

ABSTRACT

Process analyzers are often distinguished from similar analyzers used in other applications by

their physical mechanical and electrical characteristics. In particular, process analyzers are

usually intended to be installed on-line in processing units in difficult ambient conditions as

opposed to analyzers which are intended to be used indoors in office or laboratory conditions.

For example, process analyzers may be built to withstand vibration, water exposure, atmospheric

dust, atmospheric chemicals and widely varying ambient temperatures. Manufacturers conduct

tests on performance of their analyzers to these varying installation conditions and specify their

products accordingly. This paper looks specifically at testing techniques used to validate process

analyzer behavior in varying ambient temperatures. Manufacturers commonly use

environmental test chambers and other equipment to conduct tests. But, as described in this

paper, test conditions may incorrectly simulate real world conditions and tested analyzer

performance may inadequately represent actual analyzer performance after installation. This

paper describes test set up and apparatus which may enable manufacturers and users to better

predict real world performance.

INTRODUCTION

Modern industrial chemical processes computer control systems demand and rely on continuous

analysis of process chemicals. An important reason for this in most processes is the desire to

improve process profitability in several ways and the control algorithms in many processes often

use continuous composition analysis. Therefore, the payback or value of automated control

systems is strongly influenced by the reliability and availability of the analysis information. This

in turn brings attention to the reliability of the process analyzers themselves.

In this paper, the reference to “reliability of process analyzers” shall refer to the ability of an

analyzer to be operating correctly, measuring process chemicals and returning valid information

to automated or manual process controls. The higher the reliability, the greater is the percentage

of time that an analyzer is performing its functions correctly and, therefore, the higher is the

payback, or return on investment, that the owner of the analyzer receives.

In this paper, the term “process analyzer” can be applied broadly to all kinds of hardware and can

include simple devices measuring a single variable as well as more complex analyzers measuring

severable variables. However, examples used in this paper for illustration are taken from an

evaluation of a gas chromatograph which is asserted to be a relatively complex type of analyzer

containing relatively more software and electronics. The reliability discussion in this paper is

more applicable to complex analyzers.

Several factor affect analyzer reliability. This paper focuses explicitly on an understanding of

the way that ambient temperature conditions may be involved and the way ambient temperature

affects electronic reliability. This paper does not report new information or findings but is

intended to help owners of process analyzers understand how testing for ambient temperature

tolerance is conducted by some manufacturers and how resulting ambient temperature

specifications may be interpreted when designing analyzer installations.

SITUATION AND BACKGROUND

To resolve several application concerns, continuous process analyzers used in industrial chemical

or hydrocarbon processing plants are installed on line, often installed in the plant facility close to

the source of the chemicals being measured, rather than in a laboratory. The installation

environment is often hostile to the operating health of electromechanical analyzer equipment.

Specifically: utilities such as electrical power and consumable gases may vary more than is

normal in an office; the atmosphere around the analyzer may contain corrosive, dirty or

flammable gases; and the ambient operating temperature and humidity around the analyzer may

be unstable or extreme.

Process analyzers are designed by their manufacturers to accommodate such undesirable

installation conditions. In fact, in some situations it seems most desirable for an analyzer to be

installed directly out doors, directly on process pipelines or vessels, or mounted in some manner

without supplemental protection from the environment. Various manufactures provide products

designed to permit this. This can significantly reduce the cost of an installation.

However there are limits to the ability of manufacturers to enable this. Particularly, as it regards

ambient temperature, electronic and mechanical components used in the analyzers construction

may become unstable or fail if temperature extremes are too great. Additionally, routine

maintenance of process analyzers may be impaired if working conditions for human technicians

become too extreme.

Therefore, the owners of process analyzers often find that they must provide protected local

conditions for the installed equipment. Such protection can be provided by buildings which shall

be referred to here as “analyzer shelters” or simply “shelters.” Analyzer shelters are usually a

small building located in a process plant which provides protection from weather. Protection

may be simply shielding from direct sunlight, wind, snow and rain or it may include modified

climate through heating and air conditioning of the shelter interior.

FIGURE 1 – EXAMPLE SHELTER FIGURE 2 – EXAMPLE PROCESS

INSTALLATIONS INSTALLATION

Clearly, the requirement for an analyzer shelter adds significantly to the cost of a total analyzer

installation. It has been reported variously that a shelter support for a complex process analyzer

will add project cost of 2 or 3 times the base cost of just the analyzer. With such a significant

potential cost impact, analyzer owners are motivated to minimize shelter design and rely on the

capabilities of the analyzer itself as much as possible.

