8

A recent Handelsblatt Global Edition article, How electric cars will charge Germany's job market, talks about changes in the job market due to the introduction of electric cars.

One of the first points it makes is about electric car production requires less of a workforce than production of combustion engine cars. Concretely, the article says:

While an electric car consists of some 200 parts, there are more than 1,000 parts in gasoline or diesel vehicles.

Is there really an inherent difference in "part count" of that order of magnitude between gasoline/diesel cars and electrical cars?

I would have expected that any such difference is caused by changes that happen to coincide with the switch to developing electric vehicles, such as new production techniques and designs (enabling more complex single-piece structures instead of assemblies) or increased outsourcing (such that more ready-made assemblies of "sub-parts" are bought from suppliers and treated as "one part" once they arrive at the actual car manufacturing plant).

EDIT: As is noted in several of the comments, a clearer definition of what constitutes a "part" is required to make answers comparable. Going by the context of the article, it seems that whenever two pieces enter the production process in a separate way such that they need to be assembled at some point during the production process, these two pieces must be considered "two parts".

With this said, the question whether the cited statement holds true only when excluding the assembly of components that the car manufacturer buys pre-assembled from suppliers is very valid. The article appears to insinuate that there is no such restriction to the statement (as it derives conclusions about the overall required workforce in car production), but answers may well come to the conclusion that the restriction is necessary for the claim to hold (or maybe not even then).

EDIT2: To make the definition of parts even clearer: This is not about "moving parts", or "parts that can be detached"; it is about any parts that were separate at any point during the production process, at any place in the supply chain. Two parts that are welded together for the finished car to be counted separately, just as the wire and the insulation of a cable are to be counted separately.

