Make conservative repair-or-replace decisions
Generated from
content/lms/fabrication-and-composites/06-inspect-repair-and-escalate-responsibly/02-make-conservative-repair-or-replace-decisions.md; edit the source file, not this page.
Source path: content/lms/fabrication-and-composites/06-inspect-repair-and-escalate-responsibly/02-make-conservative-repair-or-replace-decisions.md
Course: Fabricate composite race-car parts with workshop discipline
Module: Inspect, repair, and escalate with restraint
Estimated duration: 48 minutes
A composite repair decision is not a cosmetic decision first. It is a risk decision first.
When you find damage in a laminate, your job is not to prove that the part can be saved. Your job is to decide whether the part can return to service with enough confidence for the role it plays on the car. That difference matters. A small defect in a non-structural body panel can often be cleaned up, filled, finished, and run. The same size mark in a suspension link, wing mount, bonded insert, or impact structure may be a reason to remove the part from service until it has been tested, repaired by someone with the right process control, or replaced.
The conservative repair-or-replace rule for this lesson is simple: repair only when the damage is minor, visible, local, and in a low-consequence part; replace or escalate when the part is safety-critical, highly stressed, bonded, cored, impact-loaded, heat-exposed, or uncertain.
That rule is not pessimism. It follows from how composites earn their strength. A laminate is not one material in the same way a piece of mild steel is one material. It is fibres, resin, lay-up direction, cure quality, consolidation, core material when present, and sometimes adhesive bonds to metal or honeycomb. McBeath describes professional laminate design as a deliberate mix of fibre types and fabric forms: unidirectional carbon for directional strength and stiffness, woven carbon for all-round performance, and sometimes aramid as a fail-safe or anti-abrasion layer. That tells you something important as a repairer: the original part may be doing several jobs at once, and a simple surface patch may not restore the same balance of stiffness, toughness, abrasion resistance, and failure behaviour.
The second reason for conservatism is that damage can be hidden. Ultrasound can detect contaminant inclusions, voids, and delamination in composite laminates, and McBeath presents it as useful both for quality checking and post-impact inspection. The old coin-tap method still has a place for the DIY tester because a sharp tap suggests sound laminate and a dull thud suggests trouble, but the same passage makes clear that ultrasound is the more advanced method. If your decision depends on what is happening below the surface and you do not have a way to inspect below the surface, you do not have enough evidence to make an aggressive repair decision.
The third reason is consequence. Van Valkenburgh separates life-or-death components from parts whose failure may only cost a race or damage other mechanical assemblies. Axles, hubs, spindles, hub carriers, and steering arms are in the first category. In a composite context, the same thinking belongs on suspension links, steering track rods, pushrods, aerofoil mounts, monocoque structure, and impact structures. Smith expected composites to move into highly loaded parts such as hub carriers, wheels, flywheels, and brake discs as racing technology advanced. McBeath shows that Formula 1 composite suspension pushrods and wishbones were already serious professional test items. You should treat composite parts in those roles with the same respect you would give any other life-safety component, not as bodywork with a more expensive cloth.
This lesson teaches the decision skill. It does not try to replace the separate inspection lesson, the steel-versus-composite failure-mode lesson, or the outsourcing lesson. The skill here is what you do after you have found something: decide whether to make a conservative local repair, replace the part, or escalate the part to a professional inspection and test process.
The decision ladder
Use a five-step ladder every time you discover a defect. Do not start by asking whether you can fix it. Start by asking what the part is and what confidence you can honestly have after the work.
Step one: classify the part by consequence. Ask what happens if the repair is wrong. If the part is bodywork, a low-stress duct, a minor cover, or a cosmetic surface, the consequence of a small local repair failing may be a loose edge, a paint flaw, or a return to the workshop. McBeath specifically treats minor defects in a newly made component as repairable by removing loose gel or fibres and filling with ordinary automotive body filler before sanding and finishing. That is the repair zone: minor, local, visible defects in parts where the filler is not expected to carry structural load.
If the part is a suspension link, steering component, hub carrier, aerofoil mount, bonded metal-joint assembly, monocoque skin, or impact structure, the consequence changes. McBeath says that for safety-critical components like suspension links and aerofoil mounts, quality standards have to be set very high. He also describes proof testing and ultimate tensile testing for composite suspension pushrods with bonded metallic joints, plus corner test rigs that can apply vertical, lateral, longitudinal, and braking loads. Those are not decorative test procedures. They exist because the loads are serious and because confidence has to be earned.
