Whenever I see an explanation of an old technology, I don't wonder "How do that work" but rather "How did they even build this"
How did they make vacuum and be sure it was indeed a vacuum ?
How did they melted that metal alloy into the right shape ?
How did they built those precise mirrors ?
How many times did they fail before succeeding ?
How did they lubricate and seal properly all components ?
I know I could explain 80% of those inventions easily by studying them for a few days each. I know I would never be able to build 1% of those machines myself, I'd constantly run into technical issues and it'd take me years for each
Author here. The beam engine created power from steam, and was a key part of the early Industrial Revolution. The article is a deep dive into the engine: how it works, the history behind it, and the engineering tradeoffs the builders faced. There are quite a few interactive figures that I hope make the concepts easier to see.
Despite all the details, you still left out a lot. Safety valve, pressure gauge(s), regulation of the water pump, water level gauge for the boiler, lubrication (you only have two of the glass/metal can reservoir types for the main axle), and then maybe also the valve gear (the control mechanism for the valves which takes into account where in the cycle the machine is and maybe also how fast it is going, how fast it should be going, and in which direction). And a way to drain water from the bottom of the cylinder + maybe a way to make the steam as dry as possible.
And each of those could be expanded to much more than the page you already made.
Things are almost always a lot more complicated than they seem. It seems so simple -- "water + heat makes steam, steam pushes piston, back and forth motion makes rotary motion" -- but it is really anything but.
Excellent work, glinscott. Thank you for sharing! I am about to send this to all my pals who are parents to schoolkids (um, and also pals who are still, like me, basically a schoolkid :p XD).
Building things up from first principles, and ideally also tactile / visceral examples, is my absolute favourite genre of explanations. Alas, a rarefied genre, because it is so much harder to do well (without doing disservice to the learner's intellect and ability) than expert talking to (presumed) expert. So thanks for contributing to it.
People like George Polya [0], Richard Feynman [1], Grace Hopper [2] are exemplars of this tradition.
(And I too try in my own meagre way through my blog and source code, albeit it's all plain text because visual explainers, animated ones that too are ridiculously difficult. So I have some idea of how much of a labour of love your piece is. <Claps hands, Whistles, Doffs hat>.)
You are welcome! I agree, it's a really nice way to learn about something. Digging into the transfer of power with belts may have been one of my favorite parts here, even though it's not really steam directly :).
The key thing to understand about early steam engine technology is that they didn't have steel. Good steel in quantity wasn't available until the Bessemer converter around 1880. The converter itself is simple. It took about 10,000 melts to get the metallurgy right, and analytical chemistry to get consistent ingredient mixes.
Without steel, the early engines were cast or wrought iron. So everything was very low pressure, or blew up. That's why "atmospheric" engines such as Newcomen's were built. That doesn't use steam pressure at all. It just uses condensing steam to create a partial vacuum so atmospheric pressure can push the piston. The technology got stuck there for 75 years.
Higher pressure engines made of iron were tried. They blew up frequently at first. This led to a useful institution, The Hartford Steam Boiler Inspection and Insurance Company, founded in 1866.[1] They still exist, owned by Munich Re. Hartford Steam Boiler had a tough approach to insurance. They'd insure risky things such as steamboats, but only after their inspectors had inspected them. Their insurance policies gave them the right to inspect at any time, which they used. The usual arrangement was that they inspected something, produced a list of things which had to be fixed, and came back for a second inspection after the fixes. Only then did they provide coverage. Steamboats mostly stopped blowing up.
(Today, Hartford Steam Boiler also sells business interruption insurance against cyberattacks, and even AI liability insurance. They probably still inspect first. That may be the good path to AI safety - liability lawyers suing for damages on one side, and an insurance company into tough inspections on the other.)
> This push on every wall is pressure, and we will measure it in atmospheres, multiples of the ordinary pressure of the air around us.
No, don't do that. Because the pressure of the atmosphere isn't constant. It varies with weather and how high you are.
Instead use Pascal (Pa) or Hectopascal (hPA). I really pays back to use SI units for physical things. Suddenly you have formulas without weird constants, like 1 Pa = 1 N/m^2
It's a sad thing that new articles still propose units that are (since many, many years) outdated: the metric system is from 1793, the SI is from 1960.
BACK then people didn't knew better and used the varying atmosphere as base. But today ?!?!?