O. R. Mapper
  • 673
  • 1
  • 6
  • 13
  • 12
    Electric engines are indeed far simpler machines than multi-piston four-stroke engines. But that claim still seems dubious to me, because most of the parts of a car aren't related to the engine at all. Also, "part count" is not as clear as a measure as it sounds. How do you count pre-assembled component which is assembled from multiple components in a different factory which in turn got assembled in yet another different company? – Philipp Jan 11 '18 at 12:38
  • In a less detailed form the claim is also made by high profile sources like The Economist, https://www.economist.com/news/leaders/21726071-it-had-good-run-end-sight-machine-changed-world-death so it's probably worth investigating how much truth there is behind it. I can think for instance that while you cast the block engine of a gas/petrol car, you need assemble every stator and rotor core from to laminates. Granted, there are different types of complexity behind these two manufacturing techniques. – Fizz Jan 11 '18 at 12:55
  • Rotor/stator making machine: https://www.youtube.com/watch?v=y0feSEkR3K4 It's true that these are not *spare* parts; you can't change just one lamination after they're put together, especially if the electric motor is balanced, which for anything but cheap toys, it is.. – Fizz Jan 11 '18 at 13:02
  • 2
    @Philipp _because most of the parts of a car aren't related to the engine at all_ - This claim is also dubious. Do you have numbers supporting this? – Common Guy Jan 11 '18 at 13:24
  • Slightly differernt comparison in terms of moving parts: https://avt.inl.gov/sites/default/files/pdf/fsev/compare.pdf Unrelated discussion on HN : https://news.ycombinator.com/item?id=7583758 – Fizz Jan 11 '18 at 13:37
  • 6
    Yeah, you gotta define "part". An internal combustion engine has on the order of 10-20 parts per piston, plus maybe another 100 one-off pieces. And, if it uses a timing chain, one might be able to count that as another 100-200 pieces. A common electric motor, as Fizz says, contains several hundred laminations, plus how do you count the hundreds of feet of wire? And then figure all the parts unrelated to propulsion -- radio, seat belts, hood latch, wheels and tires. – Daniel R Hicks Jan 11 '18 at 13:40
  • @Philipp: "Also, 'part count' is not as clear as a measure as it sounds." - I absolutely agree. That's why I had also mentioned the issue with counting pre-assembled components that you point out in my question. – O. R. Mapper Jan 11 '18 at 14:17
  • 1
    As rough guide to "what are parts?" I'd offer two ideas which I admit are not without grey areas: "if two things move with respect to one another, they are different parts" and "If a thing can be 'easily' removed and refitted it is a distinct part". I happily note that a chain falls between these definitions, it is composed of "things that move with respect to each other" (so is many parts) but those sub-things cannot (really) be "removed and refitted" (so links are *not* parts - but the chain is). – Grimm The Opiner Jan 11 '18 at 14:48
  • 2
    Usually the main reason people say an electrical car is simpler and has fewer (moving?) parts is that a transmission is unnecessary in an electric car. – Batman Jan 11 '18 at 15:31
  • @DanielRHicks From a car manufacturers point of view a combustion engine is probably many parts but an electric engine might just be one (they can replace and modify many combustion engine parts but probably get electric motors as a single, unmodifiable unit). Clearly defining what they mean by "parts" is very important. – matt_black Jan 11 '18 at 16:47
  • 1
    It's much more common to hear people to talk about electric cars having fewer *moving* parts, which is relevant because less movement -> less friction -> less wear -> in theory, less maintenance. And *moving parts* is definitely more countable – user56reinstatemonica8 Jan 11 '18 at 16:53
  • @GrimmTheOpiner: Unfortunately, both your definitions miss the context of car manufacturing. Even two pieces of metal that are welded together for the final car have to be considered two parts in the context of the article, because welding together these two parts would be more work (*if* we decide to take the underlying assumption from the article, more parts = more labour, at face value) than having just one part from the get-go. I have edited the question to include a tentative definition of "part" that I think matches what must have been the intention of the article's author. – O. R. Mapper Jan 11 '18 at 18:16
  • @DanielRHicks: Both a ICE and an electric car have radio, seat belts, hood latch, wheels and tires. It is quite valid to focus on the differences -- engine and power train. – DevSolar Jan 11 '18 at 20:50
  • 1
    @DevSolar - Yes, and arguably there are more "parts" in the onboard computers than in either power plant. The point is that, without a definition of "parts", the question is meaningless. – Daniel R Hicks Jan 11 '18 at 22:30
  • @DanielRHicks: The question asks about the truthfulness of the sentence in the context of the article it was cited from. I have meanwhile edited the question to explicitly clarify how to count parts the same way it was presumeably intended by the article. – O. R. Mapper Jan 12 '18 at 09:27
  • It would make sense that an electric engine would be much simpler - it's not doing the same functions as a gas engine. The electric engine is taking the power and transferring that power into the work that a car needs to do. A gas combustion engine also has to have the equipment to generate that power in the first place, as well as then transferring that power to work. – PoloHoleSet Jan 12 '18 at 14:48
  • The individual laminations in an electric motor were at one time individual pieces. I've never had an EV motor apart, but I would guess that they contain on the order of 500-1000 laminations. (Of course, the alternator in a standard IC vehicle likely contains 200 laminations.) – Daniel R Hicks Jan 13 '18 at 22:13
  • `just as the wire and the insulation of a cable are to be counted separately.` Wire is manufactured in bulk and incredibly cheaply, a car factory doesn't put insulation a wire, assemble the laminations in an electric motor, or put the insulator on a spark plug. If it arrives in the car factory in one piece it is almost certainly a "part". If it would be replaced it is a "part". Let's not forget that any cast metal item was, once, just a molten load of atoms from several bought in ingots that were allowed to cool in a mould - but we don't count the atoms, we don't *even* count the ingots. – Grimm The Opiner Jan 15 '18 at 11:32
  • @GrimmTheOpiner: For the statement to make sense in the context of the article, manufacturing steps that happen *before* the cable or otherwise pre-assembled component arrives at the car factory have to be taken into account. – O. R. Mapper Jan 15 '18 at 11:38

2 Answers2

16

Yes, electric cars are much simpler.

You completely eliminate the combustion cycle - so no fuel delivery system (fuel pump, secondary fuel pump, fuel lines, fuel filter, fuel rail, injectors), no air intake (filter, mass air flow sensor, etc), the cooling system (radiator & hoses, fluid tank, water pump) is eliminated and the heater core is replaced by electric heater.

Electric motors are much simpler too. No crank or cam shafts, no pistons, no flywheel. Here is how AutoWeek puts it:

Compared with a typical four-cylinder internal combustion engine, which has hundreds of moving parts, the electric motor is very simple. Motors have a rotor, stator, armature, commutator, windings and bearings.

