Starter Replacement Cost: The Part That Gets Blamed

There is a particular kind of repair bill that arrives with a second, invisible cost attached to it, and this is one of them. Somebody turns a key, the car does not crank, a starter goes on the car, and the fault comes back three weeks later. The starter was fine. What was wrong was a ground strap, or an ignition switch, or a battery that had lost enough capacity to run the dashboard and nothing more. The money did not go on a bad quote. It went on a good quote for the wrong part.

So the useful shape for this page is not a price followed by some caveats. It is an order — the sequence in which the possible causes should be knocked out, arranged so that the cheapest and quickest checks come first, and the starter itself comes last because it is the most expensive thing on the list and the least likely to be at fault on any given car.

Everything below expands that order and then does the two things a used-car buyer actually needs from it: what the job costs when the starter genuinely is the answer, and how to find out at a viewing whether a car has a starting problem it is not showing you.

What the starter and the solenoid actually do

Most people know the starter turns the engine over. Fewer know that the component bolted to the engine is two devices sharing a housing, and that most of the ways it fails belong to the smaller of the two.

The motor is an ordinary direct-current machine, and a crude one by design. Current goes in through a pair of carbon brushes that press against a copper commutator on the spinning armature; the commutator reverses the current in each winding as it goes round, which is what keeps the shaft turning in one direction. Subtlety was never the design brief. What the thing has to do is produce an enormous amount of torque for a second or two and then stop, and everything about its construction follows from that.

The solenoid sitting on top of it is doing two jobs at once, and understanding that is the key to most of this page. When you turn the key, a small current runs through the solenoid’s windings and pulls an iron plunger along its bore. That plunger is mechanically linked to a lever that throws the starter’s small pinion gear forward into mesh with the ring gear on the flywheel — the flexplate, on an automatic. Only at the end of that travel does the plunger push a heavy copper disc across two large studs, closing the circuit that lets battery current into the motor itself.

The sequence matters: the gears mesh first, and the motor is powered second. That is why a healthy start begins with a single crisp engagement rather than a scrape. It is also why the starting system contains two completely different electrical circuits sharing one part. A small control circuit runs from the ignition switch, through the interlocks and usually a relay, to the solenoid winding — thin wire, modest current, the sort of thing a corroded connector can interrupt. A separate heavy circuit runs directly from the battery’s positive terminal to the solenoid stud, and from the solenoid into the motor, carrying more current than anything else on the vehicle will ever ask for.

Two more parts of the assembly earn a mention because they produce distinctive noises. The drive has a one-way clutch in it, so that the instant the engine fires and the ring gear starts turning faster than the pinion, the clutch releases rather than letting the engine drive the motor backwards through its own gearing. When that clutch wears, you get the classic high whirr with no cranking: the motor is spinning happily and the engine is not moving. And many modern starters are gear-reduction designs, with a small planetary set between the armature and the drive, which is why a compact starter can turn a large engine and why the pitch of a modern start sounds different from an older one.

The figure worth worrying about is not the quote

This site declines to print repair prices at all, and the broad argument for that position is made elsewhere on it. The version that applies here deserves stating on its own terms, because on this repair a published range is not merely perishable. It is aimed at the wrong target altogether.

Think about what a starter quote is actually competing against. Where the elimination has not been done, the live alternatives to a starter are a battery, a set of cleaned terminals, a length of braided strap, a relay, a switch, or nothing whatever beyond tightening a bolt somebody left loose. The distance between the top and the bottom of that list dwarfs the distance between one workshop’s labour rate and the next one’s. Shopping around carefully for a part you did not need is losing the argument by a wide margin while feeling as though you won it.

The job itself supplies the second problem. On one engine the starter sits high on the bell housing, reachable from above with the bonnet open, held by a couple of bolts and two cables. On another — common on transverse layouts, and on some V-configurations where the space in the middle of the vee was too useful to leave empty — it lives beneath the intake manifold, and getting to it means the manifold, its gaskets, and every hose, bracket and loom that crosses over the top. That is one part, one line on an invoice, and a completely different day’s work. Nothing about the badge on the bonnet announces in advance which of those two cars is yours.