One of the unexpected pleasures of having a 3 year old has been learning far more than I ever planned to about steam locomotives. It turns out steam engineering is fascinating, this article emphasized it for me even more. It really amazes me how much ingenuity there is in just connecting the right shapes and joints up so that the right thing happens at the right time, creating perfect cycles, regulators, etc. Seeing it explained in an iterative way like this is fantastic because you can understand what problem is solved by each part. I would love to see a piece like this strictly about locomotives.
One of my treasured toys - probably from a few years older than your 3 year old though - was a rocking cylinder steam engine my father gave me. It was a present from his father as part of a steam powered boat they built together probably in the late 1940s.
I bought one for a friend's 5 year old a while back and became BEST UNCLE EVER!
The run well on compressed air (like an electric car tyre inflator or small air compressor), or for a bit more high temperature danger they run great attached to the steam wand of an espresso machine. When I was a kid dad built me a steam boiler out of a steal aerosol can and some copper tubing with a folded brass sheetmetal alcohol burner. That's not a 5 year old friendly thing though, from memory I'd have been maybe 10 or 12 when he started letting my light alcohol on fire in that garage or backyard. (Modern parenting may disapprove...)
Thank you! I have a few ideas for the next article. Locomotives are up there on the list, but they are the final challenge of steam engineering, there is so much to cover.
Also considering the origins of precision - the intertwined evolution of machining tools and measuring systems that allowed these machines to be built.
RE: 3 year old - yes, mechanical systems are wonderful for children (and adults). There is something incredibly satisfying about watching them in motion doing work.
High performance "modern" steam engines are fascinating, they're not necessarily simpler than their internal combustion successors, just different. Monotube boilers create superheated steam that pushes 1200°F, which is right around the same temperature as the exhaust gasses in truck's diesel engine pulling a large hill at maximum power. Hot enough to melt aluminum save for the presence of oil jets to cool the piston and a jacket of cooling water surrounding the cylinder. However, since there's no combustion happening in the cylinder of a steam engine, you don't have any cooling water surrounding it--that would be totally counterproductive. So steam engines operate in a thermal regime that is far more extreme than internal combustion engines. Consequently, bores are often cut slightly tapered to account for the temperature gradient across the stroke, in a double acting engine the cylinder bore has a slight hourglass shape, otherwise the thermal expansion would cause the piston to bind at the ends.
For a fun rabbit hole to dive down, check out Doble steam car technology or the Besler airplane engine.
Thanks for the pointers! That is an area I've not explored much - reciprocating steam using modern technology. Using modern electronics, sensors and valves to optimize steam expansion would be a fun project. Perfect cutoff every time :).
Ever hear the phrase "Balls out" when referring to pushing something to the max? Comes form the balls of the centrifugal engine governor. At maximum speed the balls extended all the way outward giving us the term "Running balls out."
edit: I once said this phrase at a work meeting and someone objected to my use of a crude term which I then explained much to everyone's surprise. They thought it had something to do with testicles.
While fun and satisfying, this is extremely unlikely to be true.
The first attested use is from 1945, on the nose of a P-47 Thunderbolt flown by Captain Milton Thompson of the 509th Fighter Squadron. And his art was exactly what you'd expect: a charging bull with the words "balls out".
Definitely a fun note :). I think I heard it first from a Jay Leno youtube video where he was showing his amazing collection of steam powered machinery.
This is from the article, but governors also played a role in kicking off modern control theory!
"""
In 1868, James Clerk Maxwell studied when these oscillations grow or die away in his paper “On Governors”[https://www.damtp.cam.ac.uk/user/gold/pdfs/teaching/old_lite...]. This became one of the beginnings of modern control theory.
"""
Example, throttle controls on a Dogulas DC3. All engine controls forward is the typical convention for full power. (At sea level, neglecting the altitude effect on required engine fuel/air mixture.)
There is also a steam engineers term called "high-balling" which referred the little ball in the steam pressure gauge used in the early days. So high-balling was travelling at high speed.
Euphemisms are weird, what was originally intended as a way of saying a thing without actually saying that thing, becomes the unspeakable thing in the first place. and the original meaning fades.
The one I like is why do we use birds to refer to female anatomy? I also heard something the other day and had an epiphany, but the sort you can't really tell anyone. "Oh... it means pouch that makes so much more sense, it's not a cat."
Ancient Greeks had a steam engine. They called it a Hero's engine. They used it as a toy.
They couldn't use it as anything else than a toy. The reason they couldn't because they didn't have the metallurgy and the machining to create pistons, valves, etc.