Step two: classify the damage by visibility. If the defect is a loose gel edge, a small fibre lift, a local surface chip, or a spiky trim edge on a non-structural part, you can see the repair boundary. If the damage came from impact, crush, hard mounting load, heat exposure, or a dull coin-tap area, you cannot assume the visible mark is the whole defect. McBeath notes that ultrasound can detect flaws like voids and delamination and can be effective for likely damage after impact. That means impact history alone should move you toward escalation unless the part is clearly low consequence.
Step three: classify the construction. Solid laminate, honeycomb core, bonded metal inserts, film adhesive, mixed fibre lay-ups, and heat-cured parts are not equivalent repair problems. Honeycomb structures may crush progressively under impact, which is useful for energy absorption, but that same property means a part can have sacrificed internal structure to protect the driver. A carbon-skinned aluminium honeycomb monocoque or impact structure is not a fair place for a cosmetic repair mentality. Bonded metallic joints in composite suspension links also deserve special caution. McBeath notes that modern structural adhesives can be effective enough that the bond is not usually expected to fail in tensile testing, but some part of the assembly will fail at a sufficient load. Your garage repair should not be the weak link in a load path that was originally designed, cured, and tested as a system.
Step four: classify your process control. Ask whether you can reproduce the original material, lay-up, fibre direction, resin system, cure, consolidation, and inspection standard. McBeath is clear that autoclave curing gives consistency, consolidation, and confidence for highly stressed components, but even autoclave pressure is no guarantee of a perfect component; skilled and diligent lay-up still matters, and parts may still be rejected. That is the standard you are comparing yourself against when the part is highly loaded. If you cannot reproduce the process or prove the result, replacement or professional escalation is the conservative choice.
Step five: choose the least heroic action that gives adequate confidence. There are only four outcomes. Run as-is after inspection if the condition is known and acceptable. Make a local cosmetic repair if the defect is minor and the part is low consequence. Escalate for non-destructive inspection, proof testing, or professional repair if the part matters and the answer is uncertain. Replace the part if the part is critical, the damage is structural or hidden, the original process cannot be reproduced, or inspection evidence cannot justify return to service.
A useful sentence to carry in your head is this: the repair must be appropriate to the consequence, not to your optimism.
What counts as a conservative repair
A conservative repair is narrow. It fixes a minor, local defect without pretending to restore an unknown structural margin.
The clearest supported example in the corpus is the minor-defect repair after moulding. McBeath describes inspecting a newly released component to confirm it is defect free or to find minor defects. Small defects can be repaired by removing loose gel or fibres with a knife or file and filling the defect with everyday automotive body filler, then sanding and buffing to the required finish. The same section describes washing and polishing a coloured gel coat or washing and finely abrading before paint. In that context, the repair is a surface finishing operation, not a redesign of the part.
The conservative version of that repair has boundaries. You are not using filler to bridge a crack that carries load. You are not hiding delamination. You are not filling a crushed core and calling it restored. You are not repairing a bonded suspension end with body filler because the surface looks smooth afterward. You are cleaning out loose non-structural material, stabilizing a small cosmetic defect, and finishing it so the part can do the same low-risk job it was already doing.
You also protect the part while doing the repair. McBeath’s trimming instruction is specific: cut from the gel side so the gel coat is not cracked or flaked, then clean and smooth the cut edges with a file or abrasive paper on a sanding block. That belongs in repair thinking because careless trimming can create new damage. A repair that leaves cracked gel coat, ragged fibres, unsupported edges, or poorly abraded paint prep is not conservative. It has turned a small defect into a new initiation point.
A conservative repair also includes an inspection after the work. You are not finished when the filler cures or the edge looks smooth. You recheck the local laminate, the edge, the surrounding area, and the mounting points. If the part was damaged by an incident, you also inspect the rest of the load path because impact energy may have gone somewhere else. If you cannot explain why the damage is local, your confidence is lower than your finish quality suggests.
The key boundary is structural uncertainty. If the repair would need to restore fibre continuity, core bond, adhesive joint integrity, suspension stiffness, crash energy absorption, or aerofoil mount strength, it is no longer a simple conservative repair. It is either a professional composite repair with process control and inspection, or it is a replacement decision.
What counts as a replace decision
A replace decision is not an admission that you failed to fix the part. It is the correct answer when the part’s job demands more confidence than your repair can produce.
Replace the part when failure could directly produce loss of control. Van Valkenburgh’s life-or-death list includes axles, hubs, spindles, hub carriers, and steering arms. In this lesson’s composite fabrication context, carry that same seriousness to composite suspension links, steering track rods, pushrods, wishbones, bonded clevises, highly loaded aero supports, and composite parts whose failure could put a wheel, wing, or driver protection system out of service. McBeath’s professional examples are telling: composite suspension pushrods with bonded metallic joints are proof tested or tested to ultimate strength; suspension corners are put through repeated logged-load cycles; strain gauge and suspension displacement data can be replayed in the laboratory. If a professional team would need proof or durability evidence before trusting the component, you should not trust a visual patch after impact.