So the honest sequence puts money last. Establish whether you need one; establish which version of the job your particular vehicle demands; then ask what it costs. Run backwards, which is how it usually goes, the number arrives first and quietly settles a diagnosis nobody has performed.

The elimination order, and why cheapest first is method rather than thrift

Start with a distinction, because it separates this page from a related one. Our guide to reading a refusal to start from the noise and the dashboard sorts the problem by behaviour: what you hear, what the instruments do, and which family of faults each of those points at. That is the right first cut, it is the fastest way to get from a dead car to a short list, and this page does not repeat it.

What follows assumes you have already made that cut and landed in the starting-circuit family — the engine is not turning over, or is turning over badly. From inside that family, behaviour stops being a good sorting principle, because half a dozen different faults produce indistinguishable silence. So the sort changes. Inside the family, the useful ordering is by what each test costs to run against how much it removes from the list. That is not the same as ordering by likelihood, and the difference is the whole method.

Consider the two extremes. Checking whether the battery can hold voltage under load takes minutes, is free at a great many parts counters, and eliminates the most common cause of everything on this page. Replacing a starter buried under an intake manifold takes most of a day and eliminates exactly one candidate. Even if those two were equally likely to be the answer — and they are nowhere near equally likely — you would still do the first one first. That is not being careful with money. It is the same reasoning any diagnostician uses: run the test that cuts the field fastest per unit of effort, and keep running it until the field is one item long.

Stage one: is there enough current to do the job?

Everything the starting system does is downstream of the battery, and a battery that is merely low behaves nothing like a battery that is flat. It will light the dashboard convincingly, run the radio, unlock the doors and then collapse the instant the solenoid closes. This is where the largest number of unnecessary starters come from, because from the driver’s seat the car looks electrically alive right up to the moment it does not crank.

A resting voltage reading answers only whether the battery is full, and full is not the same as sound — a cell can carry a convincing surface charge and still collapse the moment real current is demanded of it. A load test is what separates those two, and it takes minutes. The testing, the service life question and the six things that decide what a replacement costs are covered on our page about what actually sets a battery replacement bill, so they are not repeated here. What belongs on this page is only the ordering rule: the battery is tested before anything else is suspected, every time, without exception, including on cars where somebody has already fitted a new battery. Especially on those.

Stage two: is the current getting there, and getting back?

A battery in perfect condition delivers nothing useful through a bad connection. This stage is the one most often skipped, partly because it involves no parts and therefore feels like it cannot be the answer.

The heavy circuit has four places to fail: the positive terminal clamp, the cable from that clamp to the solenoid stud, the ground path from the battery’s negative terminal to the engine block, and the ground strap between the engine and the body. That last one is the classic. Current has to return as well as arrive, and the braided strap that carries it corrodes at its bolted ends where salt and water collect. A car with a failing ground strap hands you a symptom you cannot tell apart from a dead starter, for a very small fraction of the money.

The measurement that settles this whole stage is a voltage drop test taken while the engine is actually being cranked — comparing what the battery has with what arrives at the starter, and separately, what the difference is across the ground path. Resistance in a connection is invisible with no current flowing and obvious with several hundred amps trying to pass. A connection that measures fine on a static check and fails under load is precisely the fault that gets a starter fitted, because every static check anybody ran came back clean.

Corrosion that has crept up inside the insulation of a cable belongs here too. The cable looks perfect from the outside, and the green powder under the sheath is doing the same job as an open circuit. It is worth a mention that if the car is being started with jump leads to get it moving, the same principle applies to the leads themselves — our guide to the jump start procedure and the reasoning behind its order covers why undersized cable produces a healthy-sounding donor and a car that still will not turn.

Stage three: is the car willing?

The control circuit is separate, thin, and full of switches whose entire purpose is to refuse. An automatic will not crank unless the transmission range switch agrees it is in park or neutral; a manual will not crank unless the clutch pedal switch says the pedal is down. Both of those are adjustable, both drift out of adjustment, and both fail. The ignition switch itself is a wearing mechanical contact set at the end of a barrel that gets a fair amount of use, and the starter relay is a component that costs almost nothing and fails like any other relay.

All of these produce a car that does absolutely nothing at the key while looking electrically perfect. All of them are cheaper than a starter, and one of them is free to test.

