Fitting a Radon Sump and Fan to an Existing Slab

The result is back and somebody has already bought sealant

A detector sat in the basement for three days over the winter, the laboratory returned a figure north of the action level, and by the time anyone with a core drill is involved the owner has been to a merchant and come home with a case of cartridges. The cracks get filled, the joint between slab and wall gets raked out and gunned, the hatch over the old sump gets screwed down, and the retest comes back within noise of the first. Nobody has done anything foolish. They have simply treated the slab as though it were a barrier.

It is worth being clear with them about what the number they are reacting to actually is. An action level is a policy trigger, not a safety cliff: the U.S. EPA recommends fixing at or above 4 pCi/L and considering it between 2 and 4, the United Kingdom works to an Action Level of 200 Bq/m³ with a Target Level of 100, and the WHO Handbook on Indoor Radon proposes a reference level of 100 Bq/m³ rising to no more than 300 where that is not achievable. The figures differ because the policy differs, not because the physics does, and none of them says anything about where the gas is getting into this particular house.

Back to the slab, then, because it is not a barrier and was never asked to be. Soil gas does not diffuse through a floor in any quantity that matters; it is drawn through it, because the house sits at a lower pressure than the ground beneath it for most of the heating season. Stack effect, a flue in operation, a tumble dryer, wind on the elevation — each subtracts from the pressure in the lowest storey, and the ground makes up the difference through a hairline shrinkage crack, the annulus around a soil stack, the gap where the slab was poured up to the wall, the open cores of a block foundation. Closing nine tenths of that entry area does not reduce the flow by nine tenths, because the same pressure difference drives more air through what is left.

Active sub-slab depressurisation gives up on sealing as a strategy and attacks the pressure instead. A fan draws continuously from a void under the floor and holds the material beneath the slab slightly below the pressure in the room, so every crack that used to be an inlet becomes an outlet: a trickle of indoor air leaving downward instead of soil gas arriving upward. Nothing is filtered and the concentration in the ground is very nearly what it always was — the gas now simply has an easier route to the roof than into the living room. Every decision that follows, from the number of suction points to where the discharge goes, is a decision about whether that reversal actually reaches the far corner of the floor.

The slab is a lid; the diagnosis is about what is under it

Everything expensive about a mitigation job is decided by material nobody can see. Over 100 mm of clean, single-sized aggregate, one suction point will hold the whole floor: the void between the stones is continuous, and a fan pulls a pressure field across it much as it would pull along a duct. A slab cast straight onto compacted clay, onto reclaimed hardcore fines, or onto sand that has since silted up is a different job entirely — the fan may hold a two-metre radius and nothing beyond it.

The way to find out is a communication test, and it is the single most valuable hour on the job. Core a test hole through the slab, seal a vacuum source into it, and drill small test holes at increasing distance: near, mid-floor, then at the far corners and behind every interior wall. With the vacuum running, read the pressure difference at each with a micromanometer.

ANSI/AARST SGM-SF, Soil Gas Mitigation Standards for Existing Homes, frames the requirement as measurable negative pressure at the extremities of the slab rather than a single number every house must hit. That is deliberate: what counts as measurable depends on an instrument resolving a fraction of a pascal, and practitioners generally report something around a pascal at the furthest point as the sign a field has genuinely arrived. Zero at the far corner with the vacuum flat out is not a marginal result. It is the answer, and it means a second point.

Read the flow as well as the pressure, because the two together name the soil. High flow and low suction is open aggregate. Low flow and high suction is tight material with a small void. High flow and no pressure at any remote hole means the vacuum has found a route to outdoors — an untrapped floor drain, a void in the footing, a slab edge that never met the wall — and is short-circuiting rather than depressurising. That third result is the one most often misread as good news.

How many points, and which quadrant each one owns

Suction points are not allocated by floor area. They are allocated by what the diagnostic showed, and the thing that most often forces a second one is not distance but an obstruction. A strip footing under an interior bearing wall is a dam cast down into the sub-slab material; a thickened slab under a masonry partition does the same. A garage floor poured a year after the house, a basement extended under a later addition, a step in level between an original cellar and a converted store — each is a separate slab with its own field, whatever the floor covering suggests.

Where the house has an existing drainage sump with a lid that can be sealed, it is very often the cheapest point available: the pit is already excavated and already tied into the perimeter drain ringing the footing, so a riser taken off it pulls along the whole drain line — exactly the distribution a mitigator would otherwise spend a day building. The trade-off is that the pump keeps working, so the lid needs a gasketed, mechanically fixed cover with sealed grommets for the discharge and the power lead, and the float must stay serviceable through it.

Set the layout out on the floor before anything is cut. The point wants to be as central to the field it serves as the room allows, clear of furniture and future partitions, and somewhere a vertical riser can leave without crossing a habitable room. That last constraint decides more layouts than the diagnostic does, because a perfect suction point with no route upward is not a suction point.

1. Scan for reinforcement, conduits and under-slab heating before marking anything.
2. Trace footings and thickened slab lines from the original drawings where they exist, and from a hammer survey where they do not.
3. Core a diagnostic hole at the proposed position and read flow and suction there before committing to it.
4. Drill remote test holes at the far corners, behind every interior bearing wall and in each separately poured area, and record the pressure at each with the vacuum running.
5. Where a quadrant reads nothing, choose between a second point and a larger excavation at the first, then test again rather than assuming.
6. Fix the riser route upward before the pit is cut, and patch every diagnostic hole once the design is settled — an open test hole is a short circuit into the room it was drilled in.

