24 June 2026 · 12 min read

Where your CEMS goes: siting, sample ports and the measurement plane DOE accepts

Buying a CEMS — Part 4 of 5. A practical series for industry, drawn from the DOE CEMS Guidelines (Version 8, 2025), which we helped develop.

In Part 3 we matched the sampling method to your flue gas. But a perfectly chosen system bolted to the wrong spot on the stack will still give you bad data — and, worse, can fail its calibration through no fault of the analyser. This part is about the one decision you can’t quietly fix later: where the system physically goes.

Siting has to be designed in before the stack steel is fabricated, because it dictates where the sampling ports, the platform and the cabling all land. Get it onto the drawings early and it costs nothing; discover it after the stack is up and you’re cutting new ports at height, or arguing with DOE about whether your data is representative.

The three things every location must satisfy

The Guidelines open Chapter 4 with three non-negotiable principles for any CEMS install (§4.2.1):

  • The measurement must be representative of the actual stack emission. (Dilution to hit a number is prohibited under Reg 14 of CAR 2014.)
  • The effluent gases must be well mixed at the point of measurement.
  • The location must be accessible for maintenance, repairs and calibration.

Everything else in this article is really just the practical detail of how you prove those three things to the regulator.

The straight-length rule

The single most important siting number is how far the measurement plane sits from any flow disturbance — a bend, a fan, a junction, a damper, an expansion. Right after a disturbance the gas is swirling and stratified: pockets of high and low concentration that haven’t blended yet. Measure there and your reading depends on exactly where in the swirl you happened to sample. Give the flow enough straight duct to settle and it becomes uniform — well mixed, as principle two requires.

The Guidelines set the target as a run of straight duct measured in equivalent duct diameters (D) — the hydraulic diameter of the duct. The working requirement (§4.2.2.1) is a plane at least 5 D downstream of the disturbance behind it and at least 2 D upstream of the next disturbance ahead of it, within an overall ideal of 5 D either side.

The 5D / 2D straight-length rule for a CEMS measurement plane A horizontal duct with a flow disturbance at each end. The measurement plane sits at least five duct diameters downstream of the upstream disturbance and at least two duct diameters upstream of the next disturbance. upstream disturbance next disturbance flow D duct dia. measurement plane sampling port ≥ 5 D downstream of disturbance ≥ 2 D before the next
The straight-length rule (Guidelines V8 §4.2.2.1): give the gas at least five duct diameters to settle after a disturbance, and keep at least two before the next one, so the plane sees well-mixed flow.

The rule is about flow, not orientation. The Guidelines define upstream and downstream by the direction the gas travels — from the stack inlet toward the sampling port, then from the port toward the outlet — so the same 5 D / 2 D applies whether the gas runs along a horizontal duct or straight up a vertical stack. The vertical chimney is in fact the more common case:

The 5D / 2D straight-length rule on a vertical stack A vertical stack with gas flowing upward, a disturbance at the inlet below and the outlet above. The measurement plane sits at least five duct diameters above the inlet disturbance and at least two duct diameters below the outlet. upstream disturbance (inlet · bend) next disturbance (outlet) flow D duct dia. measurement plane sampling port ≥ 5 D downstream of disturbance ≥ 2 D before the next
The same rule on a vertical stack (Guidelines V8 §4.2.2.1). Upstream and downstream follow the gas — up from the inlet, on toward the outlet — so the 5 D / 2 D spacing is identical to the horizontal case.

The same plane serves the CEMS and the manual stack test — under §4.2.8 the Standard Reference Method location is prepared to the same criteria, because the SRM is what your CEMS is calibrated against. One well-chosen plane does both jobs.

Where exactly in the cross-section

Straight-length gets you a good plane; now the question is where in that plane each measurement sits. There are three answers, depending on what’s measuring.

Where the analyser and the SRM sit in the duct cross-section Three duct cross-sections. A point CEMS measures at one spot at least one metre from the wall or over the centroidal area. A path CEMS sends a double-pass beam across the duct, at least 70 percent of it within the inner half of the area. The manual SRM traverses points placed at the centres of equal areas. Point CEMS one point — ≥ 1 m from the wall, or over the centroidal area Path CEMS beam across — double-pass, ≥ 70% within the inner 50% of area SRM traverse manual method — points at the centres of equal areas
Three things share the plane: the continuous analyser as a point or a path (Guidelines V8 §4.2.6–4.2.7), and the manual SRM's grid of equal-area traverse points (ISO 9096:2017, Table 1). The bigger the duct, the more traverse points the SRM needs.

A point CEMS measures at a single spot, which must be at least 1.0 m from the duct wall or sit over the centroidal area (§4.2.6). A path CEMS shoots a beam across the duct — and remember from Part 3 that DOE accepts double-pass only; the path must lie within 1 m of the wall, with at least 70% of it inside the inner 50% of the cross-section (§4.2.7). The manual SRM, meanwhile, doesn’t measure at one place at all: it traverses a grid of points set at the centres of equal areas, and ISO 9096 scales that grid with duct size — a small duct may need a handful of points, a duct over 2 m needs seventeen.

That last detail is why the plane has to be genuinely well mixed: the CEMS reads one point or one line, but it has to agree with the SRM’s full-section average at QAL2. A stratified plane is exactly where that agreement breaks down.

Round stack or square duct?