Many tradeoffs arise in this discussion. These have been discussed in the literature and are

outside the scope of this paper. However, to provide background, a summary list of issues

follows:

For a single analyzer, taken alone, the total cost of installation with no shelter is lower

than with a shelter. On the surface, this seems to define the economic cost/benefit

analysis.

However, even if a certain analyzer is designed not to require separate shelter, other

analyzers or apparatus other than the analyzer itself may be present which does require

shelter. Therefore the incremental cost of making an existing shelter large enough to

accommodate an analyzer may be relatively low, changing the overall economic

cost/benefit analysis.

Apparatus that must be provided for an analyzer include utilities such as power and

consumable gases as well as communication and maintenance interfaces. Bottles,

regulators and gauges, electrical connectors and other equipment must be provided

regardless of the shelter situation.

The suitability of an analyzer for mounting without shelter is also governed by its design

for maintainability. If an analyzer requires significant or detailed human attention for

periodic or routine maintenance or for abnormal repair maintenance, then the cost, length

and effectiveness of human time may increase if the analyzer is located in undesirable

conditions for the human.

The suitability of a particular analyzer installation for mounting without shelter is

influenced by the location of the sample extraction point. If the extraction point is

difficult to access, such as in a high location or near heavy machinery, it may be

necessary to provide shelter for the analyzer for other reasons.

EFFECT OF AMBIENT CONDITIONS ON THE ANALYZER

In all situations regardless of the installation with or without shelter the performance and

reliability of the analyzer will be affected by varying ambient conditions. Ambient temperature

has several effects but these can be grouped in two broad categories.

EFFECT ON MEASUREMENT RESULTS

The operational behavior and measurement results reported by the analyzer are affected by

ambient temperature. This occurs because of hardware responses to changing temperature and

because of chemical responses affecting the fluids being analyzed. Following is a partial list of

and brief description of some of these effects.

Mechanical components such as regulators, flow controllers, flow restrictors, reference

junctions and other may physically change dimensions and flexibility as temperatures

change. These changes alter the way in which the analyzer handles the sample being

measured and can result in differences in reported results.

Electronic components such as amplifiers and voltage and current reference circuits are

affected by the temperature of the electronic devices. Changes in current and voltages

can directly affect the apparent measurement values and reported results.

Sample fluids being delivered to the analyzer may be exposed to ambient temperature

changes and experience physical or chemical changes. Good sample handling and

sampling system design is intended to minimize these effects through the use of heat

tracing, insulation and other techniques.

Utility gases including air, carrier and calibration or reference standards delivered to the

analyzer may also be affected by changes in ambient temperature. Reference standards

are often heated or insulated to minimize ambient temperature influences. But other

utility gases, even those not directly involved in the measurement, may affect analyzer

performance. For example, instrument air which changes in temperature may couple

those changes to the analyzer hardware.

EFFECT ON ANALYZER ELECTRONICS RELIABILITY AND RESULTS

The reliability of the analyzer is directly affected by ambient temperature. Electronic

components in particular have a life time expectation that is affected by operating temperature.

(Mechanical components, notably moving parts, can be affected as well. However, the focus of

this paper is on electronic issues.) In general, the issues regarding electronics are:

a) failure at high or low temperatures; and;

b) life reduction at high temperatures.

In a process analyzer, the term “electronics” is commonly thought of as a zone or compartment

of the analyzer which contains electronics. Or the term may be thought of as an individual

circuit board. However, in considering the effect of temperature on process analyzer reliability,

one must begin by considering the specific electronic components and devices, or “chips,” which

are mounted on a circuit board or other assembly.

Electronic component manufacturers specify their devices, such as semi-conductor chips or fuses

or transformers or discrete resistors and capacitors, to have a “maximum operating temperature”

which must be considered by the electronics assembly designer working for the process analyzer

manufacturer. Usually, a device maximum operating temperature is regarded as a “failure point”

at which, if exceeded, the chip or other device will cease to operate correctly. In that regard, the

process analyzer designer attempts to allow an operating safety margin at a few degrees lower

temperature. (See also below for a further discussion on testing issues.)

Again, in common thought, “failure” of the electronic device is taken to mean total failure in

which the device simply ceases to operate. While that may be true, in practice “failure” is often

much less well defined. In fact, failure may be hidden to the point that it is not even observed

casually. Furthermore, a device which has failed due to high temperature may become fully

operational again when temperature is lowered. Here are some examples of temperature related

failures.

A device may “burn out” and cease to operate entirely and permanently and it may not

recover when temperature is lowered.