Replace the part when an impact structure has done its job. Honeycomb core can crush gradually under severe impact, and McBeath explains that this makes honeycomb useful in racecar impact structures because it provides progressive energy absorption and less severe deceleration values for the driver. A crushed energy absorber is not a failed repair candidate in the usual sense; it may be a consumed safety device. The more it has protected the driver, the less you should expect it to be reusable without a defined inspection and replacement standard.
Replace the part when core damage is suspected and you cannot inspect or rebuild the core correctly. Honeycomb and sandwich construction depend on the relationship between skins, adhesive film, and core. McBeath’s glossary describes film adhesive as a dry adhesive film used to bond core materials to laminate skins. If the skin is cracked, the core is crushed, or the skin-to-core bond is uncertain, a surface patch may restore appearance while leaving the sandwich unable to carry load as designed.
Replace the part when the part depends on cure quality you cannot verify. Autoclave curing under pressure and elevated temperature is described as a must for consistency, consolidation, and confidence in highly stressed components such as monocoque chassis, high-downforce aerofoils, and suspension parts. Room-temperature repair work may be entirely appropriate for some low-stress bodywork, but it does not automatically recreate a high-quality pressure-cured laminate. If your confidence depends on pretending those processes are equivalent, replace or escalate.
Replace the part when heat, load, and adhesive performance are part of the original design. McBeath describes testing components in high-temperature environments when competition car parts must function in hot areas. He also describes specialised adhesives for bonded metallic joints in composite suspension links. If a part lives near heat or carries load through an adhesive joint, the repair decision has to respect that environment. A repair that looks acceptable at room temperature in the paddock may not be acceptable under heat and cyclic load.
Replace the part when inspection evidence is ambiguous. Ambiguity is not neutral. If a tap test gives a dull area, if the mark came from an impact, if the part sounds different from its mirror image, if the edge has fibre lift that continues under paint, or if mounting hardware shows load transfer into the laminate, treat that as an evidence problem. The correct conservative move is not to explain the uncertainty away. It is to escalate or replace.
What counts as escalation
Escalation is the middle path between a garage cosmetic repair and immediate replacement. It is appropriate when the part is valuable or difficult to replace, the consequence is significant, and there is a realistic inspection or proof route that can create confidence.
The corpus gives you several escalation models. One is proof testing: a component is loaded to a predetermined limit and, if it passes, may be put into service. Another is ultimate tensile testing: the sample is loaded until it fails so the failure strength can be measured. McBeath uses a composite suspension pushrod with bonded metallic joints as an example of a component that could be evaluated either way. For an in-service race part, proof testing is the relevant concept because you are trying to verify that the part can meet a required limit without being destroyed.
Another model is coupon testing. McBeath describes laminate coupons made to the same lay-up as a component and cured at the same time, then tested to establish that the component will perform to the required standard. That is quality assurance, not guesswork. For a club-level builder, the practical lesson is that confidence is strongest when you built a testable path into the process before the part was needed. If you made no coupons, saved no lay-up schedule, and have no process record, later repair decisions become much harder.
A third model is durability testing. McBeath describes servo-hydraulic machines and corner rigs that can apply vertical, lateral, longitudinal, and braking loads and replay data logged from real circuits using strain gauges and suspension displacement measurements. The advantage is that testing can be repeated in a laboratory, stopped for inspection, and performed without risking a driver or valuable car. At the club level, you may not have that rig, but the decision lesson remains: if the real concern is cyclic load durability, a static look in the paddock is weak evidence.
A fourth model is non-destructive inspection. Ultrasound can detect inclusions, voids, and delamination in composite laminates and can help check damage after impact. A coin tap is more rudimentary, but it still has a place as a screening tool. The conservative use of a tap test is to find reasons to escalate, not to overrule a bad history. A sharp response in one area does not prove the whole structure. A dull response, especially near an impact mark or bonded joint, is a reason to stop.
Escalation should be specific. Do not write vague notes such as have someone look at it. Write what question must be answered. Is the laminate delaminated around the impact? Is the core crushed? Is the bonded insert still sound? Did the mount see enough load to crack the surrounding skin? Does the part need proof testing before reuse? Clear questions help the professional decide what inspection method or test is appropriate.
The decision matrix
When you are tired after a session and looking at a damaged part, use a simple matrix.
