Reverse engineering the 59-pound printer onboard the Space Shuttle

http://www.righto.com/2024/08/space-shuttle-interim-teleprinter.html

by chmaynard • 2 years ago

340 96 2 years ago

righto.com

kens15• 2 years ago

Author here if anyone has questions...

hobo_in_library• 2 years ago

This is really cool. How did you discover all this information?

kens15• 2 years ago

I had to go to the Library of Congress to get some of the information...

ASalazarMX2• 2 years ago

Was this printer radiation-hardened? As others have pointed out, a commercial dot-matrix printer would have saved many valuable kilos of weight. There must have been other priorities besides weight.

kens15• 2 years ago

No radiation hardening. Their main priority was building a system in 7 months that meet their toxicity and flammability standards. The original military teleprinter weighed 100 pounds. They cut it down to 59 pounds, so that's a win, I guess.

magnat• 2 years ago

Is the drum rotating smoothly at constant RPM, or is it stopping briefly (using gears [1] similar to those found in analog movie projector) when hammers hit the paper to prevent them from tearing it?

[1] https://en.wikipedia.org/wiki/Geneva_drive

jmole• 2 years ago

it's stopping, like all printers of the era, e.g. https://youtu.be/A_vXA058EDY?&t=41

edit: I see now you were asking about the drum, rather than the paper

tim333• 2 years ago

It looks like a regular line printer where the drum keeps spinning. See wikipedia:

Because the drum carrying the letterforms (characters) remains in constant motion, the strike-and-retreat action of the hammers has to be very fast. https://en.wikipedia.org/wiki/Line_printer

nyrikki2• 2 years ago

The platen and paper feed stops line for line.

the drum or similarly the chains used on the hammer printers were in constant motion.

I am just old enough that I had to repair both types at the beginning of my career. Although typically rebranded Data Products and other OEMs, which I am probably wrong but vaguely remember being a supplier for the DEC L20?(maybe). But different than this printer.

To these hammer action printer, the ribbon and the paper weren't even a consideration.

If you have a Newton's cradle, put a piece of paper between the inner balls and it will still mostly work if you release a single ball.

The high speed drum printers they typically had to rotate twice for each line (at minimum) so a 600lpm printer would have the drum rotating at about 1200rpm.

If you look at the video posted in another comment, you can see the ragged vertical alignment of the chars. IIRC that is why IBM preferred chains in their hammer printers, because the human eye was more forgiving of vertical misalignment compared to the vertical misalignment that was a natural result of the mechanical differences.

Edited to add link to video from page:

https://www.youtube.com/watch?v=EDeL15amsus

zelphirkalt2• 2 years ago

Why do you make it impossible to highlight text on the website?

EDIT: Apparently not the whole text cannot be highlighted, but only the initial lines of text.

slater• 2 years ago

Works for me? macOS, Firefox

eichin• 2 years ago

How much does the paper move when the hammer hits it? even with the padding of the ribbon (and having seen the video clips of it in action - ps. would love to see high-speed closeups if you get a chance) I'm not sure why the paper doesn't tear when the hammer hits it - or if it's a very short distance, does it instead smudge?

cbhl• 2 years ago

I think the paper gets sandwiched between the hammer on one side and the drum on the other side.

This would be not too dissimilar to how a typewriter would have worked, I think?

userbinator• 2 years ago

Less than 100 thou, I think. It's also a very rapid process, so the drum would have barely moved meanwhile.

kens15• 2 years ago

You'd think that hammering the paper against the spinning drum would result in a blurry mess at best, but the hammers are fast enough that it works. It's the same principle as a chain printer such as the vintage IBM 1403, where the chain is zipping along at 90.3 inches per second when it gets hammered.

dTal• 2 years ago

often printing thousands of lines per flight
with a Shuttle flight costing $27,000 per pound, putting the 59-pound teleprinter in space cost over $1.5 million per flight.

That budget calculation is just the weight of the printer itself. I am curious how much blank paper was flown to feed it, and how this was decided? From the font size shown in photo we can assume it was at least several more pounds of paper for "thousands of lines" to be true.

kens15• 2 years ago

The printer took a big roll of paper, maybe 5 inches in diameter (although I don't have it here to check). I was a bit surprised that they didn't trim the roll down to the minimum size necessary, but just used a full roll. I guess they didn't want to waste paper :-) Also, I'm surprised that they used cheap-looking yellow paper rather than white paper, but maybe there's a reason teletypes used yellow paper. (The paper hasn't yellowed due to age; it's yellow in contemporary photos.)

neaanopri• 2 years ago

I'm curious what the actual marginal cost of an additional pound on the shuttle is - surely you can't just divide?

metadat4• 2 years ago

How does a hammer strike cause only one character to be printed? For example, striking the leftmost hammer not also causing the neighboring character to also be printed.

jdlshore• 2 years ago

There’s 80 hammers, one for each character. They’re small.

skissane2• 2 years ago

and a ROM that holds its program code

Did you manage to dump the ROMs? (I understand from your article that article there are 3, a 4KB ROM on the CPU card, and two 8KB ROMs on the word processor card)

kens15• 2 years ago

I'd like to dump the ROMs, but they are soldered in and covered with conformal coating, so it would be moderately destructive to remove them and dump the contents.

chiph• 2 years ago

Greetings Ken. Thanks to you and Marc for another great dive into obscure NASA hardware.