Cutting the pit, and why size pays for itself here

The hole through the slab is trivial and the void beneath it is the job. A core through 150 mm of concrete takes minutes; scraping aggregate out through that hole with a bent bar and a shop vacuum takes an hour and decides what the fan will achieve. Every litre removed reduces the entry loss between pipe and field, moving the operating point along the fan curve towards more flow at less suction. EPA/625/R-93/011, Radon Reduction Techniques for Existing Detached Houses, describes forming an excavated void beneath the opening for exactly this reason, and the practice figure quoted most often — a bucketful and upward — is a floor rather than a target. In tight material it is worth taking out several times that.

Core rather than break wherever the slab will allow it. A hammered hole leaves a ragged annulus that is hard to seal cleanly and starts the crack radiating away from the opening that becomes the first air path back into the room. Where reinforcement is hit, cut it back cleanly inside the hole rather than leaving a bar bridging the void with sealant packed around it.

The riser is then set into the hole and sealed to the slab around its full circumference. That collar is a structural detail as much as an airtightness one: it carries the weight of the pipe, it resists whatever gets kicked into it, and it is the only part of the system a vacuum cleaner will eventually be driven against. Use a sealant with movement capability rather than a rigid mortar collar alone, and do not bring a concrete patch up hard against the pipe without a bond breaker — a shrinkage crack around a rigid collar is an inlet reopened three months after handover.

A radon fan lives in the opposite corner of the curve

Almost every other fan in a house is chosen for flow and asked to work against very little pressure. A bathroom extractor is a few tens of litres per second at perhaps fifty pascals; a range hood is chosen on the number printed on the carton. A radon fan is the inverse. It moves a small flow — tens of cubic feet per minute, not hundreds — against a suction that in tight ground can run to several hundred pascals, and it does it continuously, wet, and often at outdoor temperature, for years without being switched off.

Ventilation instinct applied to that duty produces two classic mistakes. The first is fitting the biggest fan on the shelf to a tight floor: on low-permeability ground a high-flow fan simply runs near its shut-off point, making noise and heat, drawing its full current, delivering almost nothing and holding no more suction than a properly matched high-vacuum unit would. The second is fitting a high-vacuum fan to an open aggregate bed, where the field needs volume rather than depth and the curve has nothing to bite on. Match the fan to what the diagnostic told you: high suction and low flow for clay and fines, higher flow and modest suction for a clean stone bed.

Read the manufacturer's published curve rather than the headline rating. RadonAway, Fantech and Festa all publish flow against static pressure for each model, and the point that matters is where your system's total resistance crosses that line. A fan quoted at an impressive free-air figure will be doing a fraction of it at the operating point your pit and pipe impose — which is why the curve is published at all.

Specify for the duty as well as the point. The fan is rated for continuous operation and for the moisture it carries, its housing must be sealed at the seams because everything downstream of the impeller is above ambient pressure and full of soil gas, and it must be listed for where it is mounted. The supply is governed by the adopted wiring code — NFPA 70, the National Electrical Code, in the United States and BS 7671 in the United Kingdom — and the detail installers regret skipping is the disconnect: it has to exist for servicing, and it has to be somewhere nobody will flick it off to quieten the house down.

Sealing, now that it has a job to do

Sealing was useless as a remedy and is essential as a support. Once a fan is holding the sub-slab material below room pressure, every unsealed opening in the floor steals suction from the far corner and spends fan capacity dragging warm indoor air into the ground. The order changed, not the work: the fan goes in, then the openings that measurably rob it get closed, which is why ASTM E2121, Standard Practice for Installing Radon Mitigation Systems in Existing Low-Rise Residential Buildings, carries sealing provisions alongside the depressurisation ones. The list of what matters is shorter than the list of what gets caulked. The perimeter joint between slab and wall is normally the largest single opening in a basement floor and often the only one worth a full day. An untrapped floor drain is a direct short to the drainage system. The open top course of a hollow block foundation wall connects the whole cavity to the sub-slab material and to the room, and capping it moves readings in a way that hairline cracks in an otherwise sound floor never will.

Where there is no slab at all — an earth-floored cellar, or a crawlspace under part of the footprint — the technique changes name and the sealing becomes the field. A membrane laid over the ground with a suction point beneath it is sub-membrane depressurisation, and it succeeds or fails on the continuity of the sheet at its laps, perimeter and penetrations rather than on the fan. That work, and the encapsulation it sits inside, is covered properly in the crawlspace and ground gas membrane guides.

Commissioning, and the five years after it

Two things get fitted at handover and both are cheap. A manometer on the riser, mounted where somebody will look at it, with the commissioned operating point marked on the tube in permanent ink — an unmarked gauge tells nobody anything, because what matters is not the reading but its distance from the day the system worked. And a label carrying the installer, the date, the fan model and the instruction not to switch it off.

The confirmatory measurement comes after a deliberate delay rather than the same afternoon. The room does not jump to its new concentration when the fan starts; it decays towards it at a rate set by the air change of the space, so a basement in a reasonably tight house takes the better part of a day to get most of the way there. That is why the standard sets both a minimum delay after activation and an outside limit on how long the test may be left — read the window in the edition being worked to. Test under the same closed-house conditions as the original, and record the fan model, the manometer reading and where each suction point is, or the next person is hunting for a pit under a carpet.

Failures over the following years are predictable and each writes its own signature. A manometer gone level means the fan has stopped or the pipe has separated, and that reading is the only warning anyone will get. A manometer noticeably higher than the mark with the fan still running means the system is pulling harder against something that has closed: ice at the discharge in a hard winter, or a horizontal leg holding water because its fall was lost when somebody boarded the loft.

A reading lower than the mark means the opposite — the field has opened up. A sump lid unscrewed and never refitted, a floor drain trap dried out over a summer, a new penetration cut through the slab for a shower waste. The fan needs replacing in none of those cases, and it usually gets replaced anyway.