I'm not sure if I understood that correctly. It sounds flat out wrong, and it presumably would not be general practice because an Ethernet segment with a missing terminator either doesn't work at all, or works very poorly. An Ethernet segment must have a 50 ohm terminator at each end (and nowhere else). Always. The spec is crystal clear about this, and so is elementary transmission line physics. If you have an intrusive (connectorized) tap, installed at the very end of the cable, the other side of the tap must have a terminator since that's the end of the segment. I once saw an article in a magazine ("DEC Professional" or "RSTS Professional", a US based rag for PDP11 users) that talked about thick and thin Ethernet. It mentioned that you can connect the two -- which is correct. But it showed the wrong way -- with a T connector and a terminator attached to one of the legs of the T. In other words, a "terminator" in the MIDDLE of the segment. The correct way, of course, is a BNC to N adapter, and you have to use the thin Ethernet length and station limits. So it sounds like your instructor was just as ignorant as the writers of that magazine. paul

On Tue, Jun 26, 2018 at 3:07 PM, systems_glitch via cctalk <cctalk at classiccmp.org> wrote: Handy. I'll have to get some pictures this weekend of the stuff I've collected so far. I have 1-2 of the ST-500-01 Transceivers with the non-intrusive taps, and 4-5 of what I am pretty sure now are called ST-500-02 Transceivers, with a single male BNC sticking up where the ST-500-01 has the tap "fang". I'm now thinking for the cost of a few BNC-to-N couplers (couple of bucks from banggood.com) and a one BNC T each (of which I have an abundance), it might be cheapest to get the ST-500-02 acting like an intrusive tap. It might not meet the spec for a full-sized network, but for a 15-20m segment, I can't imagine it's going to break things. It's probably not a terrible thing to get some of the cable, cut it into some 5m segments and terminate the ends with N connectors. Yes, there will be impedance bumps at every joint, but I'm not trying to run 100m, just a segment I can set up somewhere to show how it was once done. Fortunately, I also have an abundance of DEC and desktop hosts with 15-pin AUI as an option (as well as a 10BaseT hub with an AUI, that's at the center of my low-speed network at home). I think the reason is you won't have enough distance between the tap and the endpoint of the transmission line. -ethan

On 06/26/2018 01:07 PM, systems_glitch via cctalk wrote: I'm taking that to mean that the terminator shouldn't be placed immediately adjacent to an intrusive tap. Or rather that you should have some length of cable between the intrusive tap and the terminator. I can only guess that having a terminator too close interferes with or weakens the signal too much in some way. -- Grant. . . . unix || die

On Tue, Jun 26, 2018 at 4:37 PM, Chuck Guzis via cctalk < cctalk at classiccmp.org> wrote: If the termination resistance matches the characteristic impedance of the cable, there should be no difference. When terminated properly, there is no reflection from the terminator, so it looks equivalent to an infinitely long cable, though in practical terms with less leakage than an "actual" infinitely long cable would have. Of course, in reality it will never be terminated perfectly, so there will always be a small reflection, which can be seen with a TDR. If the termination resistance is pretty close, the reflection will be small enough not to matter at all for Ethernet.

So then we're in agreement that screwing the N terminator directly to an intrusive tap shouldn't make a difference? No need for a jumper off the end of the tap for the terminator to live on? As a high school CCNA hopeful, I accepted this as, "it's what you do," and I hadn't really given it any thought since then, as I hadn't had to mess with thicknet since then. I agree that it doesn't really make sense when you actually *think* about it, and, like I said to start with, it's what I recall being told back then anyway -- I could be remembering wrong to start with :) Thanks, Jonathan On Tue, Jun 26, 2018 at 8:02 PM, Paul Koning via cctalk < cctalk at classiccmp.org> wrote:

On 27/06/18 02:56, Chuck Guzis via cctalk wrote: No idea. But on thickwire the taps were all supposed to be made at specifically marked positions (for the reason given earlier). Perhaps someone (incorrectly) thought that the terminator should also be at such a position and so a terminator could not be located at a tap? Antonio -- Antonio Carlini arcarlini at iee.org