They must have considered that the spinning drum had a gyroscopic effect on the shuttle. One of the reasons why the shuttle computers used tape drives and not hard drives was because they would have affected it's attitude in microgravity. Perhaps there wasn't enough mass there to be a problem, and also that it wasn't spinning 100% of the time like hard drives would have been?

I'm also surprised they didn't consider the Teletype Model 40. It was lighter, didn't have a spinning drum, came in a receive-only configuration (no need to hack off the keyboard), was ASCII-based, and was already flight qualified on military aircraft. Maybe my timing is off - I first saw one in 1984 and it was considered "new" then, so it might not have been available in 1981.

The Model 40 used a rotating belt with individual character pallets in it. It also used 80 hammers that would fire at the correct time to produce a printed character. You can see the pulleys for the belt spinning (one is behind the red plastic part on the left) in the video. It would also power down after a period of inactivity. One of the jokes you could play was to swap pallets around. They weren't tracked individually, but the belt had a flag that would indicate the start of the pallet sequence. So any monkey-business afterwards would result in M's for N's etc. being printed.

https://www.youtube.com/watch?v=whKVGefscro

ars• 2 years ago

The gyroscopic effect of a hard drive is easily removed by having two identical ones attached together back-to-back. You would anyway want two for redundancy.

Pinus• 2 years ago

Now I want to see a print sample! =)

johnklos• 2 years ago

There's a short video of the printer printing towards the bottom of the article. You can see the printing in that.

On the other hand, a Snoopy calendar printed on that would be cool :)

ASalazarMX2• 2 years ago

Wonder if it's as incredibly noisy as a regular dot-matrix printer. In the video it sounds almost pleasant.

beardyw• 2 years ago

I worked with line printers for many years and they were as noisy as hell. Pretty fast though.

SoftTalker4• 2 years ago

I'm pretty sure we had commercial/consumer dot-matrix printers in 1981. Something like that would have been much lighter and lower-power than a drum line printer.

treyd• 2 years ago

But would it have worked after the intense vibrations and acceleration of launch and in zero gravity?

cyberax2• 2 years ago

Why wouldn't it? Matrix printers are extremely robust.

michaelcampbell2• 2 years ago

Are they printing during launch?

Not sure what part of a dot-matrix required gravity to work; the tractor feeds for the paper (if they used that sort), and nothing in the print mechanism requires a "down" other than possibly paper "exhaust" collection?

lysace2• 2 years ago

Weren't they primarily Japan-made at the time? (I imagine that could present a political problem.)

darreninthenet• 2 years ago

Centronics (American...) built the first "true" dot matrix in 1970 although IBM built a line printer that was similar in the late 50s

metabagel• 2 years ago

https://news.ycombinator.com/item?id=39769157

And when shuttle was developed, printers barely existed. Both inkjet and laser desktop printers were introduced commercially 1-3 years before the shuttles first flight in 1981, and weren’t very reliable yet. Desktop printers still aren’t as reliable as a teletype or dot matrix printer. There’s a reason airlines use dot matrix for printing flight manifests at the gate.

Ink plotters, teleprinters, and fax machines ruled the world. But plotters are dreadfully slow at writing text. Radio fax machines may have been viable if they were rugged enough. But they probably weighed as much as the teletype and were much slower - only real advantage is printing diagrams and photos.

Mistletoe• 2 years ago

And imagine how cool that matrix printer sound would be ringing out against the walls of the shuttle in space. Cyberpunk as hell. I miss dot matrix printers so much.

https://youtube.com/watch?v=A_vXA058EDY

ralferoo• 2 years ago

A line printer sounds nothing like a dot matrix printer. They're much, much louder, and it's a continuous buzzing noise, like a chainsaw. Also, the output appears at a similar speed to a dot matrix doing a form feed.

While the sound of a dot matrix printer can be kind of therapeutic - I remember debugging by printing, where the sound of printing and the time waiting could be used to mentally debug the problem, so you almost knew where the bug was without even looking at the print out, a line printer is the complete opposite as it's so loud it's distracting and the output gets spat out so quickly you don't have much, if any thinking time!

jonathaneunice2• 2 years ago

In 1981, maybe, but there's a long time lag between design and flight for spacecraft. Cheap/light/sturdy dot matrix printers weren't yet available in the 1970s when the Shuttle was being designed. Nor had the idea of using commercial/off-the-shelf (COTS) components yet taken root. That would come years after the STS was already built and in service.

SoftTalker4• 2 years ago

But the story says the drum printer was selected at the last minute as an "interim" solution?

cyberax2• 2 years ago

I wonder why they didn't use a matrix printer.