Only then, the motor

When the supply is proven, the path is proven, and the control circuit has delivered current to the solenoid terminal while the key is held, and the motor still does not turn, you have arrived at the starter by elimination rather than by assumption. That is the position from which it is reasonable to buy one.

It is also the position from which the number becomes worth asking about, because now you are pricing a part you know you need rather than a guess.

The tap on the starter, and what it is actually telling you

Somebody will always suggest hitting it. The suggestion is sound, the trick works often enough to have survived several generations of drivers, and almost everyone who uses it draws the wrong conclusion from the result.

Here is what a knock does. Inside the motor, the carbon brushes are held against the commutator by springs. As the brushes wear down, the springs extend, the pressure they can apply falls, and eventually the contact between brush and commutator becomes marginal. Where the armature happens to have stopped now matters: on a worn or burnt commutator segment, the brush may be sitting on the one spot where it cannot pass enough current to break the motor out of standstill. A sharp knock jars the brushes back into contact, or shifts the armature fractionally off the dead spot, and the motor goes.

The same physical logic applies to the solenoid. A plunger that has begun to stick in its bore — corrosion, wear, a weak return spring — can hang partway, close the contacts imperfectly or not at all, and be freed by a jolt. Solenoid contacts themselves fail differently: the copper disc and the two studs it bridges pit and burn where they arc, until the resistance across them is high enough to starve the motor. Tapping helps that condition less reliably, but it is part of the same family of faults, and it is the other thing a starter that has become unreliable is usually doing.

So the trick works, and the reason it works is the important part. It works because there is genuine wear inside the unit. That is a diagnosis. It is a good one, and it is the sort of confirmation that would cost money to obtain any other way.

Two cautions about actually doing it, since people will. The thing to strike is the body of the motor, firmly but not wildly, with something with a bit of weight rather than a full swing of a club hammer; the solenoid, the terminals and the wiring are not targets, and a cracked housing turns a diagnostic tap into a new part. And look at where your arm is going. The starter is bolted to the engine, which means the exhaust manifold and its pipework are frequently right beside it, and the heavy battery cable terminates on the solenoid a few inches away and is live whether the ignition is on or not. Parking brake on, transmission in park or neutral, and nobody turning the key while your hand is in there.

There is a used-car observation attached to all of this that matters more than the mechanics. A seller who knows the trick has a car that starts on demand, and you will never see the tap.

What sets the bill, with access at the top of the list

Assume the elimination has been done and the starter is genuinely the answer. What separates a modest bill from a serious one is mostly not the part.

One more item does not fit a table row. Look at what is above the starter. A valve cover gasket weeping down the back of the engine, a leaking power steering line, or oil tracking along a bell housing seam will drip onto the unit and cook into it. A starter that has been marinating in oil for a year has a reason for failing, and fitting a new one underneath the same leak buys you a repeat performance. That is a question to ask when a car you are looking at has had a starter recently and nobody has mentioned why.

New, remanufactured, and the question that separates them

A starter is a better candidate for remanufacture than most components on a car, and the reason is the crudeness mentioned earlier. It is a simple machine with a small number of identifiable wear items: brushes, bushings or bearings, the solenoid plunger and contacts, and the drive assembly with its one-way clutch. A rebuilder who replaces all of those has genuinely renewed the unit, because there is very little else in it to renew.

What makes this awkward is that nothing visible from outside separates a thorough rebuild from a cursory one. A full strip with new contacts, a new drive, fresh bushings and a bench test is one product. A wipe, a brush set and a coat of paint is a second, and it turns up in a box the first would be proud of.

So the question worth asking is narrower than it is on most parts, and it is specific to this component: was the solenoid rebuilt or reused? On a great many starters the solenoid is what failed — burnt contacts, a worn plunger — and a rebuilt motor carrying its original solenoid across is a unit with the failed half still attached. A supplier who can answer that question is telling you something about the whole operation.

The same knowledge cuts the other way and produces the one genuine saving on this page. Because burnt contacts are such a common failure, contact-and-plunger repair kits exist for many common starters, and on a unit that is easy to reach that is a real and sensible repair rather than a bodge. Where it stops making sense is on the car whose starter lives under the intake manifold. When the labour dominates the bill, doing the cheap fix means paying the expensive part of the job twice if it turns out the motor was tired as well — which, on a starter that has been struggling long enough to burn its own contacts, it very often is.

