INSTRUMENT SERVANT A MESURER L'EFFICACITE DES APPAREILS A COMBUSTION

APPARATUS FOR MEASURING THE EFFICIENCY OF COMBUSTION APPLIANCES

Abrégé anglais

-1-
A B S T R A C T
"APPARATUS FOR MEASURING THE EFFICIENCY OF
COMBUSTION APPLIANCES"
This disclosure relates to apparatus of the kind
suitable for taking spot measurements of the heat loss
or stack loss and/or efficiency (?) in flue gases
(stack loss) and comprises respective sensors (5, 3)
for producing output signals which vary with the temp-
erature and the concentration of a constituent gas
e.g. O2 of the flue gases and microprocessor-based
computation means (10) arranged to derive measurement
values of (and numerically equal to) the measured
temperature and constituent gas concentration, from the
two sensors and to apply these measurement values in the
computation of a predetermined formula relating the
stack loss or efficiency to the measured quantities.
In accordance with the invention, the apparatus is
arranged to automatically calibrate the or each sensor
from a test measurement prior to deriving the measurement
values. This may be achieved in the case of a sensor
having a non-linear response, by performing a calculation
of a formula defining the non-linear response of the
sensor using a coefficient derived from the test
measurement. The sensor is thus automatically calibrated
and "linearised" from a single test measurement.
The predetermined stack loss or efficiency formula may be
modified for different types of fuel, and temperature and
O2 (or CO2) values used in the calculation as well as

-2-
the result of the calculation may be
presented on a visual display (24). The apparatus
requires a minimum of operator expertise and
reduces the possibility of human error associated with
known stack loss measurement apparatus of the kind
in which the operator is required to refer to charts
to determine the stack loss from separate temperature
and oxygen concentration readings.

Revendications

The embodiments of the invention in which an exclusive property
or privilege is claimed are defined as follows:-
1. Apparatus for measuring the degree of efficiency
of a combustion appliance, comprising a first sensor for producing
an output signal which varies with the concentration of a constit-
uent gas of the exhaust gases of the appliance, a second sensor
for producing an output signal which varies with the temperature
of the exhaust gases, and computation means adapted to receive
the sensor output signals to derive therefrom measurement values
representing the concentration of said constituent gas and the
temperature of the exhaust gases and to apply these measurement
values in the computation of a predetermined formula relating the
degree of combustion efficiency to the temperature of, and the
concentration of said constituent gas in the exhaust gases,
wherein the computation means is operable to calibrate at least
one of the sensors from a test measurement made with that sensor
prior to deriving a measurement value therefrom.
2. Apparatus as claimed in claim 1, wherein the
computation means is operable to calibrate the first sensor from
a test measurement of a test gas having a known concentration
of said constituent gas.
3. Apparatus as claimed in claim 1, wherein the comput-
ation means is operable to derive a said measurement value
from the sensor output signal in accordance with a formula
defining the relationship between variations of the sensor
output signal with the measured quantity, and the measured
quantity.
4. Apparatus according to claim 3, wherein the
computation means is operable to compute the value of a
coefficient of the formula defining the relationship between
variations in the sensor output signal with the measured

-26-
quantity and the measured quantity for subsequent use in
deriving a said measurement value.
5. Apparatus as claimed in claim 4, wherein the following
relationship defines the variation between the output
signal of the first sensor and the measured quantity:
Fractional concentration = 1 - exp - S/k
where S is the output signal value of the first sensor and
k is said coefficient.
6. Apparatus as claimed in claim 5, wherein the
computation means is operable to compute a value for the
coefficient k in accordance with the following formula:
<IMG>
where S' represents the value of the output signal produced
by the first sensor from the test measurement and C' represents the
fractional constituent gas concentration of the test
gas.
7. Apparatus as claimed in claim 6, wherein the computation
means is adapted to calibrate the first sensor from a
test measurement of ambient air, the value of C' being
set to correspond to the nominal concentration of
said constituent gas in ambient air.
8. Apparatus as claimed in claim 5, wherein the computation
means is operable to derive a measurement value of the
constituent gas concentration in accordance with the
following formula:
Fractional Constituent S/k.(1 - S/2k)
gas concentration
where S represents the output signal of the first sensor
produced by measurement of the exhaust gases.

-27-
9. Apparatus as claimed in claim 1,
wherein the computation means is operable, in carrying
out a test measurement of the said constituent gas
concentration of ambient air or other test gas,
automatically to sample the value of the output signal
of the first sensor, to compare its value with that of
a stored value representing the estimated value of the
sensor output signal for the nominal or known concentration
of said constituent gas in the test gas, and if the
difference between the compared values is above a predetermined
limit to repeat the comparison after a predetermined interval
until the difference between the compared values falls
within the limit.
10. Apparatus as claimed in claim 9, including means
for visually and/or audibly indicating when a test
measurement of the constituent gas concentration has been
obtained.
11. Apparatus as claimed in claim 1,
wherein the first sensor is an electrochemical oxygen sensor.
12. Apparatus as claimed in claim 1, wherein the
temperature sensor is arranged to produce an output
signal which varies with the difference between the exhaust
gas temperature and a reference temperature.
13. Apparatus as claimed in claim 12 wherein the
temperature sensor is a thermocouple.
14. Apparatus as claimed in claim 13 wherein the
thermocouple is a type K alloy thermocouple having a
substantially linear variation in its output signal with
the measured quantity over the range of interest.