Electric motors deliver very high torque so no need for transmissions. From the Road&Track article:

Electric motors have a much larger RPM range than the typical internal combustion engine. And unlike a gas or diesel engine, an electric motor makes its best power output over an incredibly broad RPM range. So instead of packing the car with numerous gears to keep the engine in its happy zone, designers of electric cars just pick a gear ratio that provides a good compromise between acceleration and top speed. And with the typical electric motor capable of sustaining 20,000 RPM, the top speed often isn't even a limiting factor.

Electric cars also don't have differentials or a drive shaft. Instead, each wheel that's powered gets its own motor and torque is controlled by electronics. That's called "electronic differential".

Here is how Wikipedia describes them:

The classical automobile drivetrain is composed by a single motor providing torque to one or more driving wheels. The most common solution is to use a mechanical device to distribute torque to the wheels. This mechanical differential allows different wheel speeds when cornering. With the emergence of electric vehicles new drive train configurations are possible. Multi-drive systems become easy to implement due to the large power density of electric motors. These systems, usually with one motor per driving wheel, need an additional top level controller which performs the same task as a mechanical differential.

ventsyv
  • 7,136
  • 2
  • 27
  • 44
  • 3
    A minor nitpick - an electric engine still produces waste heat, so cooling system is not eliminated, although it might be less powerful than on an internal combustion engine with comparable power output. – Danila Smirnov Jan 12 '18 at 07:03
  • 1
    The first part seems obviously biased since you need a battery; how many parts does that have... And you also need cooling for an electric car as well. So -1 just for uncritically citing that. – Fizz Jan 12 '18 at 08:43
  • 1
    Cooling is also applied to the battery and some (Tesla) use liquid cooling: https://www.technologyreview.com/s/421913/the-electric-cooling-battery-test/ – Fizz Jan 12 '18 at 08:48
  • Teslas have differentials, they use one engine per axle not an engine per wheel so a differential is needed https://cdn.shopify.com/s/files/1/0196/5170/files/Red_Teslas_schematic_grande.jpg?v=1496610796 – Rsf Jan 12 '18 at 10:11
  • @Fizz - "How many parts does a battery have?" I just popped open my iPhone and replaced the battery this week. The answer I see, looking at that battery I removed, is "zero." – PoloHoleSet Jan 12 '18 at 14:52
  • Most electric cars still have a differential, only top end stuff has a motor per wheel. But still a lot less parts mh – RemarkLima Jan 12 '18 at 18:41
  • I'll adjust the answer. Thanks for the input – ventsyv Jan 16 '18 at 15:59
7

The claim vastly depends on what you consider parts... and even on the model of car in question. Given that Tesla S used AA-sized cells for its battery pack... how many of those did it take to power the car? (I bet more than 200.)

Straubel pointed to the wide variety of lithium-ion battery cells—the parts of a battery pack that actually store energy—that the company is testing. This included a row of small cylindrical cells about the size of AA batteries—the kind Tesla uses in the Model S.

By choosing smaller, cylindrical cells, Tesla saved on manufacturing costs—their costs have been driven down by economies of scale for the laptop industry, for which the cells were developed.

Count them in this photo if you want (source):

enter image description here

(The source actually did that for us: 444 or 384 cells depending on pack model: 85kWh or 60kWh.)

The same is true for Tesla Model 3 basically, although a slightly thicker and taller (custom) cell dubbed 2170 is used instead of the industry statndard 18650.

The advantage claim is a bit more believable for moving parts (as covered in the other answer).

Also Toyota says

A single car has about 30,000 parts, counting every part down to the smallest screws.

So while one may make valid claims about the superiority of electric cars in various ways, the comparison quoted in the question is just not believable to me.

Likewise, an (academic) source says:

A typical car requires 20,000 to 30,000 parts, each usually manufactured in a different facility and by different company groups.

HBR does have slightly different view:

The typical car contains about 2,000 functional components, 30,000 parts, and 10 million lines of software code.

So maybe the OP's page is talking about these "functional components" as opposed to parts... but from that HBR page I can't tell how these are defined or counted.