There's an important lesson that can be applied to every technology, I think: technological progress is always iterated, never revolutionized. If your novelty - a programming language, a framework, a methodology, whatever - doesn't stand on the shoulders of the giants, then you have simply nothing to apply this novelty to.
That was a reaction engine, something like a steam rocket confined to rotate in a circle, and not a positive-displacement engine: https://en.wikipedia.org/wiki/Aeolipile
They couldn't use it as anything else than a toy. The reason they couldn't because they didn't have the metallurgy and the machining to create pistons, valves, etc.
IMHO they definitely had the processes to basically add a PTO to one, but it was probably dismissed as impractical.
This reminds me a lot of https://ciechanow.ski, down to the body typeface. (IBM Plex Sans?) Also, the SI units everywhere here make me beam. Thank you :)
100% - his articles were a major inspiration, and first line in the credits:
- Thanks to Bartosz Ciechanowski, whose exceptionally well written and designed articles inspired this one.
Fred Dibnah's BBC series "Age of Steam" is a fine overview of the development of steam engine technology and the industrial revolution in the United Kingdom.
Quinn (Blondihacks) on youtube is a model engineer machinist and has, IMO, the most outsider-watchable content on building boilers and steam engines like this.
One more: a little engine I'd never encountered before called a wobbler is, apparently, the hello world of machining model engines: https://www.youtube.com/watch?v=HrHDrIms7-0
I'm only ~5 minutes into that first video and every single thing she has said so far rings so stupid true that its almost painful hearing it said so bluntly. Making things takes a lot of time, requires lots of fundamental skills, is expensive to do yourself, generally isn't going to be better than something other people are selling, and should be fun the entire time you're doing it.
This should be the comment that people read on this post. Watch these videos.
There’s also the amazing Crossness, which does have occasional steaming days. And if you’re in Manchester the Museum of Science and Industry has a great collection of small and large engines which they run
The first steam engines were vacuum engines, fill a cylinder full of steam and as it condenses it sucks the piston down, Watt's improvement and patent was on a separated condenser unit so you did not have to reheat the whole cylinder each time. This doubled the efficiency, that is, reduced the amount of coal you had to burn by half. But Watt would not sell manufacturing rights to his patent he sold operating licenses, If you had an engine that utilized Watts patent idea you owed him half the price of the coal you did not have to burn by using it. Consequently there were a lot of "pirate" engines, the most common method of working around Watts patent was to claim that your condenser was integrated into(as opposed to separated from) the cylinder some how.
Yes, great details! Not sure if you've read it, but "Richard L. Hills, Power from Steam: A History of the Stationary Steam Engine (1989)." is a truly wonderful overview, that really digs deep into the business. It has many letters directly from Watt, including the problems they had getting these huge engines built (before any decent roads!).
Engineering drawings of small model engines used in model aircraft making. These designs are intended for manufacturing and practical operation, rather than merely for display, and can be used to build fully functional engines.
Cool! If you're in New Zealand, there is a massive, two story Double Woolf beam engine at MOTAT museum in Auckland, along with an array of other steam driven engines that they run from time to time.
Awesome! It's pretty amazing all the concepts that need to come together for even these very old engines. I found I built a better understanding of what was going on as I built out the animations as well.
RE: units - Good catch - there is almost certainly a better way to phrase this one. I always tried to give physical examples to make things more concrete. But who knows, maybe the bathtub will take off as a unit :).
One nit: All three animations that show the governor and the flywheel have the governor turning the wrong way, causing the gears that join them to do physically impossible things.
Oh my. Yes, you are totally right. This was tricky because I needed to merge two different animations (the governor moves independently from the rotation of the wheel), and I seem to have messed that up.
When I clicked on the link, I first thought it was going to be about, well, steam engines. Then I realized this is Hacker News, so it's probably about the Erlang VM. Turns out my first thought was actually the correct one!
Glad to not be the only person to expect Erlang instead of steam.
It was a mix of techniques, this article has been in progress for quite a while :). The first part was modeling all the relationships for an accurate beam engine in onshape (great cad tool). Then, I built a custom exporter to cleanly export it to three.js. This ensured physical realism for the hero of the article. All of these are linked from the article.
For the figures, I had prototyped some by hand, and it was slow going. Fable was the first LLM I found that was actually able to understand and model mechanical linkages correctly enough to be useful. I built a style guide, and iterated many times on each figure and animation until we arrived at something useful. There are also quite a few backing tests to ensure physical correctness of the animations (as far as reasonably possible :).
How did they make vacuum and be sure it was indeed a vacuum ?
How did they melted that metal alloy into the right shape ?