-28-
15. Apparatus as claimed in claim 1, wherein
the computation means is operable, in deriving a
measurement value from the temperature sensor output
signal, automatically to sample the value of the
temperature sensor output signal to store the sampled value
for a predetermined period and then to re-sample the value
of the temperature sensor output signal, to compare the
two sampled values, and to repeat the procedure until
the difference between two successive sampled values
is below a predetermined limit whereupon to retain one
of these two values as the said measurement value.
16. Apparatus as claimed in claim 1, including means
for visually and/or audibly indicating when measurement
values of the temperature and constituent gas concentration
of the exhaust gases have been derived.
17. Apparatus as claimed in claim 1 wherein the
computation means is operable to calibrate the
temperature sensor from a test measurement of said
reference temperature.
18. Apparatus as claimed in claim 1, wherein the
measure of the degree of combustion efficiency is provided
by an indication of the heat loss or stack loss of the
heating appliance.
19. Apparatus as claimed in claim 18, wherein the
first sensor is an oxygen sensor, and the predetermined
formula for computing the heat loss or stack loss is:
<IMG>
where K3 is constant related to the fuel used by the heating

-29-
appliance, T1 and T2 are the exhaust gas temperature and
a reference temperature respectively, and % O2IN and
%O2OUT are the percentage O2 concentrations of the
combustion air supplied and the exhaust gases
respectively.
20. Apparatus as claimed in claim l, wherein
the measure of the degree of combustion efficiency
is provided by an indication of the operating efficiency
(?).
21. Apparatus as claimed in claim 20, wherein
the first sensor is an oxygen sensor and the predetermined
formula for computing the operating efficiency:
<IMG>
where R is a constant related to the type of heating
appliance, K3 and K4 are constants related to the type
of fuel used, P is a constant related to the moisture
and hydrogen content of the combustion gases supplied to
the heating appliance, T1 and T2 are the temperature of
the flue gases and a reference temperature respectively,
and %O2IN and %O2OUT are the percentage O2 concentrations
of the combustion air supplied and of the exhaust gases
respectively.
22. Apparatus as claimed in claim 19 or 21,
wherein different values of the constants K3 and/or
K4 are stored in the computation means for different
types of fuels, and the apparatus includes means for
selecting said different values for use in the computation
of the predetermined formula.

-30-
23. Apparatus as claimed in claim 1, including
display means for displaying measurement values
derived by the computation means for the constituent
gas concentration and temperature of the exhaust gases,
and the result of the computation of said predetermined
formula using these measurement values.
24. Apparatus for measuring the degree of efficiency
of a combustion appliance comprising a first sensor
for producing an output signal the value of which varies
with the concentration of a constituent gas of the
exhaust gases of the appliance, a second sensor for
producing an output signal the value of which varies
with the temperature of the exhaust gases, at least one
of said output signals varying non-linearly with the
measured quantity, and the computation means adapted to
receive the sensor output signals, the improvement
consisting in that the computation means is adapted
to correct any such non-linearity by applying a correction
derived from the known relationship between the sensor
output signal and the measured quantity to produce
respective output values which vary substantially linearly
with different values of the measured quantities, and
to apply said output values in the computation of a
predetermined formula relating the degree of combustion
efficiency to the temperature of, and the concentration of
said constituent gas in the exhaust gases.

25. Apparatus for measuring the degree of efficiency of
a combustion appliance according to claim 1, wherein the
computation means is operable to derive a said measurement
value from the sensor output signal in accordance with a formula
defining the relationship between variations of the sensor
output signal with the measured quantity and the measured
quantity; and the computation means is operable to compute the
value of a coefficient of the said formula defining the relation-
ship between variations in the sensor output signal with the
measured quantity and the measured quantity for subsequent use
in deriving a said measurement value.
26. Apparatus for measuring the degree of efficiency
of a combustion appliance according to claim 1, wherein the
computation means is operable, in carrying out a test measurement
of the said constituent gas concentration of ambient air or
other test gas, automatically to sample the value of the output
signal of the first sensor, to compare its value with that of a
stored value representing the estimated value of the sensor output
signal for the nominal or known concentration of said constituent
gas in the test gas, and if the difference between the compared
values is above a predetermined limit to repeat the comparison
after a predetermined interval until the difference between the
compared values falls within the limit.
27. Apparatus for measuring the degree of efficiency of a
combustion appliance according to claim 1 wherein the computation
means is operable, in deriving a measurement value from the
temperature sensor output signal, automatically to sample the
value of the temperature sensor output signal to store the sampled
value for a predetermined period and then to re-sample the value
of the temperature sensor output signal, to compare the two
sampled values, and to repeat the procedure until the difference
between two successive sampled values is below a predetermined
limit whereupon to retain one of these two values as the said
measurement value.
28. Apparatus as claimed in claim 12 wherein the temperature
sensor is a type K alloy thermocouple having a substantially
linear variation in its output signal with the measured quantity
over the range of interest.
31

29. Apparatus for measuring the degree of efficiency of
a combustion appliance comprising a first sensor for producing
an output signal the value of which varies with the concentration
of a constituent gas of the exhaust gases of the appliance, a
second sensor for producing an output signal the value of which
varies with the temperature of the exhaust gases, at least one
of said output signals varying non-linearly with the measured
quantity, and computation means adapted to receive the sensor
output signals, the improvement consisting in that the computation
means is adapted to correct any such non-linearity by applying a
correction derived from the known relationship between the sensor
output signal and the measured quantity to produce respective
output values which vary substantially linearly with different
values of the measured quantities, and to apply said output
values in the computation of a predetermined formula relating
the degree of combustion efficiency to the temperature of, and
the concentration of said constituent gas in the exhaust gases;
and in that the computation means is operable to calibrate at
least one of the sensors from a test measurement made with the
said at least one of the sensors.

