54
shows the "relative response curves" of the instrument to
several different gases.
CALIBRATION
STANDARD
100
80
60
40
20
20
40
60
80
100
CALIBRATION
STANDARD
OTHER
GASES
OTHER
GASES
RELATIVE
LEL METER
RESPONSE
ACTUAL LEL CONCENTRATION
Figure 2.2. Relative response curves
Note that the response to the gas to which the instrument
was calibrated, the "calibration standard”, is still precisely
accurate. For the other gases the responses are a little
off.
In the case of some gases the readings are a little high.
This results in the instrument going into alarm a little bit
early. This type of error is not dangerous, since it results
in workers exiting the affected area sooner than they
absolutely have to.
Gases which produce lower relative readings than the
calibration standard can result in a more potentially
dangerous sort of error. In the chart example above the
"worst case" gas only produces a meter reading of 50
percent LEL even when the actual concentration is 100
percent explosive. If the alarm were set to go off when
the display reads 50 percent LEL, the alarm would sound
simultaneously with the explosion!
If on the other hand the alarm is set to go off when the
display reads 20 percent LEL, a 50 percent concentration
of the same "worst case" gas is enough to cause an
alarm.
It may be seen from the graph that the amount of relative
error decreases the lower the alarm point is set. If the
alarm point is set at 10 percent LEL, the differences due
to relative response of the combustible sensor are
minimal.
When it is not possible to calibrate directly to the gas
to be measured, or when the combustible gas is an
unknown, an alarm set point of 10 percent LEL or
less should be selected.
In the new standard for "Permit Required Confined Space
Entry" (29 CFR 1910.146) OSHA has determined that a
combustible hazard exists whenever the concentration of
combustible gas or vapor exceeds 10 percent LEL. Per
this standard confined spaces with concentrations which
exceed 10 percent LEL may not be entered. Likewise,
workers are required to immediately leave anytime
readings exceed 10 percent LEL.
The standard combustible alarm set-point for the
PhD Ultra is 10 percent LEL.
2.1. Calculating relative responses
There are theoretical ways to estimate the relative
response of a sensor calibrated on one combustible gas
to exposure to another gas. This is done by taking the
actual instrument reading, and multiplying it by a
correction factor.
It is very important to understand that if an error is made
in determining the specific kind of gas present, and the
wrong correction factor is used, the accuracy of the
calculation may be significantly affected.
In actual practice, the relative response varies
somewhat from sensor to sensor.
The response ratios may also shift over the life of a
particular sensor, especially in the event the sensor
loses sensitivity as a consequence of being
“poisoned”.
It is very important to treat gas concentration
calculations based on theoretical relative response
ratios cautiously. Correction factors for PhD Ultra
combustible gas sensors:
Combustible
Gas / Vapor
Correction factor
when instrument
is calibrated on
Propane
Correction factor
when instrument is
calibrated on Methane
Hydrogen 0.54
0.83
Methane 0.65
1.0
Propane 1.0
1.5
n-Butane 1.0
1.5
n-Pentane 1.1
1.7
n-Hexane 1.2
1.8
n-Heptane 1.3
2.0
n-Octane 1.6
2.5
Methanol 0.65
1.0
Ethanol 0.76
1.2
Isopropyl
Alcohol
1.0 1.5
Acetone 0.93
1.4
Ammonia 0.46
0.71
Toluene 1.6
2.5
Methyl Ethyl
Ketone
1.2 1.8
Ethyl Acetate
1.2
1.8
Gasoline
(Unleaded)
1.1 1.7
2.1.1. Using correction factors
As an illustration, consider a PhD Ultra calibrated on
methane, which is then used to monitor ethanol. When
calibrated to methane, the instrument is actually less