A device may begin to operate abnormally and in an observable manner. For example, a

digital chip may operate too fast or too slowly or may switch intermittently. The device

may resume normal operation when cooled or it might not.

A device may fail to meet its stated operating specifications. This may be exhibited as a

loss or excessive increase in amplitude gain in analog circuits; excessive signal or

background noise; irregular switching or speed in digital circuits, random loss of memory

or state in digital circuits; altered activation thresholds in fuses or comparator circuits;

and more. These characteristics may or may not reverse when the device is cooled.

For process analyzer users, the last of these examples must be taken as particularly significant.

A simple – and possibly small - change in gain, speed or other behavior may escape notice

entirely. If electronics fail, they fail; but if electronics simply behave differently, an analyzer

may report a measurement in a way that seems proper but is in fact very wrong. For example, a

5% gain in a detector amplifier might result in a 5% error in assumed measurement. Worse, a

fluctuating gain in the amplifier might easily be seen as actual process variation.

This paper does not address the mechanisms of temperature-induced failure. However, for

background, here is provided a list of some of these mechanisms. This list is illustrative only

and not exhaustive.

Junction failure. Semiconductor devices depend on the interface between two dissimilar

materials to function properly. For example, purified silicon which is deliberately doped

with a second material is placed in close physical contact with silicon doped with a

different material. This creates a junction which has certain electronic behavior. At

elevated temperatures, the junction’s behavior fails due to atomic migration or loss of

voltage differentials.

Molecular migration. Small electronic devices often involve deposition of minute

quantities of material on an inert substrate. For example, this technique is used for fuses,

resisters and other normally passive components. In this case, the material may migrate

and separate or change other characteristics at high temperatures.

Physical failure. In most cases, electronic devices involve some way to attach

microelectronics to the outside world. This is done through physical interfaces between

semiconductor material and other material such as gold or copper. These physical

interfaces may be relatively fragile and may separate or decay at high temperatures.

EFFECT ON ANALYZER ELECTRONICS LIFE

While actual electronic device failure is significant, whether observed or not, process analyzer

users must also understand that failure is not the only way that temperature can affect analyzer

reliability and performance. Specifically, the long term reliability of any electronic component is

degraded by long term operating temperatures – even temperatures which are well below the

device’s specified maximum operating level.

In general, the life of electronic semiconductor circuits is affected adversely by temperature as

indicated by the Arrhenius Equation. This equation predicts approximately that each 10°C

increase in operating temperature reduces mean electronic component operating life by 50%.

The Arrhenius Equation is typically used to model the acceleration of the temperature dependent

physical processes that lead to functional wear-out. (1)

CR = M * e **( -Ea / k T ) [M times e raised to exponent (-Ea/kT) ]

where:

CR is the “Process Rate Coefficient” (applies to chemical as well as physical processes)

M is an experimentally determined constant specific to the materials and methods used

e = 2.718281 (natural constant)

Ea is the “Activation Energy” for the physical process(es) that lead to wear out, in eV

k is Boltzmann’s Constant, 8.617 x 10-5 eV / K (8.617E-5 in Excel notation)

T is Temperature in degrees Kelvin (Celsius degrees plus 273.15)

eV means “Electron Volt”

The general case is to determine the component’s service life operating at room temperature.

Thereafter we are interested in the probable service life when used at a higher temperature. We

use the comparative form of the Arrhenius Equation to estimate a service life acceleration factor:

Acceleration Factor = e **[ (Ea / k) (1/T1 - 1/T2)]

where

T1 is the reference temperature (e.g. 25ºC / 298.15ºK)

T2 is the actual use temperature.

The Activation Energy, Ea, is determined empirically. However, precise calculation is not

significant in this paper; this reference is intended to document the relative effect of temperature

on life. This effect is shown relatively by the following table and chart.

TABLE 3. INDICATED DEGREDATION IN TYPICAL ELECTRONIC CIRCUIT LIFE

BASED ON ARRHENIUS EQUATION

Operating

Temp ºC

Operating

Temp ºK

Acceleration

Factor

Projected Service Life

(1 / Acceleration Factor)

25 298.15 1 100.00%

45 318.15 12 8.65%

65 338.15 100 1.00%

85 358.15 679 0.15%

105 378.15 3,769 0.03%

125 398.15 17,607 0.01%

FIGURE 4 – ELECTRONIC LIFE VERSUS OPERATING TEMPERATURE

It is not uncommon for process analyzers to be operated at elevated temperatures. However, the

fact remains that analyzers operated at elevated temperatures – even temperatures that do not

exceed maximum specifications – will ultimately fail sooner than analyzers operated at cooler

temperatures.