Or even a thermal printer. Fax machines were ubiquitous at that time.

kens15• 2 years ago

NASA had toxicity and flammability constraints that limited what they could use.

ahoka2• 2 years ago

It’s in the article.

annoyingnoob2• 2 years ago

Mechanical parts likely still work, does not need an ink subscription.

cesarb• 2 years ago

Mechanical parts likely still work

Towards the end of this article, Ken says that the mechanical parts did not work:

"The printer had many mechanical problems, mainly because the rubber rollers had turned to liquid and gummed up the mechanism."

annoyingnoob2• 2 years ago

Now imagine a printer made of mostly plastic and a bunch of elastomers. Your subscription based Inkjet is not going to last even half as long.

rkagerer2• 2 years ago

The drum for the Shuttle teleprinter replaces 10 ASCII special characters with symbols that are more useful to the Shuttle, such as Greek letters for angles.

Specifically, the characters ;@[\]^!"#$ are replaced by θ~αβΔ.

Sigh... This ancient printer could render them, but apparently the Chrome browser on my modern Android cannot.

tialaramex2• 2 years ago

Somewhere between when this was originally written and when it was delivered to your browser the text here was destroyed.

I don't see how that's the fault of Chrome or Android.

Narishma• 2 years ago

It works fine in Firefox.

Stratoscope2• 2 years ago

Oddly enough, Chrome on my Samsung S24 Ultra renders them OK here on HN, but not on Ken's page. Chrome on Windows and Ubuntu and Safari on iOS are the opposite: OK on the page, not here.

Character encoding is such fun!

Update: Ken, if you see this, it's specifically the characters encoded with numeric entities that have the problem on my devices. I used the Chrome dev tools to replace these with the actual characters on my Windows and Ubuntu devices and they rendered OK.

skylanh• 2 years ago

Author font family: normal normal 14px Arial, Tahoma, Helvetica, FreeSans, sans-serif;

News YC font-family: Verdana, Geneva, sans-serif;

Per https://unicode.scarfboy.com/ these are the characters (formatting issues are expected):

U+03B8 θ GREEK SMALL LETTER THETA

U+2713 CHECK MARK

U+203E ‾ OVERLINE

U+2191 ↑ UPWARDS ARROW

U+2193 ↓ DOWNWARDS ARROW

U+007E ~ TILDE

U+03B1 α GREEK SMALL LETTER ALPHA

U+03B2 β GREEK SMALL LETTER BETA

U+0394 Δ GREEK CAPITAL LETTER DELTA

U+03D5 ϕ GREEK PHI SYMBOL

Specific characters that aren't showing up:

U+2713 CHECK MARK, U+203E ‾ OVERLINE, U+2191 ↑ UPWARDS ARROW, U+2193 ↓ DOWNWARDS ARROW, U+03D5 ϕ GREEK PHI SYMBOL

Arial contains most of these, Verdana does not. That's the limit of my knowledge of webified things.

kens15• 2 years ago

I think you posted this just as I was removing the <code> tag, so you're looking at the (hopefully) working fonts, not the bad fixed-width font.

2Gkashmiri• 2 years ago

Is it not possible for a website to tell the os, here use the maximum compatible font you have. Instead of trying to use Google fonts ?

kens15• 2 years ago

It looks like the <code> tag is the problem. I guess the Android Chrome code font doesn't include those special characters. I took out the <code> tag and now the characters render on Android (at least for me).

rkagerer2• 2 years ago

That fixed it, thanks Ken.

Taniwha• 2 years ago

This was a drum printer, the other common fast printed from the same era was a chain printer, essentially the same mechanism with a row of hammers one for each column and a chain of characters that would get whacked on as it sped past the hammers, they were really fast!

The reason why they probably didn't choose a chain printer for the shuttle is likely because they had a failure mode where if the chain broke it would tend to fly out one side of the printer and through the wall - our operators were cautioned not to stand beside it ....

dugmartin• 2 years ago

We had to get printouts from the mainframe for my freshman CS class and they came out of a chain printer. It was in the next room separated by a set of cubbyholes for the printouts and it sounded terrifying when it was printing, even 20 feet away.

michaelcampbell2• 2 years ago

My college had a chain printer in "the computer lab" (obsolete now, but not everyone HAD a home computer then). The "joke" in the lab was to run a or write a program that printed a line of "-" or "=" with no newline so it would overstrike and eventually shear the paper, and often cause the printer to jam.

This was especially 'funny' on Sunday nights, when printed programs for courses were due before class on Mondays.

jonsen• 2 years ago

?: "Since the word processor was irrelevant to the Shuttle, I wonder why this card wasn't removed to reduce weight."

!: "We obtained access to a Shuttle teleprinter (probably a development system that remained on the ground)"

creativenolo• 2 years ago

~£59~

Reverse engineering the 59-pound printer onboard the Space Shuttle

Related Stories