Description

-1
"APPARATUS FOPc MEASURING THE EFFICIENCY' OF
COMBUSTI()N APPLIANCES"
, " , ~
This invention relates to apparatus for measuring
the degree of efficiency of combustion appliances using
fossil fuels.
A known method of determining the degree of eff-
5 iciency of a combustion appliance9 e~g. a boiler or a
furnace9 involves measurement of the oxygen or C02
concentration and the temperature of the exhaust gases, and
the measured values are then referred to a standard chart
showing either the'btack loss" (proportion of heat loss)
10 or efficiency values for different temperature and
oxygen or C02 concentration values. For different types
of fuel, e.gO solid fuel, fuel oil or natural gas, a
different chart must be referred to. Apart from accuracy
limitations inherent in the use of such charts, a possibility
15 exists of the operator referring to the wrong chart i.e.
to the chart of the wrong fuel, or referring to the wrong
line or column of the chart~ or possibly even misinterpret-
ing or misreading the readings on one or both of the
measurement instruments.
~`~
. .
.
-

Various forms of
apparatus have been propQsed in an at~empt to overcome
these disadvantages by providing means for automatically
determining the degree of efficiency from temperature
5 and 2 or C02 concentration measurement of the exhaust
gases.
_For example, U.K. Patent Application No. 2016707 (pub-
lished September 26,1979) discloses apparatus for perLorming
predetermined algorithms relating the operating efficiency (~ )
to the output signals of temperature and oxygen concentration
sensors of the exhaust gases and embodyiny corr~ctions for the
non-linearity of the 2 and temperature sensors used.
This~ however~ has the disadvantage that separate indications
; of the actual 2 concentration and exhaust gas temperature
15 cannot be separately indicated since the non-linearity
corrections for the sensors are integrally embodied
in the single algorithm.
Another form of apparatus is disclosed in U.K.
Patent No.1,562,576 in which an electronic computing
20 device receives output signals from an 2 or C02 concentration
sensor and a temperature sensor and computes the efficiency
in accordance with a predetermined formula relating
this quantity to the 2 or C02 concentration and the
tempera~ure of the exhaust gases. However, such apparatus
25 assumes a linearity in the variation of the received sensor
signals with the measured quantities. This can rarely be
achieved, particularly in the case of gas concentration
sensors.
0~ ,
.

The problem of calibrating an 2 concentration
sensor is discussed in "Improving the measurement of
2 in flue gases", an article by Alan M.Crossley in
"Power & Works E~gineering", October, 1979.
The proposed solution~however, lnvolves taking
a number of test measurements of test gAses having different
known concentrations of 2 over the range of interest,
and is accordingly cumbersome and time-consuming~
It is an aim of the present invention to provide
10 apparatus for measuring the degree of efficiency of a
combustion appliance which enables one or more of the
above-mentioned disadvantages to be overcome or at least
to be substantially reduced.
According to the present invention, apparatus
15 for measuring the degree of efficiency of a combustion
applia~ce comprises a first sensor for producing an
output signal which varies with the concentration o a
constituent gas of the exh~st gases of the appliance, a
second sensor for producing an output signal wh~ch varies
20 with the temperature of the exhaust gases, and computation
mea~s adapted to receive the sensor output sign~l, to
derive therefrom measurement values representing the
concentration of said constituent gas and the temperature
of the exhaust g~se and to apply these measurement values
25 in the computation of a predetermined formula relating
the degree o combustion efficiency to the temperature o,
and the concentration of said constituent gas in exhaust
gases, the computation mean~ being operable prior to

derivi.n~ a measllre~llerlt ~-alue therefro~n, to calibr~te at
least one of the sensors from a test measurement made
with that sensor3
In ~ preferred ~mbodiment, thi.s i.s achieved
5 in the case o~ a sensor having a Icno~ non~linear rel~t~on~
ship between its output signal and ~he quant~.t;y to be
measured, by computi.ng the value o a coefficient of an
- equ~tion or ~ormula defini.ng that non~lineaL rel&~i.ons~lp
from the sensor output signal produced by the test
10 measurement. This coefficient, whi.ch may conveniently
be referred to 2S a calibration coefficient~ is then
used by the computation means in deriving a measuremer.t
value of tl-le exllaust gases in accordance with the equation
defi~in~ the non-linear relationship between the .serlsor
15 output signal al1d.the quantity to be measured. In this
way, the sensor is calibrated, and ~.he~on l;nearity
of the response of the sensor automatically compensated
simultaneously from a single test measurement~
The degree of efficiency of the combustion
20 appliance may be provided by de.ter~ini.ng the heat loss
or stack loss of the eXhRUSt gases ~ or of the operating
efficiency (~ ) of the applianoeO
The particular predetenmi.n~ formula used in
: computing these val-ues fro~ the derived measur~ment
25 values of the consti.tuent gflS coneentration and the
temperature of th~ exhaust gases rQlies on a previous
computatlon of the stack loss. In the preferred case of
the constituent g~5 ~eing oxyyen, ~he stack loss is prefer--
ably compute~ in accordance with th~ ~ollo~in~ formul~:
.
~ , ~