The connection nobody makes: weak batteries kill starters

These two components are sold, quoted and discussed as though they had nothing to do with one another beyond sitting at either end of a cable. The relationship is closer than that, and it runs in a direction most drivers find counter-intuitive.

A direct-current motor draws current according to how fast it is turning. Spinning freely, it draws relatively little. Loaded down and turning slowly, it draws a great deal, and at the instant of standstill it draws the most it ever will. That is why a starter is designed for seconds of work at a time: the current it pulls while it is getting an engine moving is enormous, and essentially all of the energy that does not become rotation becomes heat.

Now take a car with a battery that has lost a third of its capacity. Every start becomes a longer, slower, hotter event. The motor labours instead of spinning, so it draws harder for longer. The brushes and commutator run hotter and wear faster. The solenoid contacts arc more heavily each time they close on a higher current. Insulation ages. And because the car is slow to fire, the driver holds the key longer, which is exactly the wrong instinct and doubles the exposure.

Do that through one winter and the starter has aged more than the calendar says. The owner, entirely reasonably, replaces the battery when it finally gives up. The car starts properly again. And then the starter fails in the spring, apparently out of nowhere, and nobody connects the two events because they happened months apart to two different parts.

The same argument applies to every high-resistance connection on the heavy circuit. From the motor’s point of view there is no difference between a battery that cannot supply the current and a cable that will not pass it — both arrive as a lower voltage at the terminal, and the response to a lower voltage is a slower turn, a bigger current draw and more heat. A corroded battery clamp is not only a no-start risk. It is a starter-life risk, and it is free to clean.

The third member of this group is the charging system, because a battery that is never properly refilled produces exactly the same slow, hot cranking as a battery that has worn out. Telling those two apart before either one is replaced is what our guide to the variables behind a charging system repair exists for, and it belongs first in the queue, since replacing a starter on a car whose real fault is upstream of the battery achieves nothing at all.

What the federal records can and cannot say about this

Nobody publishes starter motor failure rates. There is no dataset that says how many miles a starter lasts, and any page that produces such a number has either invented it or borrowed it from somebody who did. What can be looked at is narrower and still worth having.

Within that archive, the electrical system heading holds 145 records carrying an odometer reading, and their midpoint sits at 49,000 miles. Pool all fifteen headings together and the file’s midpoint is 60,000. The engine heading, the largest single group in it with 200 records behind it, does not arrive at its own midpoint until 58,000.

Taken as an ordering rather than a measurement, what that says is that owners hit reportable electrical trouble sooner than they hit reportable engine trouble. It predicts nothing about an individual vehicle. What it does is support the shape of this page: on a used car with a five-figure reading on the clock, the electrical side is not something to file behind the mechanical side and get to later. It turns up first.

What the recall side adds, and the one line in it that names this part

A complaint is a single owner reporting something that happened to them. A recall campaign is a manufacturer admitting a fault was designed or built into a run of cars, and agreeing to correct it at its own expense. That says a great deal about the run and comparatively little about the particular car on the driveway.

That archive contains 145 campaigns once duplicates are stripped out, spread over 24 component families. Electrical system takes 10 of them, contributed by 7 manufacturers with no engineering in common. The contents are the part worth reading. The biggest cluster is software rather than hardware of any description; wiring supplies the next three; a body control module’s software accounts for one more; and exactly one campaign in the family names a starter assembly.

One campaign is not a trend and should not be presented as one. What it is, is a useful demonstration of the boundary between the two kinds of evidence. Regulators act on defects — parts specified or built wrongly, which the manufacturer is obliged to remedy. Brushes worn thin over a decade are not a defect. They are a component doing the one thing it was always destined to do, on roughly the schedule it was always destined to do it. So starter trouble on a used car is overwhelmingly a maintenance and wear question rather than a recall question, and the single campaign that does appear is the exception that marks where the line falls.