How did they built those precise mirrors ?
How many times did they fail before succeeding ?
How did they lubricate and seal properly all components ?
I know I could explain 80% of those inventions easily by studying them for a few days each. I know I would never be able to build 1% of those machines myself, I'd constantly run into technical issues and it'd take me years for each
Not just in terms of presentation, but also quality.
Amazing!
And 100% inspired by Bartosz's articles - he is the first line in the credits section.
And each of those could be expanded to much more than the page you already made.
Things are almost always a lot more complicated than they seem. It seems so simple -- "water + heat makes steam, steam pushes piston, back and forth motion makes rotary motion" -- but it is really anything but.
Building things up from first principles, and ideally also tactile / visceral examples, is my absolute favourite genre of explanations. Alas, a rarefied genre, because it is so much harder to do well (without doing disservice to the learner's intellect and ability) than expert talking to (presumed) expert. So thanks for contributing to it.
People like George Polya [0], Richard Feynman [1], Grace Hopper [2] are exemplars of this tradition.
[0] https://www.youtube.com/watch?v=h0gbw-Ur_do
[1] https://www.youtube.com/watch?v=EKWGGDXe5MA
[2] https://www.youtube.com/watch?v=ZR0ujwlvbkQ
(And I too try in my own meagre way through my blog and source code, albeit it's all plain text because visual explainers, animated ones that too are ridiculously difficult. So I have some idea of how much of a labour of love your piece is. <Claps hands, Whistles, Doffs hat>.)
The key thing to understand about early steam engine technology is that they didn't have steel. Good steel in quantity wasn't available until the Bessemer converter around 1880. The converter itself is simple. It took about 10,000 melts to get the metallurgy right, and analytical chemistry to get consistent ingredient mixes.
Without steel, the early engines were cast or wrought iron. So everything was very low pressure, or blew up. That's why "atmospheric" engines such as Newcomen's were built. That doesn't use steam pressure at all. It just uses condensing steam to create a partial vacuum so atmospheric pressure can push the piston. The technology got stuck there for 75 years.
Higher pressure engines made of iron were tried. They blew up frequently at first. This led to a useful institution, The Hartford Steam Boiler Inspection and Insurance Company, founded in 1866.[1] They still exist, owned by Munich Re. Hartford Steam Boiler had a tough approach to insurance. They'd insure risky things such as steamboats, but only after their inspectors had inspected them. Their insurance policies gave them the right to inspect at any time, which they used. The usual arrangement was that they inspected something, produced a list of things which had to be fixed, and came back for a second inspection after the fixes. Only then did they provide coverage. Steamboats mostly stopped blowing up.
(Today, Hartford Steam Boiler also sells business interruption insurance against cyberattacks, and even AI liability insurance. They probably still inspect first. That may be the good path to AI safety - liability lawyers suing for damages on one side, and an insurance company into tough inspections on the other.)
[1] https://www.munichre.com/hsb/en.html
For other animations I put together a style guide and worked together with Fable to build them.
No, don't do that. Because the pressure of the atmosphere isn't constant. It varies with weather and how high you are.
Instead use Pascal (Pa) or Hectopascal (hPA). I really pays back to use SI units for physical things. Suddenly you have formulas without weird constants, like 1 Pa = 1 N/m^2
It's a sad thing that new articles still propose units that are (since many, many years) outdated: the metric system is from 1793, the SI is from 1960.
BACK then people didn't knew better and used the varying atmosphere as base. But today ?!?!?
Something similar to this: https://en.wikipedia.org/wiki/Oscillating_cylinder_steam_eng...
And you can still get them reasonably inexpensively: https://www.aliexpress.com/item/1005007964198565.html
I bought one for a friend's 5 year old a while back and became BEST UNCLE EVER!
The run well on compressed air (like an electric car tyre inflator or small air compressor), or for a bit more high temperature danger they run great attached to the steam wand of an espresso machine. When I was a kid dad built me a steam boiler out of a steal aerosol can and some copper tubing with a folded brass sheetmetal alcohol burner. That's not a 5 year old friendly thing though, from memory I'd have been maybe 10 or 12 when he started letting my light alcohol on fire in that garage or backyard. (Modern parenting may disapprove...)
Also considering the origins of precision - the intertwined evolution of machining tools and measuring systems that allowed these machines to be built.
RE: 3 year old - yes, mechanical systems are wonderful for children (and adults). There is something incredibly satisfying about watching them in motion doing work.