3 ( 1 2)
Stack Loss
%02IN ~ %0~0uT
where K3 is a constant related to the type of fuel 9 Tl
and T2 are the temperature of the flue or exhaust gases
5 and a reference temperature, e.g. the ambient temperature
respecti~ely, and %02IN and %020UT
percentage oxygen concentrations of the air supplied
to the heating appliance, which may be nominally set
substantially equal to 21, and the exhaust gases~
Alternatively, the constituent gas, the
concentration o which is measured, may be C02, in which
case the stack loss may be calculated substantially as
follows:
1 ( 1 2)
Stack Loss = - ~C~ -
where Tl and T2 are as previously, Kl is again a
constant related to the type of fuel used, and ~/OC02
is that of the exhaust gases.
: Whichever stack loss formula is used, and
20 whether based on the %2 or ~/OC02 of the exhaust gases,
the following preferred formula may be used in determining
: the operating efficiency (~ ) of the heating appliance:
Efficiency - 100 - ~ + (Stack loss) ~ K4 ~P +(Tl -T2)~
where R and P are constants related to the type of heating
25 appliance and the moisture and hydrogen content of the
; exhaust gases, K3 and K4 are constants related to the
:::
- ~ . .

-6-
type of fuel used, and Tl and T2 are as previously
defined.
Preferably, therefore, the temperature sensor
may be adapted to produce an output signal related to
5 the value of (T~ - T2)for example, a thermocouple the
cold or reference junction of which in use is placed
in the ambient atmosphere, while the 'hot' junction
is placed in the flue gasesO Preferablyy a Type K alloy
thermocouple is used~
10Preferably, different values for the constants
K1, K3 and K4 as appropriate for different types
of fuels, such as solid fuel 9 fuel oil or natural gas,
are stored in the computation means, and the apparatus
: includes means for selecting which value of the constant
15 is ~o be used in the computation of the predetermined
formula.
To simplify operation of the apparatus and to
reduce the possibility of error, the computation means
is preferably operable to automatically perform individual
20 stages of the operating procedure from test measurements
for calibration purposes, through derivation of the
measurement values of the constituent gas concentration
and temperature from measurements taken of the exhaust
gases, to final computation of the stack loss and/or
25 operating efficiency, upon receiving instructions from
the operator, conveniently by means of switching or like
devices. Accordingly, the apparatus may also include means
for visually or audibly indicating when some or each of
.
:,

theseindividual stages has been successfully completed,
so that the appropriate switching device can be
actuated to enable the next operating stage to be
performed.
An embodiment of the invent~on will now be
described in greater detail; by way ofe~ample only,
with reference to the accompanying drawings, in which:
Figure 1 shows a block schematic diagram of
a fuel efficiency monitor in accordance with the
10 invention;
Figure 2 shows the internal architecture of a
microprocessor fonming part of the monitor of Figure l;
Figures 3(a~ and 3(b) are circuit diagrams
of two fonms of amplifier suitable for usP in the
15 monitor shown in Figure l; and
Figures 4 and 5 are flow charts of differQnt
parts ~ a programme which ~-ontrols the operation of the
apparatus shown in Fig~re 1.
Referring to Figure 1 the apparatus compr~es an
20 analogue module 1 having a first lnput terminal 2 for
receiving the output from an oxygen sensor 3 and a second
input tenminal 4 for receiving the output from a
temper~ture sensor 5. The analogue module 1 includes
respective Amplifiers for linearly amplifylng the input
25 volt~ge~ received at its two input terminal3 to
product full scale output voltages At re~pect~ve output
- tenminals 7, 8 each equal to the value of a re~erence
voltage, typically 3 volts, produced at a third ou~pu~
tenminal 9.
. ~
.

8-
The output tenninals 7, 8, 9 of the analogue module 1
are connected to respect~ve input tenmlnal~ design~ted
A/D 1 " A/D 2 and A/D REF o~ an ~bit mit~roprocessor chip
10 incorporating ~n internal 8-bit A/D converter h~ving
two input ch~nnels. The tenminals A/D 1, A/D 2 provide the
input connections for the two A/~ converter channels
while the voltage applied ~o the input termin~l A~D ~EF
from the ~nalogue module 1 determines the upper limit of the
conversion range. The microprocessor chip 10 u~ed
1~ in the pre~ent example i9 a commercial ly Avall~ble
\ componentD sold under the designation INTEL (~eglstered
Trade Mark) 8022 and manufactured by Intel Corpor~tion of Am-
: erica, U.S.A.It is particularly suited to the present
~pplication ~s the required two-channel A~D conver~ion
15 acility is built in obvi~ting the need for ex~ernal
A/D conver~icn of the output signals of ~he oxygen and
temperature sensors 3,5.
The microprocessor chip 10 also has three input~
output ~I/0) ports 123 13, 14 of 8 lines each. Three
20 of the I/0 line~ of the first port 12 are connected
to a three-way fuel select switch 16~ while another
three lines are connected to respective switches 17, 18~ /
19, for selecting an appropriate one o three differe~t
oper~ting progr~mmes stored in a read-only-memory (ROM)
25 having ~Gspaci~y of 2K (8-~lt) words contained
in ~he microprocessor chip 10.