I don't know if any car-industry standards exist for counting "functional components", but a textbook on such decompositions in general (using cars as example), essentially says the number is arbitrary and not related to manufacturing:

enter image description here

Figure 7.4 exhibits the black-box model of a car. It corresponds with the driver’s perspective on a car. So, the driver is the using system and the car is the used system. Through changing the values of the input variables (e.g., the position of the steering wheel) the driver is able to change the values of the output variables (e.g., the direction of the car). Ideally, as we have seen, the (transfer) function is a mathematical relationship between the input variables and the output variables. However, for most concrete systems, e.g. for cars, this is hardly possible. Therefore, in practice, the notion of function is a rather informal, loosely defined notion of what kinds of (functional) behavior can be caused through manipulating the input variables. In this respect one also often uses the term “functionality”, a term that we prefer to “function”. If the transfer function is too complicated to understand, the technique of functional decomposition can be applied, through which the black-box model of a system is replaced by a structure of submodels of with more readily understandable functions. Figure 7.4 shows a possible decomposition of a car. If a component is still too complicated, it can be further decomposed. The exhibited decomposition of a car could be very helpful for a driving instructor to explain its functionality to a new student. Note, however, that the knowledge that one acquires about a system by means of functional decomposition is only functional knowledge, nothing less and nothing more. The only thing one does in functionally decomposing is help explain the functionality of a system, as we showed for the car example. The idea that functional decomposition ultimately leads to knowing the construction of the system is a widespread misunderstanding.

To elaborate this, a BB model is a purely conceptual division of the function or functionality of a system, independent of its construction and operation. Therefore, one can make virtually any decomposition one likes, and one can freely add or remove functional components.

(From J. Dietz Enterprise Ontology: Theory and Methodology, Springer, 2006).

Britannica has a similar presentation

enter image description here

The stuff that ventsyv is talking about is based on these aggregate functional notions rather than actual part counts. In this view a battery is one thing regardless how many cells it has etc.


Finally, does any of this overall part (or subsystem) counting matter in economic terms? I think the answer is no, because a 95-page UBS comparison of internal combustion and electric vehicles never mentions total (or component/subsystem) part count as an economic advantage (either way). It does mention moving/powertrain parts though (extensively):

Mechanical complexity is much lower [in EVs], whereas electronic complexity is higher. We counted 24 moving parts in the Bolt's powertrain, versus 149 in the Golf. The powertrain electronics content is $4k higher on the tier-1 level, motor included.

enter image description here

And it also mentions spare/replaceable parts being fewer in EV's:

Also, revenues from the lucrative spare parts business, which accounts for ~20% of EBIT, are likely to drop by ~60% in the long term in an EV world. However, this scenario is several decades away.

And again (later):

A separate point not to be ignored: Because EVs have much fewer moving and wearing parts, the attractive spare parts business, which represents ~10-15% of an OEM's EBIT, is likely to shrink considerably long-term. However, this should take another 15-20 years longer, due to the replacement cycle of the existing car parc

and one more time

The Bolt is almost maintenance-free. Not only do fewer parts need to be replaced over the car's life, it also does not require a regular change of fluids, such as engine oil. On our analysis, the after-sales revenue pool could drop by ~60% or >$400 per vehicle per year. This should pose a major challenge for dealerships, which typically generate >40% of their gross profit pool in service and maintenance.

enter image description here

But the overall part (or component/system) count at assembly seems irrelevant, at least at this depth of analysis... which is still quite substantial.

The power train is considered cheaper overall in EVs (this is currently offset by the battery in the figure 6 above).

enter image description here

The Bolt's powertrain is much simpler than the Golf's from a mechanical point of view:

  • The e-motor itself is much less complex than the combustion engine. Bearings aside, there are only three moving parts. Modern e-motors are brushless, ie, maintenance-free. The Golf's 4-cylinder engine has 113 moving parts. On top, spark plugs need to be replaced and engine oil needs to be changed regularly.

  • The combustion engine has a limited usable rotation range, between c800- 6,000 rpm. Also, its torque is not constant over the usable rpm range (unlike the e-motor). Therefore, a complex gearbox and clutch (or torque converter) are needed. The Golf's 6-speed automatic transmission has 27 moving parts. Gearboxes and clutches also wear. After mileage of 150k kilometres, gearbox replacements begin to rise significantly. In contrast, the Bolt has a very simple single-speed gearbox with only four gear wheels. We expect no maintenance or replacement to be required over the life of the car.

