Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

righto.com

Post reply on HN
Print view View original

21–30 of 44 posts

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#21

Post by dfox »

CamperBob2 wrote: post #17
kens wrote: post #7

Earlier quoted context omitted.

The transceiver basically has three parts. There's the logic side connected to the Alto, which XORs the sent data and received data to see if there's a collision. The cable-side circuitry has some analog stuff to put signals in and out of the coax. These two sides are connected via transformers for isolation. Finally there's a simple switching power supply to provide isolated 5.7V to the cable-side circuitry. (You wa…

Interesting, I wonder what the funky magnetically-coupled circuit with Q1 and Q2 is for.

That is the aforementioned switching power supply.

On the other hand what I see as more funky is IC1 and IC02, what exactly is that? Some kind of gated comparator? (Also of note is how IC3A and B has different, but equivalent, logic symbol to IC3C and D)

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#22

Post by jacquesm »

jonsen wrote: post #18
jacquesm wrote: post #15

Earlier quoted context omitted.

And because at that point the shield wires were spaced further apart so as not to short out the connection.

Really? Sounds like a major complication to the cable manufacturing process. Wouldn't all those holes in the shielding hose degrade cable specs? All for not having to drill a hole for the eventual connecting tap.

It is not that big a deal actually, either you drill a hole with a special tool and then align the tap with the center of the hole or, on some cable you align the tap with a special dot mark on the cable where the shield has been moved aside a couple of mm. It is a convenience at the slight cost of some loss. Far faster to connect to than drilling holes in insulation using yet another tool.

It really adds up if you have to make a lot of drops on a cable.

If there is one thing I am happy about then about then it is that we no longer have coax networks. Nightmare stuff. Flaky connections and weird voodoo problems.

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#23

Post by kens »

dfox wrote: post #21
CamperBob2 wrote: post #17

Earlier quoted context omitted.

Interesting, I wonder what the funky magnetically-coupled circuit with Q1 and Q2 is for.

That is the aforementioned switching power supply. On the other hand what I see as more funky is IC1 and IC02, what exactly is that? Some kind of gated comparator? (Also of note is how IC3A and B has different, but equivalent, logic symbol to IC3C and D)

IC1 is a 75140 dual line receiver, where pin 6 is a reference and pin 2 is a strobe, so pretty much what you thought.

Also of interest is IC101, a 75122 triple line receiver, which has a Schmitt trigger, two AND gates and a NOR for each line. Most of the pins are tied to ground (1, 2, 3, 4, 12) so it's used as a buffer, ignoring the rest of the logic.

For IC3A-D, I think what's going on with the symbols is a briefly-popular way of representing positive logic and negative logic by changing the symbols so a wire has bubbles on both ends. In other words, if you have active-high inputs, represent it as a NAND gate, but if you have active low inputs, represent it as an OR gate with inverters on the inputs. It's the same gate, just drawn to indicate if it's performing an AND or OR. Note the way the bubbles move on the inverters too.

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#24

Post by kens »

CamperBob2 wrote: post #20
kens wrote: post #19

Earlier quoted context omitted.

That's the switching power supply for the circuitry on the coax side of isolation. The coils are coupled to E, F, G which feed the 5.7V power supply on the previous page.

That's what I thought at first, but the pin numbers (5, 6, and 7) on the E/F/G windings at right overlap with the pin numbers for the primary, suggesting they're different parts altogether. Pretty confusing...

I think that's all "T3". Looking inside the transceiver, there are two toroids, each with 6 loose wires coming off. Since there are no pins, I think the pin numbers are fictional.

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#25

Post by ghostly_s »

DannyB2 wrote: post #3

BYTE magazine, April 1980, page 115. NEW HIGH-SPEED COMMUNICATIONS BUS: Xerox Corporation recently made a public announcement of a new concept of processor-to-processor communications intended for an office environment. This novel concept is called "Ethernet", and is a result of some of the work being done in their research labs. In this concept, a single coaxial cable is used as a high-speed communications bus betwe…

I've never considered networking as "processor-to-processor communications" before. Kind of an interesting angle.

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#26

Post by sigstoat »

bsder wrote: post #14
jonsen wrote: post #13

How did we make a similar adaptor in the 80'ies with a microprocessor running 2 MHz? Fortunately the PLA was invented. A few bits of memory and a little programmable logic. Enough to synchronize to the, in this case 2.35 MHz, Manchester coded bitstream and assemble the data bits into bytes. The processor was still not fast enough to handle the control bytes of the data stream in real time. It could actually handle th…

While I love this (and I am ECSTATIC about the PRU documentation ... thankyouthankyouthankyouthankyouthankyouthankyouthankyou ... I have been using this right now ), I would expect this would have been loads easier with even something really cheap like a MAX10 FPGA. However, he does mention needing a network stack, so maybe not.

> I would expect this would have been loads easier with even something really cheap like a MAX10 FPGA.

basically nothing is easier with an FPGA. faster, perhaps.

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#27

Post by kevin_thibedeau »

> using a 74AHCT125 level shifter chip

The 74x125 is not a level shifter. It is a tristate buffer that can be exploited for poor-man's level shifting with a suitable configuration of external components. It isn't an ideal part to do this and it would be nice if the purveyors of hacker/maker doodads didn't misrepresent what they're selling.

> Without a resistor, the open-collector signal stayed at ground. Adding the resistors to my circuit fixed that problem.

That never would have been an issue with an actual level shifter.

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#28

Post by thought_alarm »

What makes ethernet "Ethernet"?

There's 10BASE-T Ethernet, which is what we all use and has been around since the late 80s.

There's 10BASE5 and 10BASE2 Ethernet, which uses a shared coax cable between machines and was around from the early-80s to the mid-90s.

And then there's the pre-standard coax Ethernet that the Alto used in the 1970s.

Is there a common technological thread between these three things, other than the name?

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#29

Post by mhalle »

jonsen wrote: post #18
jacquesm wrote: post #15

Earlier quoted context omitted.

And because at that point the shield wires were spaced further apart so as not to short out the connection.

Really? Sounds like a major complication to the cable manufacturing process. Wouldn't all those holes in the shielding hose degrade cable specs? All for not having to drill a hole for the eventual connecting tap.

In general, there weren't "all those holes" because most sites didn't have that many networked computers. This was the pre-PC era. 10BASE5 could only have taps every 2.5 meters.

https://en.wikipedia.org/wiki/10BASE5

Re: Xerox Alto's 3 Mb/s Ethernet: Building a Gateway with a BeagleBone

#30

Post by fernly »

thought_alarm wrote: post #28

What makes ethernet "Ethernet"? There's 10BASE-T Ethernet, which is what we all use and has been around since the late 80s. There's 10BASE5 and 10BASE2 Ethernet, which uses a shared coax cable between machines and was around from the early-80s to the mid-90s. And then there's the pre-standard coax Ethernet that the Alto used in the 1970s. Is there a common technological thread between these three things, other than t…

I think the common thread is multiple sender/receivers on a single shared conductor using csma/cd.

https://en.wikipedia.org/wiki/Carrier-sense_multiple_access_...

Post reply on HN