The second port 13 has one of its I/O lines
connected to enable an oscillator 20 feeding a speaker
21 for sounding an audible warning tone when an oxygen
concentration measuremen~ h~s been t~ken, and another
5 line connected to operate the motor 22 of an air
~uction pump for drawing air into the oxygen sensvr 3
before an oxygen roncentration measurement ls to be t~enO
Another two of the I/O lines of the second port l3,
together with the 8 I/O lines of the third port 14
10 are connected to respective inputs of a displ~y module 24.
The display module 24 comprise~ three ~even-~egment
alpha-numeric displays 25~l 25b~ 25c driven by seven
of the I/O ~nes of the third I/O port 14 of the
microprocessor 10, ~ group of three indlcator lights
15 26A, 26b, 26c controlled by the eighth I/O line of p~r~ 14
for indic~ting which of three quantities, temper~ture~
% oxygen or efficiency, is being displayed on the alpha-
numeric displays 25a~ 25b, 25c and two further indicator
lights 27, 28 controlled by the two I/O line~ of port 13
20 connected to the display module 24, one 27 for indicating
when the oxygen sensor has been successfully calibrated
and the other 28 for indicating when an oxygen concentration
reading ha~ successfully been completed.
Other terminals on the micruprocessor chip
25 include power supply terminals 30, 31 across whi~h a
st~billsed ~5V supply 32 is connected, a pair of c~y~t~l
control tennin~ls 34, 35 ~cross which ~ timing element
36 i~ connected to control the timing o an lnternal
.
, .
- ,.

-10-
cryst~l-controlled oscillator and cloek ciroui~
built into the microprocessor chip 10~ In the present
example, the timing eleMent 36 consists of ~ 15k resi~tor
which sets the clocking period of the processor at 5
S Jus giving an instruction cycle time of 150~us
(30 clockln~ periods).
Figure 2 shows a block schematic diagr~m of the
intern~l archltecture of the microprocessor chip 10,
comprising a clock 40, the clocking period of which is
10 determined by the timing elernent 36 (Figure 1~, and
which controls the instructions cycle time of the m~ro-
processor via an 8-bit central processing unit (CPU)
41. The CPU 41 ~arries out v~rious arithmetic oper~tions
and controlq the operation o the remaining sections of the
15microprocexsor in ~ccordance with progr~mme instructions
stored in ~ read-only~memory (ROM) 42 mentioned earlier.
The microprocesQor also includes a dat~ memory 43 h~ving
a capacity of 64, 8~bit words, which can be ~cce~sed during
oper~tion both to have data written into it and to have
; 20data re~d from it; on 8 bit tlmer~event counter 44;
and a two-channel 8-bit A/D converter 45 which w~s
discussed earlier in connection with the analogue
module 1 (Fi~ure 1). The CPU 41, the progr~mme memory 42
the d~ta memory 43, the timer/event counter 44 and the
25 two channel AJD converter 45 are all interconnected
with one ~nother, and to the three I/0 ports 12, 13, 14
: by means of an internal bus network 47.
Referring again to Figure 1, the gnin of the
respective ~oplifiers in the analogue module 1 requlred
30 to bring the output signals of the oxygen sensor 3

~nd the temper~ture s~nsor 5 up to a f~lll scale value of
about 3 volts will depend upon the range of these
two output sign~ls. The requlred maximum range of
oxygen concentratlons to be me~sured by the oxygen
~ensor 3 in the present exarnple is from 0 to ~bout 22
(20.9X bein~ the nomlnal am~ient oxygen concentration
of air).
A p~rticularly suitable form of oxygen sensor
is th~t commercially available under the design~tion "C/S"
10 and manufactured by City Technology Limited, London,
England, which when loaded by a 47 ohm resistor gives
full scale output voltage of 47 mV for the ambient air
oxygen concentr~tion (20~9~/o)~ This form of oxygen sensor
operates on an electrolytic principle. The ~ensor is self-
powered, diffusion limited and consists basicAlly ofa met~l anode, electrolyte and an air cathode.
The diffusion of oxygen to the ~r cathode 1~ controlled
by a capillary diffus1On barrier~
:. The ~mplifier in the ~nalogue module 1 ~oci~ted
wlth the oxygen sensor may therefore h~ve ~ gain of ~bout
60 to produce a full scale output voltage of 3~olt~0
for an inpu~ signal of 50 mV representing a maximum
~; oxygen concen~ration of ~bout 22Z.
A preferred fonm of temperature ~en~or comprises a
"Type K" ~lloy thenmocouple as defined i~ Briti~h
Stand~rd ~S.No.4937 part 4:1973, the output of wh~ch
- v~ries with temper~ture at a rate of about 4.1 mV per 100 C
giving a full scale re~d~n~ of ~bout 41mV st 1000C,

-12-
the upper limit of the temperature range of
interest~ Again, an amplification factor of 60 would
be suitable or the output signal of this thermocouple,
giving a full scale output voltage of about 2.6
volts for a sensor output voltage of 45 mV.
To obtain an overall instrumental accuracy
of within 1% the amplifiers within the analogue module
which may be of a commercially available form, maintair.
an accuracy of better than ~ 1% over a working
10 temperature range of 20C, and any temperature drift is
compensated for automatically to eliminate the need for
re-calibration while in useO
Figures 3a and 3b show purely by way of example
two alternative suitable orms of D.C~ operational
15 amplifier circuitry for amplifying the outputs of the
sensors 3, 5. The circuit of Figure 3a uses a dual
transistor pre~amplifier Tl, TZ (prefera~ly both from th~
same chip) while Figure 3b uses an output diode Dl~ with
the input being ground-referenced.
The linearly amplified sensor output signals
from the analogue module 1 are then applied to respective
channels ~ the 8-bit A/D converter 45 of
the microprocessor 10 via the input terminals A/D 1, A/D 2.
Both channels of the A/D converter 45 are set to produce
25 a full scale digital reading of 256 when a m~ximum
input voltage of 3 volts, determined by the reference
..
.

