Overview
In remedial programmes, one of the most common mistakes is a generic instruction like “seal the gaps” without defining what the gap actually is.
That usually leads to:
- the wrong product used
- poor durability
- defects reappearing after follow-on trades
The short version: they are different defect types
Penetrations and joints can both look like “gaps”, but they behave differently and often need different detailing.
| Defect type | What it is | Typical example |
|---|---|---|
| Penetration | An opening created for a service passing through a fire-resisting element | Cable tray through a wall, pipe through a slab |
| Joint | A linear gap between construction elements | Wall-to-slab line, wall-to-wall interface |
Treating a joint like a penetration (or vice versa) can create repairs that look tidy on day one but fail later due to movement, poor adhesion, or incomplete continuity.
Why this matters in real buildings
Most repeat defects happen for predictable reasons:
- a quick repair doesn’t match the defect type
- durability isn’t considered (wet areas, movement, vibration, repeated access)
- scope doesn’t define the compartment line and the repair boundary
In practice, access hatches and boxing-in interfaces are where “it looks sealed” becomes “it isn’t continuous”. Treat them as their own defect category so they don’t get lost under generic scope language.
Penetrations (services)
These are openings for:
- pipes
- cables
- containment
- mixed services
What tends to matter operationally:
- what service type it is
- what the surrounding substrate is
- whether the penetration is single/multiple
- whether access is available from both sides
Common penetration mistakes
- sealing only the visible face, leaving an open route on the hidden side
- “stuffing” mixed services into a single opening without a controlled detail
- failing to consider service movement (especially pipes) and future access
Evidence that helps later (penetrations)
- defect ID and location reference
- service type(s) present
- photos before/after (where practical)
If you manage portfolios, consistent defect naming and IDs matter more than perfect wording.
Joints (linear gaps)
These are usually:
- wall-to-slab
- wall-to-wall
- around structural interfaces
What tends to matter operationally:
- movement expectations
- continuity along the line
- ensuring it’s sealed along the full length, not “patched”
Common joint mistakes
- patching small sections rather than treating the full run
- ignoring movement expectations (cracks reappear)
- sealing over debris/dust so the repair fails quickly
Evidence that helps later (joints)
- location and length/run reference (start/end points)
- photos where helpful
- notes on any access constraints
How to scope correctly (a repeatable method)
1) Identify the compartment line and construction element
Is it a wall line, a floor line, or a junction? Confirm this first, because it defines what “compliant” needs to look like.
2) Classify the defect
- penetration (single service)
- penetration (mixed services)
- linear joint/interface
- access panel/hatch interface
- boxing-in interface
3) Specify scope with clear boundaries
Avoid generic scope text. Instead include:
- defect type
- exact location reference
- whether both sides will be addressed (where relevant)
- evidence requirements at close-out
4) Build in durability
If the area is likely to be disturbed (risers, plantrooms, comms routes), include a control measure:
- labelling and IDs
- change control for follow-on works
- clear reinstatement expectations
A quick checklist for property teams
| Question | If “no” | Why it matters |
|---|---|---|
| Do we know the compartment line here? | Escalate for competent confirmation | Otherwise you may fix the wrong thing |
| Have we classified it as penetration vs joint? | Re-scope before instructing works | Different details and durability needs |
| Is access available to inspect both sides (if needed)? | Record constraint and plan | Single-sided fixes can leave open routes |
| Is the output evidence-led (IDs, photos, exceptions)? | Tighten outputs | Governance and repeatability |
Scope language: phrases to avoid (and better replacements)
| Risky phrase | Better scope language |
|---|---|
| “Seal gaps” | “Fire-stop service penetration through [element] at [location], including both sides where accessible.” |
| “Make good around” | “Seal linear joint at wall-to-slab line along [run], full length, with continuity at interfaces.” |
| “Patch as required” | “Treat all breaches along the compartment line in this area; update defect register with IDs and close-out evidence.” |
Durability and change control (why defects reappear)
Risers, plant areas, and ceiling voids are disturbed repeatedly. If you want remedials to last:
- label and ID what was addressed
- include a simple rule for follow-on trades: if you cut through a compartment line, it must be reinstated and recorded
- keep the close-out pack accessible (not buried in a one-off report)
FAQs
Can we treat joints and penetrations the same if it “looks sealed”?
It’s risky. A repair that looks fine can fail due to movement, poor continuity, or because the defect type wasn’t addressed properly. Classify first, then scope.
Do we always need product/system details in records?
Not always, but for higher-risk areas and repeat programmes, recording enough detail to support governance is helpful. If a repair method will be questioned later, it’s worth documenting.
What to ask for in the remedial output
- clear location references
- photos before/after (where practical)
- product/system details where relevant
- exceptions list for access constraints
Related pages
Note
This article is general information. Always align remedials to competent guidance and the building’s fire strategy.