Low consequence plus visible minor surface defect usually points to local repair. Example: a bodywork panel has a small loose gel edge or local fibre fuzz after trimming. You remove loose gel or fibres, fill only the minor defect, sand, buff, and recheck. The part’s job is aerodynamic shape or coverage, not primary load-bearing safety.
Low consequence plus hidden or impact damage points to inspect harder before repair. Example: a duct or cover took a strike and now has a dull-sounding patch. The part may not be safety-critical, but delamination can grow or the part can shed fragments. Tap-test, compare to known-good areas, inspect mounts and edges, and repair only if the damage is understood and local. If not, replace the low-cost part.
High consequence plus visible minor cosmetic defect points to escalation, not automatic repair. Example: a composite suspension cover or aero mount has what looks like a small chip. Because the part is highly loaded or supports a high-downforce element, the standard is higher. McBeath says safety-critical components such as suspension links and aerofoil mounts require very high quality standards. Cosmetic confidence is not enough.
High consequence plus impact, crush, bonded-joint concern, core concern, or uncertain sound points to replace or professional proof. Example: a suspension pushrod with bonded metallic ends was hit by debris or contacted another car. Professional practice includes proof tests, tensile tests, and durability rigs for such components. A club paddock repair should not return it to service on appearance alone.
Impact structure plus crush points to replacement under the governing rule or manufacturer process. Honeycomb impact structures are valuable because they crush progressively and absorb energy. That means visible or suspected crush is not a cosmetic defect. It is evidence that the structure may have spent some of its designed capacity.
The skill is not memorizing every possible composite failure. The skill is sorting the case into the correct confidence category.
Worked example: a cracked bodywork edge after trimming
You have just pulled a small composite body panel from the mould and trimmed the spiky overlap. While cleaning the cut edge, you find a short area where the gel coat has flaked and a few loose fibres are exposed. The panel is bodywork, not a suspension bracket, monocoque structure, impact attenuator, or wing mount. It carries shape and maybe airflow management, but it is not a primary safety load path.
The conservative decision is a local repair, provided the defect stays local. First, inspect around the edge. You are looking for a crack that continues into the laminate, a dull sound compared with the rest of the panel, loose plies, or damage near a fastener. If the defect is only loose gel and local fibre fuzz at the trim line, McBeath’s minor-defect process applies: remove loose gel or fibres with a knife or file, fill the defect with suitable automotive body filler, then sand and buff or prepare the surface for paint.
The technique matters. You do not grab the most aggressive grinder and chase the flaw into a larger repair. You remove only loose material. You support the edge. You keep the repair local. You smooth the edge with a file or medium to coarse abrasive paper on a sanding block. If more trimming is required, you cut from the gel side to avoid cracking or flaking the gel coat. After the repair, you inspect again, because a good finish can hide poor edge work.
What would make the decision change? If the crack runs away from the edge, if the surrounding laminate sounds dull, if the panel is actually a structural undertray or aero element with meaningful load, if the defect is at a mounting point, or if the part has been impacted rather than merely trimmed, you no longer have the simple minor-defect case. You inspect deeper, escalate, or replace.
Worked example: an impacted carbon pushrod with bonded metal ends
A carbon composite suspension pushrod has a scrape and a small impact mark after a car-to-car contact. The metallic end fittings look aligned. The surface mark is not dramatic. Someone in the paddock says the bond looks fine and suggests sanding the mark smooth.
That is the wrong decision frame. This is not body filler territory. McBeath uses a composite suspension pushrod with bonded metallic joints as an example of a component that can be proof tested to a predetermined limit or tested to ultimate tensile failure. He also notes that the structural adhesives used in such parts are effective enough that the bond is not usually expected to fail in the tensile test, meaning the composite or metal part may fail first. That is a high-confidence engineered assembly, not a cosmetic shell.
Your conservative decision is remove from service until replacement or professional proof is available. The reason is not that every mark means certain failure. The reason is that the part is safety-critical, loaded, bonded, and potentially damaged below the visible surface. The load path includes the laminate, the adhesive, the metal joint, and the geometry of the end fitting. A surface repair cannot prove that the fibres, bond line, or local crush are intact.
If the car owner wants to save the part because it is expensive, escalation must answer a specific question: can the part still carry the required load with margin after impact? A professional path could include non-destructive inspection for delamination and a proof load if the component and rules allow it. Without that evidence, the conservative answer is replacement.
Worked example: a honeycomb nose or side impact part after contact
A honeycomb-cored impact structure has a visible dent after an off. The skin is not fully torn. From five feet away the shape still looks acceptable. The tempting repair is to fill the dent, sand it smooth, and repaint.