-13
voltage applied to
tenminal A/D R~F is applied to it. Thus each increment
of the digital range corresponds to ~bout 12mV.
The result of the A/D conversion on either channel can
S be read from the A/D converter 45 via the internal bus
network 47 during the course of a programme,
.However, before these values can be used to provide
an accurate measurement, the temperature and oxygen
sensors 5, 3 must be calibrated sinceneither has an
10 output voltage which is linearly related to the qu~ntitles
they are used to measure. In both cases a correction is
performed on the digltal valu~ of the sensor signal.
The relationship between the output voltage of the
oxygen sensor and the actual oxygen level sampled
15 is exponential:
C ~ 1 - exp k ............. (I)
~- where C is the fractional oxygen concentration ~e.gØ209
i~ ambient air), S is the ou~put signal from the sen~ox
and k is ~ constant.
Expanding the exponential gives:
exp k ~ 1 ~ k ~ S 2 - S 3 ........ ,....... ~.... ,.. 7~
:~ and ~o C ~ S/k may be taken as a first approxim~tion,
k ( 2k )...... ~.......... ~.................... (III)
a~ ~ second approximation.
However, before the value of ~ the oxygen
concentra~Dn can be ~lculated .Erom one or other of these
approximations, the value of k, the callbration oonstant,
mu~t first be derived. This is done by taking an ini~iAl
: c~llbratlon reading of the oxygen ooncen~ration of
,.

a test gas, e.g. ambient air9 and assigning the resulting
A/D converted digital reading S', the nominal fractional
oxygen concentration of ambient air C' 0.209.k is then
calculated in accordance with the following equation,
derived from equation (I):
k = ~ = - ................ ~.... o~o~ (IV~
- 4.2651 x S' (for ~nbient air as the test gas)
Having determined a value f~r k rom the initial
calibration reading S', this value is stored and used
10 as the calibration constant in determining the actual
oxygen concentration using equation ~I), or rather one
of the simplified approximations given by equation
(II) or (III) (in this example Pquation (III) )
from a subsequent measurement.
The "Type K" allay thermocouple 5 produces an
output voltage which is substantially linear over most
of the range of interest i.e. 50 to 1000C~
Calibration of the thermocouple is required,
however, and this is conveniently achieved by
20 appropriate adjustment of the offset voltage of the
associated amplifier in the analogue module prior to taking
a flue measurement, based on a reading of the ambient
temperature at which the thermocouple output voltage
must be zero since both junctions are at the same (ambient)
25 temperature.
When taking a flue measurement, the thermocouple will
automatically record the difference in temperature between
the flue gases and ambient,

-15-
Alternatively, the microprocessor 10 may be arranged
to store a value representative of the thermocouple
output signal produced by the ambient temperature
test measurement and to subtract this re~erence value
from the appropriate value produced by the thermocouple
on measurement of the exhaust gases~
The measurement junction of thermocouple 5 and
the oxygen sensor 3 are both housed in hollow cavity
on a common probe~ A small motor-driven pump 22 is
actuated by the microprocessor 10 to draw air into the
cavity when an oxygen reading is to be taken.
It can be shown (British Standard BS. No.845:1972)
that the heat loss in dry flue gases based on the nett
calorific value (stack loss) is given by:
15 Stack loss= ~ ~-] ......... ~..................... (V)
where Kl is a constant dependent upon the type of fuel,
Tl is the flue gas temperature, T2 is the ambient temperature.
i The ~/~C02 is given by the formula:
%2
2 (1 21 ) x K2 ..................................... (VI)
where K2 is the maximum theoretical % C02 value. Substitu-
ting equation (VI)into equation (V) and defining
K3 , 1 gives
K2
Stack loss = r 3( 1 T2~ ......... ~ .. ~ .. o ......... (VII)
21 ~ ~/2~
The di~ference between the flue temperature and ambient
(Tl - T2) is given automatically by the thermocouple

- 1 6 -
reading ~ince its cold or reference junction ~8
at ~mbient, and the %2 value i9 given by the oxyge~ ~ensor
reading. Inserting thes e values into equation (VII), i~ used
to give the stack loss measurement, using the follswing
S val~es of K1, K2 ~nd K3 for solid fuel, for fuel
oil ~nd natural gas;
K1 K2 K3 K3 (rounde~)
Sol~d ~uel 0.65 18.5% 0.73784 0.738
Fuel Oil 0~56 15.5% 0.75871 0.759
10 N~tural Ga~ 0~3811.870 0.67~27 0,676
The above corrections of the A/D converter readings
of the thermocouple ~nd oxygen sensor outputs ~re
applied automatically within the microprocessor 10. In
the case o~ ~he thenmocouple reading thi~ simply compri8e8
multiplying the A~D converter reading by ~ con~tant a~tor
of 4. 835, while in the case of the oxygen sensor" an
initi~l calibration me~surement (S ' ) i~ carried out OT~
~mbient air to determine the value of the c~libr~tion
csn~tant k in accordance with equation (IY)
. ~0 (k o 4. 2651 x S ' ) and then thls value of the const~nt
is used to calculate the oxygen concentration of th~ f lue
gases from ~ subsequent measurement S using equation (III)
`~ a~ explained ~bove. Using these values ~he stack lo~
then calculated in ~ccordance with equ~tio~ (VII)y or the
efficiency derived from the stack loss, selecting an
appropriate value for the cons~ant K3 dependent upon the
type of fuel, the three diferent K values being stored
within the ROM 42 of the microprocessor 10.
The cali~ratlon and measurement proced~re
o the ~uel eficiency monitor will now be de~crib0d,