That said, an open campaign is worth checking on any car you are considering, for the plain reason that the remedy is free to the current keeper at a franchised dealer regardless of how the car was sold or how many times it has changed hands. Notification runs by post to a registered keeper, so a vehicle that has moved on twice has usually outrun its own letters, and the campaign simply sits there unattended. It takes a minute to pull the open campaign list attached to a VIN, and what to do with the result — how to read it, and what to put to the dealer — is covered in our guide to running an open-recall check on a VIN.

Buying a car with a starting fault it is not showing you

If you are reading this with a viewing booked rather than a car in pieces, this is the section to take with you. The argument in it is narrower than the usual advice about cold starts: it concerns intermittent cranking specifically, and what a single successful start does and does not establish.

Take a starter that has begun to fail in the ordinary way — worn brushes, a commutator with a bad spot, contacts that are burning. It does not fail every time. It fails occasionally, and the occasions are governed by where the armature happened to stop, how warm everything is, and how much the battery has in it that morning. A car in that condition might refuse to crank once a fortnight, or once a week, or twice on a cold Monday and then not again until the following month.

Now count what you learn from a viewing. You turn the key once. The car starts. You have taken a single sample from a process that fails some fraction of the time, and the fraction is one you cannot see and the seller may not have measured either. If the fault shows up one time in twenty, your one start had a nineteen in twenty chance of looking perfect. Nobody has deceived you. The test was simply far too small to detect the thing you were testing for.

That asymmetry is why intermittent starting is the easiest fault on a car to sell without disclosing. Concealing a knocking engine takes effort. Concealing this takes nothing at all: you simply have a normal morning.

Two starts, and the second one is the one nobody asks for

The cold start at the beginning of a viewing is essential, and it is not this page’s subject — our page on what a cold engine gives away in its first fifteen seconds covers what it reveals and how to arrange to see it. What belongs here is the second start, and almost nobody asks for it.

Heat is the specific enemy of a starting system. The motor is bolted to a hot engine, frequently within a few inches of exhaust pipework, and it soaks up that heat while the car sits after being driven. A starter with marginal internals, or a solenoid whose contacts are burnt, or a control wire with a high-resistance joint, will often work perfectly from cold and refuse when everything under the bonnet is at working temperature. This is a genuinely common pattern and it is invisible in the sequence a normal viewing follows, because the normal sequence is: start it, drive it, park it, agree a price, go home.

So change the sequence. Drive the car properly, park it, and leave it standing while you get through everything else — the service history, the documents, the bodywork, the tyres, the spare wheel well. Then start it again yourself, hot, with the bonnet shut and the radio off, and listen. That is the test that catches a heat-soaked starting fault, it costs you nothing but the order you do things in, and no reasonable seller can object to it because you are simply starting a car you have already been allowed to drive.

What to listen to on the first turn of the key

Turn the key yourself rather than watching somebody else do it, and pay attention to the first half second, which is where nearly all of the information is.

Then ask the question directly, because it is a fair one and the answer is informative regardless of what it is: has this car ever needed a second attempt to start, and has anyone ever had to help it along? An owner who has kept on top of a car answers it without having to think. What you are watching for is not a confession but a hesitation.

When the car is already running as you arrive

An engine that is warm or idling when you walk up has already given its most revealing performance to an empty audience. The right response is to say so plainly and ask to come back, and the reasoning behind that — together with what to do when a second visit is impossible — sits on the no-start page rather than being rebuilt here.

The starter-specific point is simply this: a car that has been running for ten minutes before you arrived will start again for you, almost certainly, and that second start tells you nothing at all. A recently run engine turns over more easily, the battery has just been charged, and any marginality in the system has been papered over by the conditions. Whatever else it may be evidence of, it is not evidence that the starting system is sound.

Put it on the inspection list by name

Almost everything on this page is settled by measurement in a few minutes on a ramp, and by nothing at all in a car park. A voltage drop test across the heavy circuit while the engine is cranking, a battery load test, and a look at what is leaking onto the starter from above will together answer questions that no amount of standing and listening can answer. Ask for those by name rather than assuming a general once-over covers them. Briefing a workshop properly, and what a thorough pre-purchase inspection ought to reach, is set out separately.

And remember which half of a car an inspection cannot reach. A ramp shows you the vehicle as it stands this afternoon. Its working life, anything that happened to it, and whatever its title says are matters for the paperwork instead, so give it two minutes and read what the records hold against that VIN before a number is agreed rather than after.