The conservative answer starts with the purpose of the structure. McBeath explains that honeycomb cores are used in stressed and unstressed components, including carbon-skinned aluminium honeycomb monocoque construction, and that honeycomb has a useful tendency to crush gradually under severe impact. That makes it valuable in racecar impact structures because it absorbs energy progressively and reduces driver deceleration.
A dent in that kind of structure may be evidence that the core has crushed. If the core has crushed, the structure may have done work. Filling the skin does not restore crushed honeycomb or re-establish the original skin-to-core relationship. The decision is therefore replacement or escalation according to the manufacturer, rule set, or professional composite inspection process. The repair question is not whether the surface can be made smooth. The question is whether the energy-absorbing structure still has the capacity it was designed to have.
For an intermediate club driver or builder, the practical rule is direct: do not turn consumed crash structure into pretty crash structure. If the part is an impact structure and you have evidence of crush, treat it as a safety item.
Worked example: a high-downforce aero mount with a small crack
A wing support or aerofoil mount has a small crack near a fastener. The car is not a prototype, and the part looks much simpler than a Formula 1 component, so the owner is tempted to drill-stop the crack and add a patch.
McBeath groups aerofoil mounts with safety-critical components when discussing the high standards required for quality. He also lists high-downforce aerofoils among the highly stressed components for which autoclave curing provides the consistency, consolidation, and confidence needed in professional work. The lesson is not that every club wing mount must be autoclaved. The lesson is that aero loads are real, and a mount failure can have consequences beyond losing a body panel.
The conservative decision is escalation or replacement unless you can show that the crack is superficial and outside the load path. A crack near a fastener is especially suspicious because fasteners concentrate load. If the mount is composite, the repair may need to restore fibre direction, bearing strength, local stiffness, and possibly bond integrity. A patch that makes the crack less visible may also change stiffness and move the next failure somewhere else.
If the part is low-cost or easy to fabricate correctly, replacement is cleaner than heroic repair. If the part is expensive or integrated, define the inspection question and involve someone with the right composite capability. Either way, the decision is not made by surface appearance alone.
Sub-skill: separating cosmetic finish from structural confidence
Intermediate drivers and builders often get into trouble because composites reward appearance. A repaired laminate can look excellent before it is trustworthy. Paint, gel coat, filler, sanding, and polish all improve the surface. None of them prove fibre continuity, core condition, bond integrity, or cure quality.
Train yourself to use two separate sentences. The first sentence is finish quality: the edge is smooth, the gel coat is repaired, the surface is ready for paint. The second sentence is structural confidence: the laminate is sound, the core is uncrushed, the bond is intact, the part can carry its load. If you only have evidence for the first sentence, you have not earned the second.
The minor-defect repair from McBeath belongs to finish quality. Proof testing, coupon testing, durability testing, and ultrasound belong to structural confidence. The conservative decision is to avoid mixing those evidence types. A good surface finish is enough for a small cosmetic defect on bodywork. It is not enough for a suspension pushrod, a bonded steering link, or an impact structure.
Sub-skill: reading part context before damage size
Damage size is a weak first question. A tiny crack in the wrong place can matter more than a large scrape on the right part. Van Valkenburgh’s inspection discussion begins with which parts are critical and how often they should be inspected, not with crack length. He points out that some components can cause serious accidents, while others may only cost the race or cause internal damage.
Apply that same hierarchy before you pick up tools. Ask what system the part belongs to. Steering, suspension, braking, hub support, aero support, monocoque structure, and impact absorption get high-consequence treatment. Bodywork, ducts, covers, and non-structural fairings get lower-consequence treatment, though they still deserve competent repair.
Then ask how the part was made. A mixed lay-up may have fibres doing different jobs. A cored panel may hide crushed core under an intact skin. A bonded insert may load the adhesive and the laminate together. A heat-exposed part may need material performance at temperature, not just room-temperature shape. The more specialized the original part, the less likely a generic repair is adequate.
Sub-skill: using tap tests without overtrusting them
The coin tap is useful because it gives quick feedback on laminate continuity. McBeath describes the old test as tapping the laminate with the edge of a coin and listening for a sharp tap as compared with a dull thud. In the same passage, he notes that ultrasound is more advanced and can detect inclusions, voids, and delamination.
Use the tap test as a screening tool. Compare symmetric areas. Compare known-good laminate to the suspicious area. Mark the boundary of any dull response. If the dull area grows beyond the visible damage, the visible mark is not the whole story. If the dull area is near a mount, bonded insert, edge, or impact mark, escalate.