The main programme is illustrated by the flow diagram in
Figure 4, consisting of a 'calibration' routine, an
'operation' routine and a stack loss (efficiency) calulation.
In additior~, there is an interrupt
pxogramme illustra~ed by ~he flow chart in Figure 5
which ~un8 continually with a cycle time of 2 m~, serving
to update the display, keep track o time and to change the
contents of the displAy to a required v~lue e.g. 2
concentration or ~t~ck loss v~lue, whenever the di~play
10 select switch 19 is oper~ted.
Referring first to Figure 4, when the
; monitor is switched on the alph~-numeric di~play 25
~nd ~he indicator lights 27, 28 rem~in off. The
.: operator then put~ the probe containing the thermocouple
15 7 hot' ~unction 5 and the oxygen sensor 3 into the
ambient atmosphere and oper~tes the callbrate switch 17. Th~
initiate~ the "calibrate" routine a~ the maln pro~r~me
outlined ~n broken lines, which switches on the pump
motor 22, wait~ for 5 second~ and reads the conver~lon resu~t
20 on ch~nnel AJD 1 vf the A/D converter 45. If thi~
converter reading lies between 220 and 250, correspo~di~g
to a 20 to 22% oxygen concentration, the motor 22 is ~witched
off. If no~ fur~her readings are taken until ~ r~ding of
between 220 and 250 i obtained whereupon the motor 22
251g ~witched of.
.
. . .

-18-
Thls re~ding (S') i8 then tak~n ~s correspond~ g
to the nominal ambient 2 concentration of 20~9~
~nd used to calculate the calibration const~nt k in
accordance with equ~tlon (IV) i.e. by multiplying lt by
4.265. The calibration con~tant thus calculated 1~
then temporarily stored in the RAM 43 and the c~libr~tion
indlcator light 27 s~tched ON . This indicate~ ~o the
operator that the calibration of the 2 qensor has been
completed9 tha~ the ~ensor probe should be lnserted
in~o the flue, and tha~ the "oper~te" swltch 1~ ~hould be
actuated. This initlates ~he 'oper~te' routine of the
programme which starts with a lS second wait
after which the conversion result app~aring on channel
A4D 2 of the converter 45 is read. A second re~ding of
conver~er channel A/D 2 is taken 5 second3 later ~nd
compared with the first. If the two values differ by 2 or
le~s~ corre~ponding to a temperature difference of ab~ut
10C or less, this temperature ~alue i9 recorded a~d the
pump motor 22 is switched on in prepar~tion for the
oxygen concentr~tion mea~urement of the flue g~3e~.
If not, the process is repeated eYery 5 ~econd~ u~til
the difference between two successive re~ding-~ 18 le~ than
. or equal to 2.
; After the sensor probe pump motor 22 h~
been swi~ched on there i5 a delay of 5 ~econd~ and th~
reading (S) on A/D 1 is taken. The "reading tak03~
i~dic~tor light 28 is then Rwitched ont and follo~n$
further del~y of 10 seconds ~n audible w~rnlng i3 sou~ded
by actuation o~ the oscill~tor 20 which power8 the
. .
.

5~
,9
speaker ~ to in~icate th~t the probe can be withdrawn
f~om the flue. The p~np motor 22 i9 then switched
of.
The values read rom She ~wo channel3 of
the A/D converter dur;ing the above-described ~ectio~ of
the programme merely provide a digital representatlon
of the value of the output signal from the respe~tive
sensor~ 3, 5O A9 described earlier, these value~ require
correction to conver~ ~hem to actual oxygen
concentr~tion and temper~ture values. In order to do
thi~ the uncorrected oxygen concentration v~lue from
; ch~nnel A/D 1 of the A~D converter 45 is divided by
the oxygen sensor cal~ration constant k detenm~ned earlier9
; to derive a value X 3 S/k (where S i8 the A/D
conversion result of the flue oxygen concentr~tlon
in equation (I), (II) and (III. This is then u~ed
; in equatlon (III~ to calculate the corrected oxygen
concentration value C as follows:
C ~ X( 1 ~ x /2) where hC ~ Stk
This v~lue of C is thus the actu~l fractional oxygen concen-
tr~tion of the flue gases; and is tempor~rily stored while
: the uncorrected temperature value from ch~nnel A/D 2 of the
converter 45 is corrected by multiplying it by ~ flxed
correction factor of 4.835 to give a digithl value
: 25 equal to the difference in temper~ture be~ween the
flue g~e8 and the ~bient temper~ture, After a delay of
10 5econds the audible warning produced by the ~pe~ker
21 i8 switched off, indic~ting th~t the fuel selec~
... . .
.