This is of course translates into an assembly advantage for the power train:

The motor in a BEV replaces the engine and transmission in an ICE vehicle and will contain a significantly lower number of moving parts. For instance, we expect that electric vehicles will have 6-7 bearings in the drive module (e-motor and mini gearbox) compared to 40-50 bearings in a traditional ICE. [...] We also expect significantly less machining will be required for the e-powertrain vs. conventional ICEs. Our channel checks indicate up to 80% of the cutting tool work needed to manufacture a car happens in the combustion engine. Significantly less machining is required for the e-motor.

The latter is because electric motors' rotor and stator are made of (electrically isolated) stamped laminates (making them of one block is a no-no because of Eddy current losses). They also talk extensively about the battery in EVs and what makes it expensive. They do note that currently there's a substantial overhead of assembling the Bolt battery pack (in Bolt is made of 288 pouches):

Finally, economies of scale and the learning curve in cell and pack assembly should bring further savings. In today's $3,600 pack mark-up, only ~25% relates to materials used. This points to high fixed costs in a sub-scale production environment. We assume a contribution from economies of scale of $10/kWh.

But with all this detailed analysis, it never gets to overall part (or system) count being an economic advantage. They only say that

General assembly is a similar process for BEVs and ICE vehicles.

Fizz
  • 57,051
  • 18
  • 175
  • 291
  • The question meanwhile contains a section about what is presumeably considered a part for the context of the cited statement. If it only applies to some electrical car models, I think that pretty much means the claim as such is false (i.e. electrical cars in general do not inherently have less parts than combustion engine cars). The distinction of moving parts, on the other hand, is rather beside the point, as the claim is related to car manufacturing - so even parts that are not supposed to move relatively to one another in the finished car might be handled separately during manufacture. – O. R. Mapper Jan 12 '18 at 11:34
  • @O.R.Mapper: Yes but Tesla assembles their own batteries, granted it is in a different plant than where the cars are finally assembled. So I'm not sure what is your point, ultimately. Which parts should we be counting? – Fizz Jan 12 '18 at 11:38
  • The cited article claims that less workforce is required to produce an electric car than a combustion engine car, based on the claim that the former contains less (by a factor of 5!) parts than the latter. I am trying to find out whether that claim is completely true (i.e. somehow, electric cars are assembled from considerably less single parts than combustion engine cars), whether that claim is true only due to coinciding external factors (e.g. because plenty of the originally single parts are typically delivered pre-assembled by external suppliers for electric cars, when their ... – O. R. Mapper Jan 12 '18 at 11:43
  • ... equivalents in combustion engine cars are typically built in-house, and the article only counts the parts that were separate some time in the car factory itself), or whether the claim is false even then. – O. R. Mapper Jan 12 '18 at 11:44
  • @O.R.Mapper: ok, but that workforce aspect is a different claim. – Fizz Jan 12 '18 at 11:46
  • The workforce aspect has been in the cited article since I posted the question. I take it you have had a peek into the article to see the context of the cited claim. (Indeed, the claim that more parts equals more workforce is separate and not necessarily literally true, but it's not the claim I'm asking about here. It's just a statement from which we can derive what definition of "part count" we need to apply here, namely parts that were separate at some point during the production process.) – O. R. Mapper Jan 12 '18 at 11:47
  • @O.R.Mapper: Mkay, but asking two questions in one, which may require different answers [is a pretty bad idea](https://meta.stackexchange.com/questions/222735/can-i-ask-only-one-question-per-post). I chose to focus on the one you've spent most of your space on, including the quote and the title. – Fizz Jan 12 '18 at 11:51
  • I have edited the question again to clearly indicate the primary "part count" I am interested in, based on the cited article. I think I have been asking only one question, but I did already anticipate that (as seems to be the case with *many* [skeptics.se] questions) answers might be more nuanced than ending up with a clear "true" or "false", exactly because the original claim is not necessarily very clearly defined, as you and many commenters correctly pointed out. – O. R. Mapper Jan 12 '18 at 12:00
  • Counting the cells of a battery pack as individual parts is like counting a bolt-washer-nut as 3 parts. – ventsyv Jan 12 '18 at 14:27
  • 3
    @ventsyv: which they are (3) if you have to manufacture them. Likewise I can argue the car is 1 part because that's how I buy it from the auto dealer. – Fizz Jan 12 '18 at 14:33
  • Most thorough answer! – Ernie Jan 15 '18 at 23:32