# BMW M5 E39

> The free buying guide from Last Real Engines, updated 2026-10-13. Web version: https://lastrealengines.com/m5-e39/free-guide/ · full guide: https://lastrealengines.com/m5-e39/
## The story

*What this car is, and the six questions that decide whether the one you are looking at is a good buy*


## 1 · Who built it, and what they built

The E39 was drawn by Joji Nagashima, a Japanese designer who joined BMW in 1988 and was sketching
this car as early as 1989 — though the drawings from that year were rejected, and look it. The
proposal that won was selected in June 1992 and then frozen for production under Chris Bangle, newly
arrived as design chief, whose later work would divide opinion far more than this car ever has.
Joji Nagashima's design was signed off in June 1992 and the car arrived in 1995. What came out was long, low and almost completely undramatic,
and it is the reason people who dislike modern BMWs still make an exception for this one.

Underneath sits a double-jointed MacPherson front with aluminium steering knuckles and outer strut
tubes, and a four-link rear axle. That front axle is the one the age list opens
with, and it is worth knowing now rather than at the garage: the arm and its ball joint are a single
part.

### The engine

The M division's part was the S62, and it was their first road-going V8. Not a tuned M62: 4,941 cc
against the M62's 4.4 litres, a bore of 94 mm and a stroke of 89 mm, 11.0:1 compression where the
M62 ran 10.0:1, a double-row timing chain instead of a single, hollow camshafts, two air intakes with
their own mass airflow meters, and a semi-dry sump with two extra scavenge pumps that come in under
hard cornering. It was also BMW's first V8 with double-VANOS, on both the intake and the exhaust
camshafts.

Then the part that still sounds like a decision nobody would sign off today: **eight individual
throttle bodies**, one per cylinder, electronically actuated, with a normal and a sport map the driver
picks himself. 400 PS at 6,600 rpm, 500 Nm at 3,800 rpm, 7,000 rpm on the dial. A Siemens MSS 52 to
run it.

The same engine went into the Z8 — a car that people now keep in climate-controlled storage. You are looking at it in a four-door saloon.

### And where it was built

Here is the detail that tells you most about the car. The E28 M5 was assembled by hand at BMW
Motorsport in Munich and later in Garching; the E34's bodyshells went to M in Garching, where each
car was put together by hand over about two weeks. The E39 was not built that way at all. It went
down the ordinary 5 Series line at Dingolfing, engine included, alongside the 523i that the regional sales manager ordered. That is the whole character of
the thing. Nothing about it announces itself, and 20,482 of them were built.

Twenty-five years later that is the reason it can still be had for sensible money — and the reason
you need to know what to look at before you go and see one.

### How the figures in this guide work

**How much work each job is, not what it costs.** Most points in this guide carry an effort — an hour
or two, half a day, a day, two or three days, or a week or more in the workshop — plus a line saying
whether the money sits in the part or in getting to it. Where a documented hour count exists for this
car, that is there too.

**There are no prices, and that is deliberate.** What a job costs depends on which country you are in
and which workshop you walk into, so a figure I cannot verify is worse than no figure. If you want
money, ask any workshop one question — *what is your hourly rate?* — and every hour in this guide
becomes your own number, in your own currency. One question, and the answer is actually true for you.

None of this is advice about one particular car, because I have not seen the car you are looking at.
Have every one of them inspected before you buy, by someone who knows this engine.

---

---

## 2 · The six questions that decide whether *this* car is a good buy

Every one of these cars is now somewhere between twenty-three and twenty-eight years old, and the
difference between a good buy and an expensive one has almost nothing to do with the asking price.
It comes down to six questions. None of them can be answered by looking at the car in a photograph,
and five of them cannot be answered by looking at the car at all.

**1 · Has the VANOS been done — and which of its two problems are we talking about?**
This is the single most misunderstood thing about the S62. It has two entirely separate failures, with two
entirely separate bills, and almost everyone treats them as one item. Worse: the symptom most people
read as VANOS can have four other causes, so the first step is a read-out, not a rebuild. Anyone quoting you a VANOS
figure without saying which of the two he means is guessing.

**2 · The rod bearings — and can anyone prove it?**
The one genuine dispute in the community. Specialists advise preventive replacement inside a mileage window; there are also documented cars
running past 300,000 km on their original bearings with nothing found on stripdown. Both camps have
evidence. What is not in dispute is what it costs when one lets go.

**3 · What happens to the cooling system, given that this is an all-aluminium engine.**
On this all-aluminium engine an overheat is documented to take the head gaskets with it, which is why the
cooling system is done as a set, before it fails. There is also a trap: the component people expect to fail first fails in a way that makes the
engine run *cold*, and the temperature gauge on this car is electronically damped and will not tell
you.

**4 · The rear subframe.**
Structural, and little known. A specialist that reinforces these subframes names a specific weak point
on it that can flex or crack under high load — and if it has cracked, the subframe comes out. It can be
checked, and almost nobody does; where to look, and what for, is in the buying guide.