A simple example is illustrative. For example, if a decision is considered to mount a

chromatograph outdoors, in a plain shelter or in a shelter with no air conditioning, one should

consider lifetime issues. Suppose the analyzer uses a certain number of electronic boards, “n.”

Then suppose the electronics have a life expectancy of 6 years at room temperature, 25°C (77°F).

In this case, one might expect to have the analyzer down for electronic maintenance “n” times in

6 years assuming an air conditioned shelter. The cost of each incident is the cost of the part plus

the cost of lost benefit – for example the cost of having the process out of full control during

maintenance. However, if the analyzer is installed without air conditioning and the average

ambient temperature raises 10°C to 35°C (95°F) or even 20° to 45°C (113°F), then the life

expectancy would fall to 3 years or even to 1-1/2 years.

In such a situation, it is easy to see that a shelter with air conditioning, while expensive up front,

might be easily justified by improved reliability of the process analyzers.

AMBIENT CONDITIONS FOR THE ANALYZER

So, in any situation, it is critical to consider the effect of ambient temperature on analyzer

performance and reliability. Process analyzer manufacturers generally make statements about

the suitability of their products for use in varying ambient conditions. The remainder of this

paper will clarify the considerations involved and will comment on some of the testing methods

behind such statements. This discussion is intended to provide the reader with a way to

understand claims being made and to use this understanding to improve analyzer installations.

-20%

0%

20%

40%

60%

80%

100%

120%

0 20 40 60 80 100 120 140R

ela

tive

Se

rvic

e L

ife

Temperature °C

Relative Service Life

SUMMARY OF ANALYZER AMBIENT SPECIFICATIONS

Process analyzer manufacturers provide their users with a statement or specification of tolerance

for ambient installation conditions. Usually this is a fairly simple statement of “minimum and

maximum ambient operating temperature.” Process analyzer users are expected to install the

units in an environment which does not exceed these limits.

Following is a table indicating the range of ambient temperature specifications in products which

are currently offered for sale by a variety of manufacturers. The information here comes from an

internet check of published product specifications. The actual source data therefore is

commercial and proprietary to the various manufacturers. Since the table is intended to be

illustrative only, the authors assert its contents and do not cite the individual references.

TABLE 5. SAMPLING OF PROCESS ANALYZER AMBIENT OPERATING

TEMPERATURE SPECIFICATIONS

Survey Item Survey Scope

Number of different analytical

technologies represented

15

List of technologies represented Gas Chromatography; Non-Dispersive Infra-Red; Ultra-

Violet; Chemo Luminescence; paramagnetic O2, Electro

Chemical (multi); Flame Ionization; Fourier Transform

Near Infra-Red; Visible; Non-Dispersive Ultra-Violet;

Tunable Diode Laser; electrochemical O2; Thermal

Conductivity; electro-chemical, ultra-low CO; electro-

chemical, ultra-low CO2

Number of different process

analyzer manufacturers represented

8

Number of different product types

or models sampled

33

Survey Results Measurement

Lowest ambient operating

temperature

(Lowest with specification

reduction or supplemental

packaging)

-20°C

(-40°C)

Average low ambient operating

temperature

-2°C

Average high ambient operating

temperature

48°C

Highest ambient operating

temperature

60°C

Other related information

commonly or occasionally provided

Ambient humidity limits; NEMA or IP packaging type;

protection required (or not) from sunlight, rain or snow

To meet acceptable ambient operating conditions, the owners of the analyzers are expected to

install the analyzers in an environment that falls within the extremes. Alternatively, the owners

are expected to create a suitable environment by installing the analyzer in a shelter with climate

control as described above.

However, since such shelters add significant cost, owners desire to minimize the design. This

leads to the need to understand what makes up the ambient temperature experienced by the

analyzer and the effect of that on analyzer reliability, life and performance.

LAYERS OF THERMAL DESIGN

It is first necessary to consider the multiple layers involved in the thermal design of a process

analyzer. A process analyzer systems designer must begin with a desired installation situation

and consider that to be the “ambient condition” for the complete analyzer. The illustration in

Figure 6 can be used to indicate the points of concern in thermal design.

FIGURE 6 – HEAT SOURCES AND HEAT DISIPATION PATHS

Of concern to the system designer is “the component subject to failure” and the temperature of

the “ambient environment.” That is, the designer needs to ensure that his complete system can

operate normally in the specified ambient environment without allowing the component subject

to failure to exceed its maximum operating conditions.

As indicated in the diagram, and as may be obvious to the reader, the temperatures of these two

design elements are not the same. Specifically the ambient environment is only one source of

heat. The other sources are these.