-20-
~witch 16 should be set to the appropriate fuel to initiate
calculation of the efficiency or stack loss value. This
determine~ w~ich of the three dlfferent value~ for the con-
stant K3 (it is 0.738 for solid fuel; 0.759 or fuel
oil; 0~676 or natural gas) should be used with
~he correc~ed temperature value (Tl-T2) and the
corrected 2 concentration(C) in the calculation of equ~t-
ion (VII)~ The warning light 27 is then switched on and
the reading taken llght 28 switched off. ~urther flue
ga~ readings can then be taken e.g. from different
section~ of the flue b~ repeated operation of the 'oper~te'
~witch 18.
Once the corrected temperature, oxygen
concentration and s~ck lo~s values have ~een determined~ -
these values ~re stored in respective buffers Bl,B2 ~nd
~3 (not shown) ~nd are made available or display on the
three~even-~egment alphA-numeric di3plays 25a, 25b, 25c
of the displ~y module 24. This facility i~ provided
by ~he 'lnterrupt' programme 9 the flow di~gram of which
i~ lllustrated in Figure 5, which is controlled by the
"di~pl~y select" switch 19.
Thi9 routine is cyclically repeated at regul~r
intervals, 2mS in the present ex~mple, and ~t the
beginning of each cycle an internal 8-bit counter is
25 incremented by a count of 1. A complete cycle of the
counter thu~ takes 256 x ~ ms ~ 0.512 seconda, and ~his
i~ u~ed as ~ clock to control the timing of the various
delay~ which occur in the opera~ion of the main
progral~ne .
- ':

~ 5 ~
The apparatus described so far is adapted to
automatically compute the heat loss or stack loss of the
exhaust gases of the heating appliance. Wowever, as
discussed earlier it may readlly be rnodified to
5 alternatively or additionally display an indication
of the percentage operating efficiency (~_) of the heating
appliance from the derived measurement values of %2
and (Tl-T~) already described.
Calculation of ~_ based on the gross calorific
10 value of the fuel may be achieved by programming the micro-
processor 10 to perform the following calculation:
[ 3( 1 T2) +K4 ~P+(Tl-T2)~ ............. (VIII)
where R is a constant related to the radiation losses of
the heating appliance (typically between 3% and 10~/o)
15 K4 ls a constant related to ~he type of fuel used, P is
a constant related to the hydrogen and moisture content
of the air applied to the appliance (in this example nominally
set at 1121.4) and K3, Tl and T2 are as previously defined.
The second term in the square brackets will be
recognised as the stack loss from equation (VII)
which value may simply be substituted into the equation,
if previou~y calculated. Again, different values for the
constant K4 in addition to those for K3 may be stored
and selected upon operation o the fuel select switch 16,
together with a range of values for P which may be selected
according to the type of heating appliance. Typical values
for K4 are as follows:

Solid fuel 0 00409
Fuel Oil 0 00512
Natural Gas 0 00828
Where it is required to measure the stack loss on the
5 basis of the ~/0CO2 concentration of the exhaust gases,
using a C02 sensor in place of the 2 sensor described,
equation (V) may be used instead of equation (VII).
Different values of the constant Kl as previously
listed for different types of fuel may then be stored
10 and selected by operation of the fuel select switch 16.
Of cour~e, the calibration and linearisation procedure
requires modification to take accoun~ o the different
relationship in the variation of the C02 sensor output
signal with the measured quantity, but similar principals
15 to those described above for calibration of the 2
sensor may be employed.
Once the stack loss has been calculated, this value
may be substituted ~s the second term in the square brackets
of equation (VIII) above to determine the percentage
20 efficiency ( ~

~L~
- 2 3 -
In ~dditiol- to this, the int~rrupt programme
monitors the input line of port 12 to which the display
select switch 19 is connected to determine whether it
has been operated since the preceding cycle, and if it has
5 it replaces the inform~tion on the display with informati~
drawn from a different one of the three buffers, B0, Bl,
B2 selected by different values of x = 0,1 or ~. For example,
if the display is showing the temperature value contained
in buffer B0, then operation of the display select switch
lQlg will change x= 0 to x = 1 causing this displayed information
to be replaced ~v that from buffer Bl, i.e. the corrected
oxygen concentratiQn value. If the display select switch
is pressed again, ~ changes from x=l to x = 2 the display
25 will be refreshed with information drawn from buffer B2,
15i.e, the stack loss value. Further operation of the dis~lay
select switch will change from ~ - 2 to ~-0, returning
the original temperature value from buffer B0.
Simultaneously, with the displayed information
a respective one of the threeindicator lights 26a, 26b~ 26c
20will light up to indicate which piece of information
is currently on display.
The purpose of comparing successive readings
of the temperature sensor output signals appearing on
channel A/D2 of the A/D converter 45 at 5-second intervals,
25until successive readingsdiffer by less than a predetermined
amount, is to ensure that the temperature of the temperature
sensor has risen and settled to the actual temperature of
the flue gases,

It wi31 be appreciated that apparatus specifically
described is only a preferred embodiment, and many
modifications may be madP to ik without departing from
the scope of the present invention, For example9 diff- -
erent forms of temperature and oxygen or C02 sensor
may be used, requiring modifi ation, inter alia to the
methods used in their calibration and ln the
linearisation of their output signals. Both sensors may
have non-~inear response, requiring a non~linear
correction also to be applied to the temperature sensor
output signal. Further, the invention is not restricted
to the particular form of microprocesssor used, other
forms of microprocessor may equally be used,
whether or not they incorporate built-in A/D conversion
means. External A/D conversion means may readily be
employed where necessary. The range of fuels for which
the apparatus is intended may be varied as appropriate
or extended to include other forms of fuel by
appropriate selection of the constant K3 and K4
in eguations (VII) and (VIII).
,~ ~