**5 · Rust, in eight specific places.**
Not general age — eight places worth checking on this bodyshell, in an order you can work through. One of
them is only visible with a piece of trim removed.

**6 · What has already been paid for.**
On this car the odometer tells you less than on almost any other. A documented 250,000 km is a better buy than an unexplained 120,000 km, and the reason is
arithmetic rather than sentiment: most of the money in a twenty-five-year-old M5 sits in work you
cannot see, and the only evidence that it has been done is paper.

### What the honest answer to all six looks like

If the engine's own items have demonstrably been done, a good car asks its owner for **roughly a third of
the age list in the first two years, and none of the heavy jobs** — not because things break, but because
things arrive that were always going to arrive. In workshop time that is a handful of half-days spread over
two years. If they have not been done, they come on top, and that is a very different thing: the VANOS
alone is two or three days, and so is the subframe if it turns out to be cracked.

That number is only useful if you know what it is made of. The buying guide takes each of these six
apart, adds twenty-four things that are simply twenty-five years old, and for most of them gives how much
work it is, whether the money sits in the part or in getting to it, what tells you about it at a viewing,
and whether it has to be done now or can wait years. It ends with a card you print and take with you, in the order you do it, including the two
moments at a viewing that come only once.

---

---

## 3 · Which years, which specification

The M5 E39 was built from 1998 to 2003 with one engine and one transmission: the **S62B50**, a
naturally aspirated 4,941 cc V8 with 400 PS, and the **Getrag 420G** six-speed manual.

**There was one gearbox, and it is a manual.** That is the fastest filter you have. If an advert says
"M5 automatic", you are almost certainly looking at a 5 Series with the M package — often a perfectly good
car, but a different one, with a different engine and a very different price.

**And one myth out of the way straight away, because this one costs buyers money.** The Nikasil bore
linings you read about on BMW V8s belong to the **M60 and the early M62, roughly 1993 to 1996**. The
S62 uses Alusil and does not have the problem. Anyone who walks away from an M5 E39 for fear of
Nikasil is walking away for the wrong reason.

**Does the model year matter?** A little, and less than you would think. Halfway through the run the
5 Series was facelifted with revised lighting and updated on-board electronics. For what you are going
to check, that changes little. One thing worth knowing: the **earliest cars** have the first version
of the throttle actuators (see the chapter on this engine's own faults) and an older navigation screen. That is no reason to rule
out a model year — it is a reason to ask, on an early car, whether those actuators have ever been done.

The main rule stands: **buy the best-maintained car, not the newest one.**

**Options can be dealt with briefly, and that is good news.** These cars came generously equipped and
there is no rare configuration to hunt for. What makes a difference day to day is a sunroof or no
sunroof, and which seats it has. What makes a difference on resale is mostly a saleable colour and
honest original condition — and that only counts the day you sell it on.

**And check that it is a real M5.** Well-built 540i conversions exist, and they are worth a
fraction. On a car at this price that is not suspicion, it is a standard check.

Beyond that, look for a car nobody has modified. Factory wheels, no lowering, no aftermarket exhaust,
no wiring someone has added themselves. Modifications are not automatically bad, but an original car
is more often a sign of maintenance than of projects.

**The wheels and tyres are a separate point here.** The M5 runs different sizes front and rear. If it
has the same size all round, or two brands on one axle, that tells you someone has been watching the
pennies. And look at the **DOT date** in the sidewall: on a car that covers few kilometres, ten-year-old
tyres are not tyres you trust, however much tread is left.

---

## 4 · What it is worth

This is the hardest thing to establish on your own and it does more than anything else to decide
whether you buy well. I am not going to give you a price list — it varies by country and by month, and
an out-of-date number is worse than no number. I am going to give you the method.

**Compare on paperwork, not on mileage.** On this car the odometer tells you less than on almost any
other. A documented 250,000 km with a folder of invoices is a better buy than an unknown 120,000 km
that nobody can account for. That is also what the age list shows: most of the money in this car sits in
work you cannot see.

**There are three kinds of car for sale, and each one deserves a different offer.**

*The documented car.* Full history, invoices tied to odometer readings, and an owner who can tell you
when what was done. Asking price at the top of the market, and usually justified. What you are buying
here is the history rather than the car.

*The tidy car with no paperwork.* Drives well, looks well, and the owner says everything has been done.
This is the hardest of the three, because there is nothing to see. You buy it on the condition it is in
today and you budget the top end of the figures in the age list.

*The project.* Known work, price adjusted to match. That can be an excellent buy if you know what you
are taking on and have the money available. Ask for a list of what is known — the owner often has one.

**How to set the three against each other.** Take the asking price, work through both lists — this engine's own faults and the age list — and
add up what is coming in the first two years. Call that the entry price. Then compare entry prices
rather than asking prices, and you will find that the most expensive car is surprisingly often the
cheapest. Not always — a documented car with an unrealistic asking price is still too expensive — but
often enough to be worth doing the sums every time.