Any external radiant source such as the sun or hot process operating equipment or

machinery. Such radiant energy will typically raise the temperature of the process

analyzer without raising the temperature of the ambient environment.

Any other internal heat sources such as other electronic components, human interfaces or

screen devices, detector sources or mechanical components in the analyzer itself. These

sources of heat raise the temperature around the component subject to failure without

affecting the analyzer’s ambient environment.

Ambient

environment

Electronics enclosure

Radiant heat source

Component

subject to failure

Other internal

components and heat sources

The “component subject to failure” itself.

A comment needs to be made on the last point. The component itself generates heat. In fact,

some individual digital electronic chips generate large amounts of heat – often several watts.

The component must dissipate this heat into its own surrounding environment. The need to

dissipate this heat is what causes the component manufacturers to specify a “maximum operating

temperature” for the chips.

However, the component’s maximum operating temperature does not equate to the analyzer’s

ambient temperature. Because of the other sources of heat, the component’s immediate

environment is invariably hotter than the analyzer’s ambient.

(It is interesting to note that electronic component specifications have improved in recent years,

meaning that the components are permitted to operate in hotter situations today than they were a

few years ago. For example, some years ago, commercial device maximums were commonly

50°C with expensive “military grade” options to get to 60° or 70°C. Today, commercial chips

are commonly specified at 70°C with many specified to 85°C or more. Several factors

contribute to this but one that is most important is the fact that electronics today operates on less

power than the equivalent electronics of before. Therefore, the component itself has less heat to

dissipate and a lower differential to its own environment is required.)

Nevertheless, the analyzer’s ambient environment is critical to the overall system design. This is

because of the need of the analyzer to dissipate heat. As indicated in the diagram, the component

must dissipate its heat to its own environment. This heat is distributed in some manner internally

and must then radiate into the ambient environment.

Therefore, the success of the overall analyzer system design is directly related to the ability of

the analyzer to dissipate electronic and other accumulated heat energy into its own environment.

Certain variations may exist on this theme. To improve heat loss, an analyzer designer may use

various techniques. These are illustrated in Figure 7 and described here briefly.

FIGURE 7 – METHODS FOR ACCELERATING HEAT DISSIPATION

The methods as indicated are as follows.

A. The analyzer manufacturer may provide or require simple shielding or insulation to

prevent heat gain from external radiant heat sources.

B. The internal electronic components may have small fans which accelerate heat

dissipation away from hot sources. For example, power supplies and microprocessors

often include miniature fans that are integral to their construction. Additionally,

however, a fan may be used internally simply to accelerate internal air contact with a

metallic outer skin for the purpose of accelerating conductive heat loss.

C. The electronics compartment may use a fan to move external air into or out of an

electronics compartment. However, this approach may be limited for fire-prevention

reasons if the compartment must be closed to an outside flammable atmosphere.

D. An external source of cold air may be used to chill the interior.

E. A heat conductor tube and radiator may be used to improve conduction of heat from

inside to outside by effectively increasing radiant area. Such tubes are often filled with a

liquid that is a good thermal conductor. The use of these kinds of devices can avoid the

problem of exposing the interior to flammable gases in the exterior.

Ambient

environment

Electronics enclosure

Radiant heat source

Component

subject to failure

Other internal

components and heat sources

A

B

C

D

Cold source

E

THE IMPACT OF INSTALLATION CONDITIONS ON HEAT DISSIPATION

Each of these approaches is straight-forward in concept. They influence the overall system

design in usability and cost. In all cases, they demonstrate the bottom-line need which is to

transport internal heat to the outside environment.

The ability to transport heat in this kind of application is most strongly influenced by two factors.

Simple thermodynamic rules apply.

1. The primary factor is the temperature differential between the analyzer outer skin and the

air adjacent to the analyzer skin.

2. The second factor – which supports the first – is the rate at which heated air near the

analyzer skin can be moved away from the analyzer, by convection or forced air

movement, so as to maintain highest available temperature differential.

This understanding is critical to the owner’s design of the installation of a process analyzer.

Specifically, when an analyzer manufacturer specifies “maximum ambient temperature”, the

owner must understand what the manufacturer intends. However, in the authors’ survey of

product specifications referred to in Table 4 above, no manufacturer provided any description of

this intent.

Therefore, the owner must assume the most severe case. The manufacturer most likely is

expecting that the analyzer’s environment be correspondent to the conditions under which the

analyzer was tested.

