10

In the early 1980's, plugs with low pin density like DB-25 were common on home computers. The industry was still relatively small, not everything in the manufacturing chain was automated. I assume a lot of those early cables were soldered by hand. I'm sure higher densities were possible, but if a cable cost $10,000 or something it wouldn't have been practical in a consumer product. So, when did the physical manufacturing technology in the consumer computer market reach a point where you could travel back in time with a USB-C diagram and use it in a product?

To scope this, I am not asking about performance or protocols. A full modern USB software stack takes more memory than existed on a home computer in the early 1980's. That's fine -- I am just asking about the manufacturing of the physical cables, plugs, and sockets. Even if a 1980's USB-C would only have run at a few KBPS, and only worked as a simple serial connection.

Edit in response to a question to add : The ability to flip USB-C is interesting, but not directly relevant to the question. I am assuming a microcontroller or other circuit could handle the flipping quite early, at the cost of expense and possibly speed. But the question is specifically about just the physical manufacturing, and how it evolved over time to make it practical. A USB-C connecter has small springs within the plug, extremely fine tolerances in the hundredths of a millimeter. Was it even practical for a cable manufacturer to source exactly the right steel for the internal springs in the connector before widespread Internet connectivity made it easier to find suppliers?

7
  • Are you specifically interested in the magic of being able to insert the cable either side up? (Or to ask differently, does en.wikipedia.org/wiki/Apple_Desktop_Bus qualify, and if not, what doesn't it do that USB-C does that you want?)
    – Foon
    Apr 27 at 19:01
  • 2
    In general, physically larger connectors are more convenient to plug and unplug than smaller ones, and are also apt to be more mechanically robust. I don't think it would have been particularly expensive to shrink connectors during the 1980s, but for most tasks many of today's tiny connectors would have been regarded as impractical even if they were no more expensive than larger ones.
    – supercat
    Apr 27 at 19:29
  • 2
    This document suddendocs.samtec.com/literature/samtec_story_original.pdf from Samtec suggests that even in the early 2000s high-density connectors didn’t go much below 1.27mm pitch. Whether they could have done of the demand was there is open to debate.
    – Frog
    Apr 27 at 20:08
  • 1
    HP-IL is another example on 2 pins per connector but is was a ring so a device had in and out for a total of 4 pins across the two connectors. From the early 80's.
    – Brian
    Apr 27 at 20:18
  • 2
    I'd rather ask "would the customers buying expensive computer equipment in the early 80ies have accepted such crappy connectors we have today" - And I think the answer would be "no.". Stuff back then had to last much longer than today.
    – tofro
    Apr 28 at 6:09
12

TL;DR: Yes, But...

Yes, it would have been possible, but what would have been the use case? Inventions without a use case will not be picked up and thus not produced.


The Details

In the early 1980s, plugs with low pin density like DB-25 were common on home computers.

That depends a lot on purpose.

  • DB25 it was for high pincount, like serial or parallel
  • DE9 for low pincount like joysticks
  • Or Atari's Sub-D-like 13 pin SIO connector somewhere in between

The industry was still relatively small, not everything in the manufacturing chain was automated. I assume a lot of those early cables were soldered by hand.

These parts were standard ones, used in millions - the PC industry was just a very small part thereof. And no-one, apart from hobbyists, was hand soldering them. At least not for straight cables and products made in reasonable quantities. That would have been ridiculously expensive.

I'm sure higher densities were possible,

Usage is not about being possible,but being useful. With computers and everything connected to them being so big at the time, why would a manufacturer choose a smaller connector?

but if a cable cost $10,000 or something it wouldn't have been practical in a consumer product.

Any reasoning for that number?

So, when did the physical manufacturing technology in the consumer computer market reach a point where you could travel back in time with a USB-C diagram and use it in a product?

There is nothing in USB-C that couldn't have been manufactured back then. But what would it be good for? There was no device so small in size that any bigger connector wouldn't fit. Think about, even a very slim notebook, like a Kyotronic 85, still measured 50 millimetres in height. That's more then enough for the 10mm a DB25 needs.

I am just asking about the manufacturing of the physical cables, plugs, and sockets.

No real issue to manufacture this in 1980 - but also no use case and thus no sales for your 'invention'. There was no 5 mm thick phone that needed a connector.

Even if a 1980's USB-C would only have run at a few KBPS, and only worked as a simple serial connection.

  • For one, USB is a simple serial connection.
  • Next, serial connection have used smaller plugs, like
    • Commodore using 13 mm DIN 45322 plugs in 1980 for their serial IEC or
    • Apple using 7 mm Mini-DIN for their 1986 ADB
  • Last but not least, several serial bus systems were used from the late 1970s all the way into the 1990s,for example
    • HP's HP-IL (which did use a connector similar in size to USB-C) (Thanks Brian)
    • The mentioned Commodore serial IEC
    • Atari's SIO
    • And Apple's ADB

The last two of those can be seen as direct predecessors to USB.

Long Story Short:

USB-like buses were already available, including quite small connectors.

14
  • 4
    What about the tolerances? The tolerance on a USB-C plug width is specified at +/- 0.03mm. That could almost certainly have been achieved with bespoke manufacturing technology, but what about mass manufacturing? Apr 27 at 23:22
  • 4
    @user1937198 Not sure how rough you assume a technology to be, that at that time already made steel tubes fly faster than sound without killing its passenger? Reaching a certain tolerance isn't a value in itself, but as well only related to what is to be achieved. But to give some random , the pins of a Cannon sub-D plugs for example were specified at +/- 0.05mm . And that has been 1952.
    – Raffzahn
    Apr 27 at 23:58
  • 3
    @user1937198 Possibly there could have been issues manufacturing plastic mouldings to that accuracy (because of there was no prior demand to develop the technology or the materials) but manufacturing metal components to 0.03mm tolerance would have been no big deal. For mass production, note that automatically controlled machine tools for metalworking were in use going back to WWI (long before computer numerical control existed!) and automated woodworking tools were used for mass production in the 19th century. A trained machinist was expected to work to 0.003mm when required, not 0.03mm.
    – alephzero
    Apr 28 at 2:34
  • 3
    "That would have been ridiculously expensive." That's a hard historical nope. Even today, my company has an outsourcer we use for cable assemblies. That's all they do, and they do it cheaply and well. IDC cables were an option to reduce assembly cost, but the connectors and cables are more expensive. You really wouldn't believe how fast a good assembly technician can work; and if you get them made in Far East (as much of these "commodity" items were even then), it becomes even cheaper.
    – Graham
    Apr 28 at 10:08
  • 3
    You also have the issue of backwards compatibility. Think about this: There's one piece of hardware on modern desktop computers that has remained basically unchanged since the 1970s: The power cable. One end of it is that rectangle with 2 corners cut off and 3 vertical contacts, and the other varies by region but plugs into whatever your standard power outlets are. If you lose yours, any cable made in the last 50 years will serve equally well because there's been just no reason to upgrade them. Same goes for signal cables - there's no need to change them until new technology forces it. Apr 28 at 13:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.