A furnace monitoring RFQ that fails to lock the gas composition, full-scale in Nm3/h, and straight-run geometry to the thermal mass flowmeter datasheet will arrive on site reading ±10–15% off, because the device's heat-transfer output scales with specific heat capacity Cp and requires a fully developed velocity profile [S1][S2].
Thermal mass flowmeters measure mass flow directly via convective heat loss from a heated RTD, with no P/T compensation required, and they hold a 100:1 turndown versus 20:1 for vortex and 5:1 for orifice meters, which is exactly why they dominate combustion-air and flue-gas loops where turndown spans the cold-start to full-fire range [S2].
Gas Composition and Cp Lock-In
Every thermal mass flowmeter RFQ line must open with the actual gas or gas mixture, because the meter equation m = Kq/(Cp·ΔT) ties output to specific heat capacity at constant pressure, and a 10% composition shift in a 60/40 CH4-CO2 biogas stream alone introduces 3–5% measurement error on an uncorrected meter [S3]. Common Cp values you will write on the RFQ: methane 2.22 J/g·K, nitrogen 1.04 J/g·K, carbon dioxide 0.85 J/g·K, argon 0.52 J/g·K, air 1.005 J/g·K, and for furnace combustion-air trim the gas is typically air plus 8–12% moisture by volume [S3][S4]. If the loop runs a mixed flue gas, request a gas correction factor (GCF) pre-loaded at the factory rather than a field entry, since most vendors calibrate on N2 or air and then apply the GCF in firmware [S4].
For pure gases the GCF table is short: H2 reads 1.004, He 1.4066, CO 0.994, CO2 0.7326, CH4 0.7147 against air = 1.000, so write the target gas and the Cp value directly on the RFQ and reject any quote that says only "calibrated for air" without a GCF entry [S4].
Flow Range, Line Size, and Straight-Run Geometry
Write the full-scale flow in Nm3/h (normal cubic meters per hour at 0°C, 1 atm), the minimum operating flow, and the line size DN, then the upstream and downstream straight-run lengths, because thermal dispersion sensors are intolerant of swirl and pulsation [S2][S3]. A typical combustion-air RFQ reads 600 Nm3/h full scale, DN80, 20×DN upstream and 5×DN downstream of any elbow, valve, or damper; the OEM datasheet also states a 0.1–120 Nm/s velocity operating window, which sets the lower bound of the turndown [S2]. Skipping the straight-run line is the single most common reason a $2,000 meter reads ±15% out of the box, because turbulence right at the heated sensor distorts the King's Law relationship Q = K·√(ρ·V)·ΔT [S2].
Insertion-style probes need a 1.5×DN ball-valve retraction assembly for hot-tap furnace duct installs, while inline (flow-through) bodies need the full 20×DN/5×DN spool and are limited to smaller line sizes (DN15–DN80 typical) where the meter coefficient stays stable [S1][S3].
Calibration: NIST-Traceable, Actual Gas, Not Air Equivalency
State on the RFQ: "Calibration NIST-traceable on actual process gas, with certificate listing Cp, density, and the test points in Nm3/h," because an air-equivalency calibration on a CO2 or CH4 stream can deliver up to ±100% error once the GCF is applied at the upper end of the range [S2]. A correctly specified thermal meter for a biogas or combustion application runs reliably for 5–10 years with annual re-checks, while a misspecified unit returns inside 90 days [S3]. Specify the calibration test points at minimum, mid, and full scale (typically 10%, 50%, 100% of full scale) and require the uncertainty statement in % of reading, not % of full scale, since %FS hides turndown degradation [S2][S3].
Output Protocol and Integration with the Furnace Control Loop
Most furnace monitoring loops accept 4-20 mA analog plus HART 7 for trim, with the HART layer carrying the secondary diagnostics (sensor temperature, flow tube ΔT, calibration date) to the DCS or burner management system [S1][S2]. The heated sensor pair, usually two Pt100 RTDs, drives a constant-ΔT control loop, and the HART variable can also expose the meter's internal Cp and density entries so the GCF change is auditable from the control room [S1]. Avoid specifying FOUNDATION Fieldbus or PROFIBUS PA on a thermal meter without confirming the OEM supports the digital stack, because most thermal mass meters ship native HART or Modbus RTU over RS-485, with fieldbus as a cost-up option that adds lead time [S1].
Where a HART-equipped thermal mass flowmeter is feeding a burner damper actuator, request burst mode enabled and the PV mapped to mass flow in Nm3/h, not volumetric, so the DCS trend matches the billing units on the gas meter upstream [S1][S2].