ANALYZER ENVIRONMENTAL TEST METHODS AND RELATED CRITERIA

Many analyzers are tested in an environmental test chamber. Such test chambers are enclosures

large enough to house the analyzer under test along with associated test equipment. The test

chambers include machinery that provides very close control of the environment inside the

enclosure. Illustration of test apparatus used with two different sizes of process gas

chromatographs is illustrated in Figures 8 and 9.

FIGURE 8 – TWO SMALL GAS FIGURE 9 – LARGE GAS

CHROMATOGRAPHS INSTALLED CHROMATOGRAPHS INSTALLED

IN A WEISS SB 11-500 TEST IN AN ESPEC TEST CHAMBER

CHAMBER

The interior of these chambers can be very tightly controlled for both temperature and humidity.

To accomplish this, the chambers usually circulate large quantities of air inside. Therefore, the

test conditions conform to “high air velocity in the working space” conditions as defined by

IEC 60068-2-1; 2007. This is air which exceeds a velocity of 0.5 m/s and may cause a

temperature decrease of up to 5°K on the surface of the unit being tested.

However, many actual installations of process analyzers are closer to “low air velocity in the

working space” conditions as defined in the IEC specification. This is a situation in which the

air at the surface of the analyzer does not move very fast. Note that low velocity conditions can

exist in many typical installation conditions including:

inside closed shelters

inside naturally ventilated shelters

inside ventilation-assisted shelters (wind turbines)

outdoors.

Furthermore, this condition can exist inside shelters with powered climate control if air

movement from the air conditioning unit is blocked before reaching the analyzer.

The difference on analyzer reliability and lifetime can be significant. Analyzer owners desiring

to maximize reliability or life should be aware of this difference.

COMPARATIVE TESTING – AN EXAMPLE

A demonstration of this effect was made through actual testing of a Siemens Maxum edition II

process gas chromatograph. This unit has a specified allowance for installation in ambient

environments from -18° to +50°C. This, along with utility and air requirements, a statement of

packaging type and humidity tolerance is the only information that is routinely provided by the

manufacturer to chromatograph owners and installers.

The upper ambient temperature limit is imposed by the device specifications of certain electronic

components inside the electronics enclosure located at the top of the chromatograph. The

electronics enclosure is normally sealed closed during operation but is purged with instrument

air. This purge air is vented through a restricted pressure relief port so as to maintain a small

positive pressure above atmosphere. It is presumed by the manufacturer that instrument air

entering the enclosure is at ambient temperature. Therefore, all electronics cooling occurs

through conduction of heat in the interior to the outside world through the metal skin of the

enclosure. It is assumed that there is no significant heating or cooling effect from incoming

instrument air.

To provide for expected heat rise in the electronics enclosure, components used inside by the

manufacturer are specified to have temperature capabilities of 70°C or higher. (At this writing, a

majority of components are rated to 85°C.) This means with an ambient temperature of 50°C

that the internal temperature rise is expected to be 20°C or less. As described in a previous

section of this paper, the component upper temperature limits are device failure or performance

specification limits – not lifetime predictors. To ensure the performance of the complete

analyzer, the manufacturer measures and alarms if internal temperature exceeds 65°C which

means that the manufacturer intends that internal temperature will not rise more than 15°C above

ambient.

Testing was conducted to verify these design constraints. As described in the previous section,

original product testing conformed to “high air velocity in the working space” conditions. In this

situation, the product conforms to its design goals and product specifications in most available

product configurations. See Figure 10. In most operating conditions measurement points inside

the electronics enclosure fall below the design criteria of 65°C. Note, certain complex product

configurations operating at extreme conditions require a reduced ambient temperature allowance

and in these situations, the manufacturer specifies a reduced ambient of -18° to 40°C. “Complex

configurations” mean for example multiple operating heated detectors and two split

chromatograph ovens each operating at 225°C or 250°C – which is a maximum application

configuration supported by the manufacturer. As stated, a large majority of practical application

configurations are not “complex” and do not require ambient temperature reduction. However,

this issue is not pertinent to the remaining testing which was conducted for this paper.

See Figure 10. The test chamber was operated to create an ambient temperature at 50°C (bottom,

green line). The hottest point inside the electronics enclosure (see also Figure 13) was observed

to be continuously at about 64°C (magenta, middle line) when the oven temperatures were set at

150°C. This is below design targets. When the ovens were operated at 225°C, the hottest point

in the electronics enclosure was just 3°C higher, or about 67°C (top, blue line.)

