Industrial ppm CO measurements

By: Scott Cowe Of Dover Gas Technologies

It is a well known and heavily publicized fact that carbon monoxide is a dangerous gas even at low exposure
levels. While measurement solutions for residential applications are fairly simple, industrial and commercial
applications require a much more sophisticated instrument. Problems with the accuracy of many measurement
devices arise due to zero instability and cross sensitivity to other gasses on the sensors especially in industrial
situations where temperature changes and the presence of other process gasses are an ongoing part of everyday
Upper exposure limits for CO in the work place are normally set to provide alarms or warnings at 25-50ppm,
sensor inaccuracies often cause alarms to be triggered when in fact no actual carbon monoxide danger exists.

When higher levels of CO are present, proper safety procedure calls for the halt of production and the
evacuation of all personnel from the area until the levels can be verified, reduced and the source
identified. These precautions taken against this potentially deadly gas show responsibility on the part of the
employer and can save the lives of many who work to make the company profitable. The resulting down time can
however have many detrimental effects including, employee stress, safety concerns from outside agencies, and
reduced production, it is therefore vitally important that we ensure that CO alarms are set off only by actual
increased levels of carbon monoxide. In an effort to reach this goal there are a number of considerations to be
looked at, the following information is provided for that purpose.

The most common types of detectors used for Carbon Monoxide measurements are,

1) NDIR or infra-red which although is very specific to the gas being measured requires a warm-up time, is fairly
large, can consume larger amounts of power making it unsuitable for small or portable instruments, and is more
expensive. For ranges of CO measurement in industrial uses other than Low ppm this technology is by far the
number one choice.

2) Solid State, while this technology is small and cost effective it is not selective enough for CO only
measurements and usually has higher temperature drift making the zero unstable.

3) and Electrochemical which is the primary choice for the majority of Carbon Monoxide analyzers on the market
today due to it’s many benefits which include, size, weight, power, cost, and proven performance.

The remainder of this article will deal only with the electrochemical sensor.

There are a number of electrochemical sensor manufacturers worldwide and each of these has a number of
sensors designed for carbon monoxide measurements in different applications, choosing the right one is the key to

The output from most of these sensors is very low, (pico amps per parts per million) so even subtle
changes or correction procedures must be dealt with using extreme care in order to preserve the integrity of the
Sensors designed for higher concentrations of CO have a lower output per ppm which can cause it to have a
higher temperature coefficient, this leads to zero instability making it unsuitable for our task.
Carbon and chemical filters are often attached to the face of the sensor or added in the sample stream to diminish
the effects of cross interference to many of the commonly encountered such as H2S, SO2, NO x, etc..
This filter, it’s efficiency and life span can be an important consideration when measuring CO in the 0-50ppm
In a number of applications background levels of Hydrogen given off by nearby processes have been found to
cause major problems with accurate low level carbon monoxide measurements and alarms. While it is possible
for hydrogen levels to cause safety concerns that should be reported, the limit is magnitudes higher than what
causes the CO alarms to trigger on most instruments. A hydrogen level of as little as 60ppm can trip the CO
alarms of many analyzers. The lower explosive limit (lel)for hydrogen is 4%, ten percent of that l.e.l. would be
4000ppm, so we see that the 60ppm is insignificant to personnel safety but remains problematic to the carbon
monoxide measurement. This cross interference can not be easily overcome with chemical filters and therefore
required a different approach. It has been found that by measuring the hydrogen separately and using that signal
in conjunction with the mixed CO/H2 signal the majority of the interference can be nulled out, again manufactures
deal with this differently with varying results. The use of two separate sensors can create a compensation lag
time and there may be temperature coefficient differences between the sensors, either of which will cause errors
in the readings and provide false alarms. Similar temperature problems can be found with units that use a single
sensor but do the electronic nulling and temperature compensation remote of the sensor itself.

There are sensors now available that have separate electrodes for the H2 and CO/H2 signals with the nulling and
temperature compensation circuit attached directly to the rear of the sensor itself. The output of this surface
mount board is in the millivolt range making it less susceptible to interference and change from wiring,
connectors and temperature effects. Each sensor and attached PCB come completely pre-calibrated for H2 cross
interference and relative CO output. Only a single calibration gas is required over the life of the sensor (2 years)
with no need for re calibration of the hydrogen signal nulling found in any of the units put into service. In
addition, the internal chemical filter provides excellent reduction of cross interference due to other common
gasses found. This customer replaceable sensor has proven itself over time to provide consistently accurate
results in demanding applications against competitors instruments.

Dover Gas Technologies Inc. incorporates such a sensor, we believe it best suits the needs
addressed above and use it in many of our portable, transmitterHealth Fitness Articles, and monitor products for low level ppm CO


» More on Technology