Do not use a sharp tap as permission to ignore impact history on a critical part. The tap test is not a proof load. It does not replay lateral, vertical, longitudinal, and braking loads. It does not tell you adhesive strength at temperature. It does not establish fatigue life. It is one clue, not a verdict.
Sub-skill: asking what proof would look like
A strong repair decision includes a proof standard. What evidence would convince a careful mechanic that the part can return to service?
For a small bodywork defect, proof may be simple: loose material removed, defect local, edge sound, surrounding laminate sharp on tap, mount area unaffected, finish restored. For a non-critical duct, proof may include confirming that the repair will not shed material or interfere with adjacent systems.
For a suspension link, proof is much more demanding. McBeath’s examples include proof testing to a predetermined limit, ultimate tensile testing, coupon testing, dynamic load testing, and complete suspension corner rigs. You may not have those tools, but you should understand what kind of evidence the part would deserve. If the only proof you can offer is that it looks fine, the part should not go back into a high-consequence role.
For an impact structure, proof may be a manufacturer replacement criterion, a rulebook requirement, or professional inspection of the sandwich structure. If the part was designed to crush, the proof question is not whether it can be cosmetically restored. It is whether the energy absorption capacity remains.
Sub-skill: documenting the decision
Conservative decisions are easier when you leave a record. Write down the part, location, damage description, incident history, inspection method, decision, and reason. Include whether the part was repaired, replaced, or escalated. If you performed a local repair, record that it was cosmetic or non-structural unless you have evidence to say otherwise.
This habit follows Van Valkenburgh’s broader inspection logic. The more the driver knows about the condition of the car and the care going into it, the more attention can go to driving. Documentation is part of that condition knowledge. It also keeps you from re-litigating the same suspect part three events later when everyone has forgotten the original impact.
Calibration cues: how you know your decisions are improving
You are improving when your repair decisions get slower at the right moments and faster at the easy ones.
You should be fast with obvious low-risk defects. A small loose gel edge on non-structural bodywork should not turn into a three-day engineering debate. You inspect, clean, fill, sand, finish, and recheck. You know the repair is cosmetic, and you do not describe it as structural.
You should slow down immediately when the part is safety-critical or the damage history includes impact. If you catch yourself asking how to save a damaged pushrod before asking how to prove it, your decision process is backwards. Improvement looks like removing the part from service, defining the inspection question, and choosing replacement or escalation without drama.
Your notes should become more specific. Early notes may say crack in carbon panel. Better notes say small gel chip on non-structural left sidepod cover edge after trimming; loose gel removed; no dull tap response around defect; filled and sanded; rechecked mount area. For a critical part, better notes say right front composite pushrod contacted wheel; visible scrape near bonded end; removed from service; replacement fitted; suspect part held for professional inspection. Specific notes show that you are classifying consequence, construction, damage, and evidence.
Your parts bin should change. You should have fewer repaired mystery parts and more clearly tagged parts: serviceable, cosmetic repair only, inspect before use, scrap, replace after impact. That is not bureaucracy. It is a way to keep uncertainty from sneaking back onto the car.
Your language should also change. Avoid saying it should be fine when what you mean is you cannot see anything worse. Say what you know, what you do not know, and what evidence would change the decision. That language protects you from optimistic repairs.
Common mistakes
Mistake one: treating carbon as magic bodywork. The wrong version is assuming that a carbon part is strong because it is carbon, then repairing surface marks as if every part were a simple fairing. The good version is asking what fibres, lay-up, cure, core, and bonds are doing in that specific part. McBeath’s discussion of mixed fibres and fabrics shows why the material choice is purposeful, not decorative.
Mistake two: using filler as a structural argument. The wrong version is filling a crack, sanding it smooth, and calling the part repaired because the crack is no longer visible. The good version is using filler only for minor local defects where the part is non-structural and the surrounding laminate is sound. If load-carrying fibres, core, or adhesive joints are involved, filler is finish work, not structural proof.
Mistake three: trusting surface appearance after impact. The wrong version is deciding from the outside that an impact-damaged composite part survived. The good version is recognizing that delamination, voids, inclusions, and core crush may be hidden, and that ultrasound or other professional inspection may be needed. A dull tap is a stop sign, not an invitation to argue.
Mistake four: ignoring the load path. The wrong version is repairing the visible skin while forgetting the fastener, insert, bracket, or bonded metal end that actually puts load into the laminate. The good version is tracing the load from the attached system into the part and inspecting the highest-stress area, not just the prettiest surface.