Hazardous Area, Materials, and Process Connection
Furnace monitoring loops sit in ATEX/IECEx Zone 1 or Zone 2 for combustion enclosures, and in NEC Class I Div 2 for North American sites, so the RFQ line must state the exact zone and gas group; the meter body and electronics must carry a dual-certified label or the inspector will red-line the package [S1][S2]. Wetted materials on insertion probes are typically 316L stainless with Hastelloy or Monel optional for SO2 or HCl-bearing flue streams, and the process connection is usually a 1.5" or 2" NPT compression fitting or ANSI 150# flanged ball-valve retractor [S1]. Operating temperature on furnace inlets routinely reaches 200–450°C, so specify the sensor thermowell rating and the electronics ambient limit (typically 70°C) separately, because remote-mount electronics with a 5 m or 10 m interconnecting cable are the only viable option above 200°C [S1].
Verification: What a Clean RFQ Looks Like vs a Requote Trigger
Compare the two RFQ styles side by side, because the difference is whether the quote lands in three days or three weeks: [S3]
1. Gas and Cp. Clean: "Combustion air, plus 10% H2O by volume, Cp = 1.05 J/g·K, GCF to be confirmed in firmware." Requote trigger: "Air, standard."
2. Flow range. Clean: "Full scale 600 Nm3/h, minimum 6 Nm3/h, DN80 line, 1600 mm upstream straight run, 400 mm downstream." Requote trigger: "Furnace air, medium flow."
3. Calibration. Clean: "NIST-traceable on actual gas mixture, ±1% of reading, certificate with Cp and density entries, test points at 10/50/100% FS." Requote trigger: "Calibrated, NIST."
4. Output. Clean: "Dual 4-20 mA + HART 7, PV = mass flow Nm3/h, burst mode enabled, remote mount 5 m." Requote trigger: "4-20 mA output."
5. Area and body. Clean: "ATEX II 2G Ex db IIC T4 Zone 1, 316L wetted, ANSI 150# flanged ball-valve retractor, sensor rated 450°C, electronics remote 70°C ambient." Requote trigger: "Explosion-proof, hot tap." [S1][S2][S3]
Common Specification Errors That Force Requote Cycles
The five patterns that generate callbacks every quarter: (a) the gas is listed as "natural gas" with no methane number, pressure, or temperature, which makes GCF selection impossible; (b) full scale is given in mass units (kg/h) without a reference pressure, so the OEM cannot confirm the meter sits in its 0.1–120 Nm/s window; (c) straight run is omitted and the vendor must assume, then prices in a flow conditioner that the buyer did not budget; (d) output is listed as "HART or fieldbus" without picking one, and the fieldbus variant carries a 4–6 week lead time on most lines; (e) calibration is listed as "factory standard" which on most thermal meters means air at three points, not actual gas at five points [S2][S3]. Each of these gaps costs one extra round-trip email and, on a typical 30-day project, can push the commissioning date by 2–3 weeks [S3].
Total Cost of Ownership: Why Spec Quality Pays Back
A correctly specified thermal mass flowmeter loop on a furnace combustion-air or flue-gas line delivers ±1–2% of reading accuracy over a 5–10 year service life with annual NIST re-checks, while a loop shipped with vague specs typically returns inside 90 days and consumes engineering hours on root-cause analysis that exceed the instrument's purchase price [S2][S3]. On combustion-control loops the meter directly drives the burner trim, so a 3% measurement bias translates into 3% excess fuel or excess air, which on a 10 MW furnace is roughly 300 kW of heat either wasted or unrecovered, equivalent to a measurable line item on the next fuel bill [S2]. Locking the gas, flow range, straight-run, calibration, and area classification on the first RFQ is therefore not paperwork overhead; it is the cheapest control point in the loop.
Trackable signals to watch on the next RFQ revision: (1) vendor moves to dual certification under ATEX 2014/34/EU and IECEx schemes on the same meter body, reducing the spares list; (2) GCF entries exposed as HART variables so the DCS can flag composition drift against a fuel-air ratio target; (3) HART 7 burst mode carrying the meter's last calibration date, which lets the asset-management system schedule re-checks automatically; (4) IEC 61508 SIL 2 capable variants becoming standard on combustion loops, which removes a separate safety-instrumented flow transmitter in some architectures [S1][S2].
For component-level specifications, see loop calibrator, and loop tester.
This topic is covered further in Slewing Bearing Selection for Automotive Production Lines.