FIGURE 10 – RESULTS OF INTERNAL TEMPERATURE TESTING OF A GC WITH

HIGH AIR VELOCITY CONDITIONS (THAT IS, NORMAL ENVIRONMENTAL

CHAMBER CONDITIONS)

However, since many real world installation situations are “low air velocity in the working

space” conditions, additional testing was conducted to determine the impact on analyzer

response. Subsequent testing was conducted tests were conducted in various “Situations.”

1. The GC was placed in the environmental chamber directly as illustrated in Figure 11 and

pictured in Figure 8. and was exposed to the ambient environment created by the

chamber. This is a “base case” that corresponds to original product testing and is a “high

air velocity” situation. In this situation, the temperature test chamber produces air

movement of 21.8 m3/min and changes air inside the chamber 234 times per minute for a

chamber volume of 85 m3.

2. An unventilated “shelter” was built around the analyzer using plywood or cardboard.

This protected the analyzer within the shelter from air movement which came from the

test chamber. Various conditions of “closure” were tested in which the open area at the

top of the cardboard “shelter” was increased from 30% to over 55%. Note, the 30%

opening is presumed to be roughly equivalent to having a natural turbine ventilator at the

top of a closed shelter. The >55% situation was considered to be equivalent to a 3-sided

shelter in the field.

3. The internal “shelter” was ventilated with forced air ventilation in which the ventilation

was forced around the analyzer as a whole but not directly on the electronics enclosure.

This forced air movement was a relatively high 11 m3/min.

EC operating temperatures APU present no purge 10 PSI oven air

50C ambient 225C air bath oven 260C 2 FID detectors

45

50

55

60

65

70

0 5 10 15 20 25 30 35 40 45 50

Time [min]

Tem

pera

ture

[°C

]

TC11 Ambient TC10

The conditions and test points are illustrated in Figures 11, 12 and 13. The last two “situations”

above created a “low air velocity” situation for the analyzer. As was demonstrated, these

situations severely restrict convective heat loss from the analyzer under test.

FIGURE 11 – TEST SITUATIONS

FIGURE 12 – PHOTOS SHOWING SIMULATED “SHELTERS” AND AIR

MOVEMENT

Electronics

Compartment

Detector

Compartment

Oven 1 Oven 2

Environmental

Test ChamberOverhead Unpowered Ventilation Simulation

(when present)

Powered fan

(when present)

Plywood or cardboard “shelter”)

(when present)

Bottom closed, top open as tested

FIGURE 13 – ILLUSTRATION OF KEY TEST POINTS

Results of testing and various test conditions are summarized in Table 14. It will be noted that

different tests were conducted at different “ambient temperatures” as set by the environmental

test chamber interior. Therefore, when reading particular test point values from Table 14, it is

important to keep those in the context of the test chamber conditions. If it is desired to know

what a particular test point result might be at a different temperature, the reader may assume

direct correspondence; that is, a 1° difference – either up or down – in the test chamber interior

would translate to 1° difference, also up or down, in the point being measured. This relationship

was demonstrated by the authors while conducting these tests. Note, that for purpose of the

conclusions in this paper, the absolute chamber temperature was not as important as the

differential between various points or between various conditions.

0.75"

1.50"

MAXUM Shelter Wall

102M luc

101M 18

108M 36

125outside

13.00"

Hot Spot

TABLE 14. SUMMARY OF PERTINENT TEST PARAMETERS AND RESULTS

To help the reader identify key measurements from this collected data, a Table of Summary

Comparisons, Table 15 is provided. Three important points shall be noted.

1. The temperature at the skin of the electronics closely matched the temperature of the

internal electronics hot spot in low velocity test situations when conduction of heat away

from the analyzer was primarily by convection. This might be expected and confirms the

importance of assisted methods of transporting heat away from the outer skin of the

electronics enclosure. In low velocity situations, the differential was 0.1° to 0.2°. In high

velocity situation, the differential was 2.7° to 4. 4°. In unsheltered conditions, the

temperature differential was the same as the difference to the ambient.

2. The temperature rise inside the analyzer electronics 14° to 24.2°C above the ambient

depending on the ventilation situation. For well ventilated conditions, this was limited to

14° to 17° and corresponded to design targets for the analyzer model under test. For ALL

shelter situations – with constrained ventilation – the temperature rise was 19.3° to 24.2°.

This means there was an ADDITIONAL rise in electronics temperature of 5.3° to 7.2°

This additional rise may be very significant when considering installation options and is

noted again in the Summary section below.

3. In situations of restricted air movement around the electronics, the temperature drop over

a short 19 mm (0.75 in) distance measured from the skin of the enclosure outwards is a

remarkably high 15°C. This relates to the ability of the analyzer to transport heat away

from itself by convection alone. It is clear that “wind-chill effect”, which means the air

movement on the skin of the analyzer improves heat removal, is significant.