Mistake five: repairing consumed impact structure. The wrong version is filling a crushed honeycomb area because the outer shape can be restored. The good version is treating crush as evidence that the structure has absorbed energy. Impact structures exist to protect the driver by controlled deformation. After they do that job, they deserve replacement or a defined professional inspection standard.
Mistake six: confusing paddock practicality with adequate proof. The wrong version is saying that because the next session is soon, the repair standard can be lower. The good version is separating schedule pressure from evidence. If the part is low consequence, make the local repair and run. If the part is high consequence, the lack of time makes replacement more attractive, not less.
Mistake seven: overreading a tap test. The wrong version is using one sharp-sounding spot to declare a critical part safe. The good version is using tap response as one screening clue and escalating any suspicious sound, especially around impact marks, mounts, cores, and bonded joints.
Mistake eight: forgetting that professional parts can still be rejected. The wrong version is assuming that because a part was autoclaved or professionally made, it must be perfect unless obviously broken. McBeath notes that autoclave pressure is not a guarantee of a perfect void-free component and that skilled lay-up remains essential. The good version is inspecting even professional parts and respecting any evidence of damage.
Drill: the three-bin repair decision exercise
Run this drill at your next shop night or event inspection. You need ten minutes, masking tape or tags, a marker, and the parts or photos available to you.
Pick five composite items around the car. Include at least one non-structural bodywork or duct piece, one mounting area, one cored or sandwich-looking part if present, one part near heat if present, and one high-consequence part or a photo of one if you are not allowed to handle it. For each item, write the part name and place it mentally into one of three bins: local repair acceptable, escalate for inspection or proof, replace if damaged.
For each item, answer four questions in writing. What happens if it fails? What construction features matter: laminate, core, bond, insert, heat exposure? What damage evidence would force escalation? What evidence would be enough to return it to service?
Then add a fake defect to each card: small gel chip, dull tap area, impact mark, cracked fastener hole, crushed edge, heat-discolored patch, or scrape near a bonded end. Re-sort the part. The success criterion is that your decision changes when consequence and damage history change. A gel chip on bodywork may stay in local repair. A scrape near a bonded suspension end should move to replace or professional proof. A dent in honeycomb impact structure should move to replace or defined professional inspection.
Repeat the drill after the next real incident, but with actual evidence. The goal is not to become timid. The goal is to make the conservative decision automatic before paddock pressure, sunk cost, or session timing starts arguing with you.
A second version of the drill uses sound. If you have a known-good non-critical laminate panel and a suspect non-critical panel, tap both with a coin edge and listen for the difference between a sharp response and a dull response. Do not practice this on a critical component as a return-to-service test. The success criterion is simply that you can identify changes in sound and mark the boundary of the suspicious area. The decision lesson is that a dull area earns more inspection; it does not earn a prettier repair.
When this principle gets nuanced
Conservatism does not mean every blemish sends every part to the trash. Racing budgets are real, and composite parts range from disposable covers to highly engineered structures. The nuance is matching the decision to consequence and evidence.
A club-level splitter endplate with a small edge chip may be repaired locally if it is not carrying a critical load and the surrounding laminate is sound. A high-downforce wing mount with a small crack near a fastener deserves a much more conservative answer. Both may be carbon. Both may be black and glossy. They are not the same decision.
A newly moulded panel with a minor defect may be repaired by removing loose material, filling, sanding, and finishing. An in-service panel damaged by impact may need deeper inspection even if the visible defect looks similar. Manufacturing defects found before installation and impact damage found after service have different histories.
A professional team may proof test a component and return it to service. That does not mean a visual garage inspection is equivalent to proof testing. It means there is a path to confidence when the right equipment, standards, and process records exist. Your decision should be conservative about the evidence you actually have, not the evidence you wish you had.
A cored structure may look less damaged than it is because the skin hides the core. A bonded joint may look aligned while the local laminate or adhesive interface has been overloaded. A heat-exposed part may look fine cold and behave differently hot. These are not reasons to panic; they are reasons to define the right inspection question.
Cross-references to related skills
Use the inspection lessons before this decision lesson. You cannot make a good repair-or-replace call if you have not inspected the laminate, edges, mounts, and surrounding load path.
Use the steel-and-composite failure-mode lesson when you are tempted to apply metal habits to laminates. Drilling, stop-holing, bending back, and welding analogies do not transfer cleanly to fibre, resin, core, and adhesive systems.
Use the outsourcing lesson when the decision requires ultrasound, proof testing, coupon interpretation, adhesive-bond repair, autoclave-level process control, or rule-governed crash structure decisions. Outsourcing is not a weakness when the part’s confidence requirement exceeds your tools.