Keeping the component in proportion

None of this makes a starter a reason to avoid a car. It is a bounded, ordinary, entirely foreseeable component with a bounded repair, and every vehicle needs one eventually. Nothing about a starter failure implies anything sinister about the rest of the car.

What a buyer gets out of it is smaller and more practical than a warning: a short list of things you can hear for yourself, a question that is fair to ask, and a rearrangement of the order of a viewing that costs nothing and catches a fault most people never test for. The starting system is also, uniquely among the expensive-sounding items on a car, one where a few minutes of careful method routinely turns a serious quote into a cheap one — which is the entire argument of the order at the top of this page.

Common questions

Why does hitting the starter with a hammer make the car start?

Because there is wear inside it. The carbon brushes that carry current to the spinning armature wear down, the springs behind them lose pressure, and where the motor happened to stop can leave a brush unable to pass enough current to break out of standstill. A knock jars the brushes back into contact or shifts the armature off the bad spot. A sticking solenoid plunger is freed the same way. The trick working is a positive diagnosis of a failing starter, not a repair, and the interval between the times it needs help gets shorter. Note also that it only proves something when it works: a starter that ignores a tap has told you nothing.

How do I know it is the starter and not the battery?

You do not, from the symptom, and that is the central problem this page is about. A low battery will light the dashboard convincingly and then collapse the instant the solenoid closes, which looks identical to a dead starter from the driver’s seat. Have the battery load-tested first — it takes minutes and is often free at a parts counter — and then have the voltage drop measured across the heavy cable and the ground path while the engine is being cranked. Those two tests between them eliminate the large majority of no-crank faults before any part is bought.

Can a bad ground cause a no-crank that looks exactly like a bad starter?

Yes, and it is one of the most commonly missed causes on the list. Current has to return to the battery as well as leave it, and the braided strap between the engine and the body corrodes at its bolted ends where water and salt collect. A failing ground produces silence or a click at the key while everything else on the car appears healthy. It is cheap to fix and it is invisible on any static electrical check, which is why the test has to be done while the starter is actually trying to turn.

What makes one starter replacement so much more expensive than another?

Access, mostly. On some engines the starter is bolted to the top of the bell housing and reachable from above in a short session; on others it sits underneath the intake manifold, or behind exhaust pipework, or is reachable only from below with the car raised, and the labour to get to it is the larger part of the bill. Neither the size nor the price of the car predicts which arrangement you have. The other escalation to know about is damage to the flywheel or flexplate ring gear, which requires the transmission to come out and changes the scale of the job entirely.

Is a remanufactured starter a sensible buy?

Often, yes. A starter has a short list of wear items — brushes, bushings, the solenoid plunger and contacts, and the drive with its one-way clutch — so a properly rebuilt unit is genuinely renewed rather than merely tidied. The variance is in the rebuilding, and the useful question is specific: was the solenoid rebuilt or carried over? A rebuilt motor wearing its original solenoid is a unit with the commonly failed half still attached. On a car where the starter is buried, the labour dominates the bill and the case for the better part gets stronger, because a second attempt costs the whole job again.

Will a weak battery damage the starter?

Over time, yes, and the mechanism surprises people. A motor turning slowly draws more current than one turning freely, not less, and almost everything that does not become rotation becomes heat. Cranking against a tired battery therefore means longer, hotter, harder work every single start — more wear on the brushes and commutator, heavier arcing at the solenoid contacts. A winter of struggling starts ages a starter well beyond what the calendar suggests, which is why a starter failing shortly after a battery was finally replaced is a sequence rather than a coincidence.

Can I tell at a viewing whether a used car has a starting problem?

Only partly, and only if you change the order of the visit. A single successful start is one sample of a fault that may only appear occasionally, so it proves very little on its own. Turn the key yourself and listen to the first half second for grinding, a whirr without cranking, or a laboured first turn from cold. Then drive the car, park it, do everything else, and start it again hot — heat-soak faults in a starting system are common and a normal viewing never tests for them. Ask directly whether the car has ever needed a second attempt, and put a voltage drop test on the list for the inspection.