The relatively high temperature differential occurring in a relatively short distance outward from

the skin of the analyzer is directly related to the speed at which natural convection can transfer

heat to the surrounding environment. Forced air movement around the analyzer accelerates this

heat transfer.

TABLE 15. SUMMARY OF TESTING CONCLUSIONS

SUMMARY AND CONCLUSIONS

This paper is intended to demonstrate the importance of considering ambient conditions when

planning the installation of process analyzers. Variations in ambient conditions can affect

measurement performance and electronics reliability. This paper has focused on issues of

electronics reliability. The following points are made.

1. Process analyzer ambient temperature specifications provided by manufacturers may be

misleading to analyzer owners unless the owners understand the test methods used by the

manufacturer. This is because in low air flow situations – including possible outdoor

installations and many real-world installations – the ability of the analyzer to shed heat by

convection alone is much less effective than it is in the high air flow situations created by

the analyzer manufacturer during product testing.

2. Therefore, analyzer installation situations should deliberately ensure high air movement

around an analyzer when operation at high ambient temperatures is possible. Any

enclosure around the process analyzer can severely restrict convective heat transport

away from the analyzer. Such enclosures can cause very significant increase in the

ambient temperature experienced by the analyzer. Additionally, no enclosure means that

the analyzer is solely dependent on natural wind or other air movement to aid convection.

3. While there may be a lower initial project cost to mounting an analyzer without a shelter

and environmental controls, a very simple analysis of the cost of maintenance may show

that the incremental up-front cost can be rapidly paid for by extended operating life and

reliability. Life and reliability are cut in half by each 10°C increase in ambient

temperature rise. This payback may be particularly clear when one considers both the

cost of electronics and the cost of lost process benefit due to analyzer down time.

Additionally, the improvement in human maintenance contributes to this offset.

4. Process analyzer performance failure may occur without notice, warning or casual

visibility if electronic component temperature limits are exceeded. If the analyzer is to be

operated at high or unregulated ambient temperatures, the analyzer user should ensure an

understanding of how the manufacturer has tested analyzer operation so that installation

conditions can be designed appropriately.

5. Process analyzer electronic life is diminished by operation at high temperatures, whether

or not failure occurs abruptly. This is indicated by the Arrhenius Equation and is

independent of a manufacturer’s specifications of ambient conditions for an analyzer.

Higher ambient operating temperatures mean shorter electronic life.

6. Other factors that may influence shelter design or a decision to use no shelter at all are

mentioned. The most important of these is accommodation for human maintenance

personnel and required maintenance activities.

REFERENCES

Reports:

1. Steve Wetterling, MSEE, and Pat Barrett, BSEE, P.E.; “Modeling Temperature Driven

Wear Out Rates for Electronic Components”; Servengr, LLC; September 2007;

http://servenger.com/Resources/Modeling_Temperature_Driven_Wearout_Rates_for_Ele

ctronic_Components_b.pdf

2.

Papers:

3. Kallis, James; Norris, Michael; Hughes Aircraft Company; “Effect Of Steady-State

Operating Temperature On Power Cycling Durability Of Electronic Assemblies”;

Proceedings 12th Biennial Conference on Reliability, Stress Analysis and Failure

Prevention, pp. 219-228, April 1997, Society for Machinery Failure Prevention

Technology; http://kaltechservices.com/Papers/Power%20Cycling%20Durability.pdf

4. Gokeler, Ulrich; Siemens Industry, Inc.; Maurer, Dr. Torsten, Siemens AG; “MicroSAM

Process Gas Chromatograph - Expanding Process Control By Lowering Cost Of

Ownership; Proceedings of ISA Analysis Division AD 2002; April 2002; Denver, CO

Standards:

5. International Standard; IEC 60068-2-1, 6th

Edition, 2007-03; Environmental testing –

Part 2-1; Tests – Test A: Cold

6. International Standard; IEC 60068-2-1, 5th

Edition, 2007-07; Environmental testing –

Part 2-2; Tests – Test B: Dry heat

7. International Standard; IEC 60068-2-30, 3rd

Edition, 2005-08; Environmental testing –

Part 2-30: Tests – Test Db: Damp heat, cyclic (12 + 12 h cycle)

Books:

8. Thermodynamics; Copyright 1959 by John Wiley & Sons, Inc.; Library Of Congress

Catalog Card Number: 59-9356; J. Van Wylen, Chairman, Department of Mechanical

Engineering, University of Michigan