For a fun rabbit hole to dive down, check out Doble steam car technology or the Besler airplane engine.
How do you make those animations?
https://en.wikipedia.org/wiki/Centrifugal_governor
edit: I once said this phrase at a work meeting and someone objected to my use of a crude term which I then explained much to everyone's surprise. They thought it had something to do with testicles.
The first attested use is from 1945, on the nose of a P-47 Thunderbolt flown by Captain Milton Thompson of the 509th Fighter Squadron. And his art was exactly what you'd expect: a charging bull with the words "balls out".
This is from the article, but governors also played a role in kicking off modern control theory! """ In 1868, James Clerk Maxwell studied when these oscillations grow or die away in his paper “On Governors”[https://www.damtp.cam.ac.uk/user/gold/pdfs/teaching/old_lite...]. This became one of the beginnings of modern control theory. """
https://www.douglasdc3.com/dc3throt/dc3throt.htm
Going balls to the wall means pushing the throttle and mixture all the way in to the firewall/instrument panel for maximum power.
The choice of wording makes me suspect that they intended both entendres. ;)
The one I like is why do we use birds to refer to female anatomy? I also heard something the other day and had an epiphany, but the sort you can't really tell anyone. "Oh... it means pouch that makes so much more sense, it's not a cat."
They couldn't use it as anything else than a toy. The reason they couldn't because they didn't have the metallurgy and the machining to create pistons, valves, etc.
There's an important lesson that can be applied to every technology, I think: technological progress is always iterated, never revolutionized. If your novelty - a programming language, a framework, a methodology, whatever - doesn't stand on the shoulders of the giants, then you have simply nothing to apply this novelty to.
They couldn't use it as anything else than a toy. The reason they couldn't because they didn't have the metallurgy and the machining to create pistons, valves, etc.
IMHO they definitely had the processes to basically add a PTO to one, but it was probably dismissed as impractical.
Glad you enjoyed the units :).
https://www.youtube.com/watch?v=kl_UA36ouzM&list=PL2vJ5Cg-wl...
Highly recommend this primer titled "The Model Engineering Learning Curve" https://www.youtube.com/watch?v=Ps_BQJEMnGA
She's been building a model locomotive for years now, there are over 100 videos: https://www.youtube.com/watch?list=PLY67-4BrEae-xhrvZX33gPk4...
One more: a little engine I'd never encountered before called a wobbler is, apparently, the hello world of machining model engines: https://www.youtube.com/watch?v=HrHDrIms7-0
It's a great hobby. Manual machining is totally different from writing code, but you get into a similar state of flow.
This should be the comment that people read on this post. Watch these videos.
https://www.youtube.com/@Wreck2Restored
It wasn't a vacuum engine, as per post the water just cooled the cylinder. And it wasn't half of the coal, it was a third.
(Sorry I am pasting my old comment.)
Some Model Engineering related resources:
Engineering drawings of small model engines used in model aircraft making. These designs are intended for manufacturing and practical operation, rather than merely for display, and can be used to build fully functional engines.
https://www.modelengineeringwebsite.com/Beam_Engine_drawings...
https://modelengineeringwebsite.com/Classic_ME_beam_engine.h...
https://outerzone.co.uk/plans.asp?cat=Engines&Xcardsperpage=...
https://modelenginenews.org/midge/index.html
https://modelengineeringwebsite.com/Midget_gas_engine_1.html
https://www.adriansmodelaeroengines.com/catalog/product.php?...
And https://modelengineeringwebsite.com/Classic_ME_beam_engine.h... is almost exactly the CAD model behind this beam engine. Designing that on paper would have been quite a challenge.
One of my favorites. Good read.
This is sending me :)
> One cup of water becomes roughly 400 litres of steam, enough to fill two bathtubs
Three units of measurement in one sentence, my favorite unit is bathtubs.
RE: units - Good catch - there is almost certainly a better way to phrase this one. I always tried to give physical examples to make things more concrete. But who knows, maybe the bathtub will take off as a unit :).
One nit: All three animations that show the governor and the flywheel have the governor turning the wrong way, causing the gears that join them to do physically impossible things.
Fix shall be deployed shortly!
Glad to not be the only person to expect Erlang instead of steam.
For the figures, I had prototyped some by hand, and it was slow going. Fable was the first LLM I found that was actually able to understand and model mechanical linkages correctly enough to be useful. I built a style guide, and iterated many times on each figure and animation until we arrived at something useful. There are also quite a few backing tests to ensure physical correctness of the animations (as far as reasonably possible :).