Use the rollover-structure lesson for cages, crash structures, monocoque areas, and anything governed by rules or engineering analysis. Driver protection parts are not places for informal repair standards.
The final check
Before you return any repaired composite part to the car, say the decision out loud in one sentence.
This is a non-structural cosmetic repair on a low-consequence part, and I inspected enough to know the damage is local.
Or: this is a high-consequence composite part with uncertain internal condition, so it is being replaced or escalated.
If you cannot honestly say one of those sentences, you are probably in the danger zone between appearance and evidence. Stop there. The conservative choice is the one that gives the driver confidence for the next session because the part has earned it, not because everyone ran out of time.
Worked example: a cracked bodywork edge after trimming
A non-structural body panel with a small gel-coat flake and a few loose fibres at the trim edge is the clean local-repair case. Inspect the surrounding laminate, confirm the defect is local, remove loose gel or fibres, fill only the minor defect, sand and finish, then recheck. If the crack runs into the laminate, the panel sounds dull, the defect sits at a mount, or the part is actually carrying meaningful aero or structural load, the decision changes to deeper inspection, escalation, or replacement.
Worked example: an impacted carbon pushrod with bonded metal ends
A composite suspension pushrod with bonded metallic joints is not a cosmetic repair candidate after impact. The professional examples in the corpus involve proof testing, ultimate tensile testing, and durability assessment because the part carries serious load and the bond, laminate, and metal ends work as a system. The conservative decision is removal from service until replacement or professional proof is available.
Worked example: a honeycomb impact part after contact
A honeycomb-cored impact structure with a visible dent may have crushed internally. Because gradual crush is part of why honeycomb is useful for energy absorption, the dent may show that the structure has spent part of its designed capacity. Filling and sanding the skin can restore appearance without restoring the core or skin-to-core relationship. Treat this as replacement or defined professional inspection, not a casual surface repair.
Common mistakes
The main mistakes are treating every carbon part like cosmetic bodywork, using filler as if it proves structural strength, trusting surface appearance after impact, ignoring fasteners and bonded load paths, repairing consumed impact structure, lowering the standard because the next session is close, overtrusting a tap test, and assuming professional manufacture makes later inspection unnecessary. Good practice separates finish quality from structural confidence and matches the decision to consequence and evidence.
Drill: the three-bin repair decision exercise
At the next shop night or event inspection, pick five composite items on the car and sort each into local repair acceptable, escalate for inspection or proof, or replace if damaged. For each item, write what happens if it fails, what construction features matter, what damage evidence would force escalation, and what evidence would be enough to return it to service. Then add a fake defect to each card and re-sort. Success means your decision changes when consequence and damage history change.
When this principle breaks down
The principle becomes nuanced when a blemish is truly minor, the part is low consequence, and the damage is fully visible and local. In that case, conservative does not mean throwing the part away. It means making a narrow cosmetic repair and documenting what you did. The principle does not soften for safety-critical parts, bonded joints, cored impact structures, high-downforce mounts, heat-exposed parts, or any case where internal damage is uncertain.
Author Review
No quiz questions are attached to this lesson.
Sources
| # | Document | Chunk | Pages | Score | Collection |
|---|---|---|---|---|---|
| 1 | Competition Car Composites Simon McBeath | 24ba3ae8-8ba0-28bd-3edb-094db2b95f90 | 179 | 1 | uio_books_raw_v1 |
| 2 | Competition Car Composites Simon McBeath | 2afc9093-4cdd-d995-d340-aac602fd741a | 176 | 1 | uio_books_raw_v1 |
| 3 | Race Car Engineering Mechanics Paul Van Valkenburgh | 6761997c-1267-f401-0671-5bfbf75c8c8d | 104 | 1 | uio_books_raw_v1 |
| 4 | Competition Car Composites Simon McBeath | 50e8919c-ef19-4354-dea8-95d9c311c69e | 178 | 1 | uio_books_raw_v1 |
| 5 | Competition Car Composites Simon McBeath | 0417d4d8-2df3-87dd-a347-0684c8b7e5b5 | 178 | 1 | uio_books_raw_v1 |
| 6 | Competition Car Composites Simon McBeath | 13ad50d9-320e-9ff6-b6a1-35cebddda495 | 111 | 1 | uio_books_raw_v1 |
| 7 | Competition Car Composites Simon McBeath | a92a57d7-66ad-7c18-c969-cf0c0d4005e9 | 204 | 1 | uio_books_raw_v1 |
| 8 | Tune To Win Carroll Smith | 7f0cfbfd-fb17-324a-2d89-e006011f5f59 | 166 | 1 | uio_books_raw_v1 |