**Where to look:** your country's big classified sites, the specialist dealers and the classic auction
sites. Follow the market for a month before you go and see anything. You do not need to know what one
is worth; you need to know what is happening — which cars sell quickly, which sit for months, and at
what price.

---

## 5 · What it is actually like to own one

Start with the thing you notice in the first ten metres. There is no turbocharger between your right
foot and the engine, and after twenty-five years of cars that all have one, that reads as a different
kind of machine rather than an older one. Eight throttle bodies open when you ask them to. Nothing
builds, nothing waits, nothing arrives a beat late.

It also asks something of you. Peak torque is at 3,800 rpm and the limiter is at 7,000, so the car
only makes sense in the upper half of the range — and using the upper half of the range on a
naturally aspirated 4.9 is the whole reason to own one. BMW knew what people would do with it: there
are two extra scavenge pumps in the oil system that come in under hard cornering. That is not a
feature you fit for a car you expect to be driven gently.

Then it goes quiet and behaves like a 5 Series. Four doors, a boot, back seats an adult fits in
without negotiating. Nobody looks at it. Colleagues do not know what it is. Of everything this car
does, that is the part owners talk about most.

**What it costs you in attention.** More than a modern car and less than you fear. It wants the right
oil, an annual look at a specialist who knows the engine, and front tyres more often than you are used
to. It does not want to sit still for months — nothing on it improves from standing.

**What it gives back.** It has stopped depreciating the way a modern car does, it does something every
time you start it that new cars no longer do, and there will be fewer of them in ten years than there
are now. That is not an investment and you should not call it one. It is a car that never asked to be noticed and never needed to be.

---

## The buying guide — €24

Everything above is what the car is. This is what you check.

Six faults that belong to this engine, taken apart one by one, with
how much work each repair is and how to tell which of them you are looking at.
Twenty-four things that are simply twenty-five years old, every one
with how much work it is, what tells you about it at a viewing, and whether it has to be done now,
within a year or two, or not at all. The eight places to check this bodyshell for rust. The rear subframe, and where you check it. The viewing in ten steps, including the
two moments that come only once. The sentences you say to the seller, and how to build an
offer out of what you found.

And a card you print and take with you — sixty lines, in the order you do them.
That comes as a separate file, because nobody wants to print a whole guide to take a few pages to a
car park.

### This is what one of the twenty-four looks like

**3 · Rear springs**
Labour ~1.5–2.5 hours · **half a day** · cheap part, easy access
A known E39 item: the rear springs can break. It is a fail point at
the annual roadworthiness test and it is visible with a torch.

A broken spring is a job for straight away.

### And this is how the card reads

- [ ] 1 · The cold start · *do nothing else*
- [ ] 2 · The engine bay, engine off
- [ ] 3 · Round the outside · *rust and wheels*
- [ ] 4 · Inside, ignition on · *keep it short, the battery*
- [ ] 5 · The fault memory · *at a garage, not with a cheap adapter*
- [ ] 6 · The test drive · *half an hour minimum*

### What €24 buys, put next to what it saves

A VANOS overhaul on this engine is **two or three days, and almost all of it labour** — and which of its two problems you are paying for changes the bill. A cracked rear subframe has to come out: **two or three days in the workshop.** The guide is **€24**. It will not stop any of that from being needed — but it does
tell you which of them you are looking at before you agree a price, and that is the whole of the
difference between paying for a car and paying for someone else's postponed maintenance.

### Why there are no prices in it

Because a price we cannot verify is worse than no price, and because a price would be wrong for most of
the people reading this anyway. Workshop rates across Europe differ widely, and so do parts prices. A quote from a specialist in one country tells you nothing in another.

**What does not change is how much work a job is** — that follows from what has to come off the car.
So every point in the guide says whether it is an hour or two, half a day, a day, two or three days, or a
week or more in the workshop, with a labour figure where a source gives one. And it says whether the money sits in
the part or in getting to it, which is what lets you tell a fair quote from a bad one.

If you want a figure, ask any workshop one question — *what is your hourly rate?* — and every effort
level in the guide becomes your own number, in your own currency. One question, and the answer is actually true
for you.

### How this guide was put together

Because you should know before you pay for it.

- Every fault and symptom comes from a source the guide names, the manufacturer's own documents first. The
  effort levels are our assessment, unless a source gives the hours.
- Where the community genuinely disagrees — and on this engine's rod bearings it does — you get both positions rather than a preference.
- The effort of every job is counted item by item and checked against the list before every edition.
- It says as plainly what you can leave alone as what you have to do. Six of the
  twenty-four age points start from *leave it*.

Leave your email address at the checkout and both files arrive straight away: one guide and one card,
as two pdfs. No subscription, no account.

[Buy the BMW M5 E39 guide — €24](https://lastrealengines.com/m5-e39/)

*Have every car inspected before you buy by someone who knows this engine. This is general information
about the model and not advice about a specific car.*