The shape of your duct changes how that traverse grid is laid out — and therefore how many ports and how much platform you’ll need to budget for. ISO 9096 divides a circular duct into equal-area rings, so the points sit on two perpendicular diameters and bunch toward the wall (Table 1: 5, 9, 13 or 17 points as the duct widens). A rectangular or square duct is instead split into a grid of equal rectangles with a point at the centre of each — 4, 9 or 16 points as the plane area grows (Table 2).

Traverse points in a circular versus a rectangular duct A circular duct divides into equal-area rings with sampling points on two diameters that cluster toward the wall. A rectangular or square duct divides into an even grid of equal rectangles with a point at the centre of each cell. Circular duct equal-area rings — points sit on diameters, clustering toward the wall Rectangular / square duct equal grid — a point at the centre of each cell (here 3 × 3 = 9)
Two ways to grid the same plane: equal-area rings for circular ducts (ISO 9096:2017, Table 1) and an even grid for rectangular or square ducts (Table 2 / Annex B). Both place a point at the centre of an equal area — the geometry just differs.

The DOE straight-length rule doesn’t change with shape: for a non-circular duct you compute the hydraulic diameter (D_H) and apply the same 5 D / 2 D spacing. The buyer’s takeaway is modest but real — a larger or rectangular stack needs more sampling ports and a bigger working platform, so it pays to know the shape before you price the access. The full traverse mechanics are the tester’s domain; we go deeper on the standards behind them in the reference-methods series.

The other location rules

Straight-length and cross-section position are the big two, but §4.2.3 adds a checklist of placement rules that catch people out:

  • Downstream of the SRM port — the CEMS sits after the manual sampling location, not before it.
  • Downstream of all abatement — measure what actually leaves the stack, after the bag filter, scrubber or ESP.
  • Downstream of any merging ducts — so you’re not reading one stream before it joins another.
  • One CEMS per stack — each system monitors a single stack or chimney.
  • PM above gas — locate the particulate CEMS above the gas CEMS where both share a stack.
  • As far from the final exhaust as practicable — away from the influence of the open top.

When the stack won’t give you the straight run

Plenty of existing stacks simply don’t have 5 D of clean duct to offer. The Guidelines don’t leave you stranded — §4.2.4 gives two routes for a stack that can’t meet the standard criteria:

  • The USEPA geometry — a plane at 8 D downstream and 2 D upstream of disturbances for dust and gas (or 2 D downstream, ½ D upstream for gas only), available to stacks that hold a Written Approval and meet the US EPA sampling-location criteria.
  • Prove the flow is good anyway — run a homogeneity or stratification test and pass it with no stratification found, using a recognised procedure (EN 15259 for homogeneity, or US EPA Chapter 2, “Bias Due to Probe Location and Stratification,” for stratification).

That second route is the important one for retrofits: if you can demonstrate the gas is well mixed at your available plane, the exact diameters matter less. It’s also why the homogeneity assessment shows up on DOE’s siting paperwork.

What DOE actually asks for

When you submit the install for approval, the siting isn’t described in prose — it goes onto a defined data sheet (Guidelines Vol II, §2.3.1–2.3.3). Knowing the shape of that form tells you exactly what to nail down at the design stage:

The siting datasheet DOE expects

AStack / chimney characteristicShape, internal diameter and stack height — the geometry the diameter rule is measured against.
BDistance from disturbancesLength of the sampling location from the downstream and upstream disturbance — your 5 D / 2 D evidence.
CSample portsPort size, number used and orientation — separately for particulate and gases.
DSample points & filtrationNumber of points / lines across the plane, and where filtration sits in the train.
EPlatform & safe accessA permanent, strong platform with ladders, staircase or elevator, weather protection, tools and spares on site, and reference gas injectable at the probe (§4.2.5, §4.2.9, §2.3.2).
FHomogeneity / stratificationThe assessment report, attached — your proof the plane is well mixed (§2.3.3).
What the DOE siting submission asks for (Guidelines V8, Vol II §2.3.1–2.3.3). Most of it is fixed the moment the stack drawings are signed off.

The platform line (row E) is the one buyers most often under-budget. A CEMS needs a permanent, strong, safely-accessed platform at the probe — not a temporary scaffold — because every future calibration, audit and repair happens there. Designing it in is cheap; bolting it on later is not.

Why it matters to you

Siting is the part of a CEMS project with the longest tail. The analyser you can swap; a badly placed plane you live with for the life of the stack. A plane too close to a bend may read fine day-to-day and then fail QAL2 or the annual surveillance test, because the one spot your CEMS watches doesn’t match the SRM’s full-section average — and as we noted across the reference-methods series, that’s the agreement your compliance data ultimately rests on.

So before you sign off a stack design or accept an installer’s proposal, make sure someone has checked: the 5 D / 2 D straight run (or a passed homogeneity test), the cross-section position for your point or path analyser, the ports and traverse the SRM will need, and a permanent platform with safe access. Get those onto the drawings now and the rest of the project — QAL2, and years of routine testing — has somewhere proper to happen. Siting is the last of the build decisions; Part 5 pulls the whole purchase together — choosing a partner, the true cost, and the order that gets you to a compliant report.

Designing a new stack, or unsure an existing one can be sited compliantly? Talk to us — we assess sampling planes, run the homogeneity tests and prepare the DOE siting submission, using the same DOE CEMS Guidelines we helped write.

Related insights

This article is general guidance, not legal advice. For obligations specific to your facility, refer to the current Environmental Quality (Clean Air) Regulations 2014, the EQA 1974, the DOE CEMS Guidelines, and the current editions of MS 1596, ISO 9096 and EN 13284, or speak with us directly.

← All insights