On 2018-06-27 19:34:38 -07:00, Chuck Guzis via cctalk wrote: I guess one of the keys to a successful networking technology is that it should be possible to specify how to install it in a way that people not familiar with the inner workings of the technology can readily deal with. There are advantages to keeping the instructions as simple, short and easy to follow as possible with a minimum of exceptions and special cases. It might make sense to state that everything should be spaced 2.5m apart even when there is no advantage to this in the special case of terminators. The advantage is in reducing the complexity of the instructions. The disadvantage is it might lead to difficult cases like this one. On a slightly different point, didn't the thickwire spec call for the outer conductor of the cable to be earthed at exactly one point, presumably for safety reasons in case the cable contacted something at high voltage? This requirement was somehow not carried forward into thinwire, perhaps because the entirity of a thinwire network, including the connectors was supposed to be insulated and therefore not a danger to anyone? DEC produced insulated thinwire connectors and terminators but other than that I think this requirement was honoured more in the breach. Regards, Peter Coghlan.

Yes, and indeed the Ethernet spec does that. The spec is fine. What seems to happen is that people who don't understand EE made up their own additional rules for no good reason. Section 7.6.1 talks about cable lengths; 7.6.2 describes transceiver placement. Those rules are clear and sufficient, but neither says anything about terminator placement. Yes, Ethernet spec section 7.6.3. Also for static discharge, though it doesn't say that explicitly. Looking at 802.3, it says that a Thinwire segment MAY be grounded at one point, but not at multiple points. It also requires a static discharge path at each transceiver, 1 Mohm to ground. So you don't necessarily have a hard ground for the case where the cable is shorted to an AC power line -- I assume the reasoning was that this is unlikely enough it doesn't need to be considered. The Ethernet spec doesn't have anything analogous for 10Base-5 transceivers, so there the hard ground is necessary for a static discharge path. Both coax types, of course, require termination at each end. And both have a stated requirement for all connectors to be insulated. In practice, you can be a little loose with that if you place things so they stay away from other metal objects. You may not see the terminator at both ends on thinwire, if you're dealing with repeaters that are designed to sit at the segment end. Those have the terminator for that end inside the box. paul

On Thu, Jun 28, 2018 at 7:18 AM, Paul Koning via cctalk < cctalk at classiccmp.org> wrote: In case it may not be obvious to some readers, the reason you should NEVER ground an Ethernet cable (of any kind) at two points is that the ground potential at two different points is unlikely to be the same, so that will cause a DC current flow through the cable. If there was a DC flow of a only a few mA, it might not be a big problem, but in practice the resulting DC flow could be _many_ amps, and could both damage equipment and be a safety hazard for personnel. Of course, this isn't specific to Ethernet. It can happen with e.g. TIA/EIA-232 (formerly RS-232) as well, and in fact that is even more common in practice, because Ethernet is transformer-isolated at each station, but TIA/EIA-232 is usually not. It's perfectly fine to have the 1 Mohm static discharge path at multiple points, because the 1 Mohm resistance prevents there from being any significant current. If the cable is not grounded at all, having at least one such static discharge path is important to ensure that no static charge builds up on the cable.

On 06/28/2018 01:22 PM, Liam Proven via cctalk wrote: The US Electrical Code has not allowed any kind of signal wire in the same conduit with any kind of power wiring for as far back as I can remember. bill

On 06/28/2018 02:37 PM, Liam Proven wrote: I realize that, just pointing out that the threat was already well known by the 1980's. bill

tor 2018-06-28 klockan 20:37 +0200 skrev Liam Proven via cctalk: The telephone companies knows about it yes ... Two things: - induced voltages in telephone lines due to solar activity (only at northern and southern latitudes like in northern sweden and norway or Canada/Alaska.) - induced voltages (around 300 V) in telephone lines due to the new railway electrification. Forced Televerket to move their telephone lines away from Malmbanan in 1915 and SJ to add current transformers along the line.

On Fri, 29 Jun 2018 at 04:39, Jon Elson <elson at pico-systems.com> wrote: Fascinating. I did not know that. Thanks for the info! -- Liam Proven - Profile: https://about.me/liamproven Email: lproven at cix.co.uk - Google Mail/Hangouts/Plus: lproven at gmail.com Twitter/Facebook/Flickr: lproven - Skype/LinkedIn: liamproven UK: +44 7939-087884 - ?R (+ WhatsApp/Telegram/Signal): +420 702 829 053

On 06/28/2018 09:45 AM, Eric Smith via cctalk wrote: So, on a TIA/EIA/RS-232C DB-25 connector, what's the official position on pin 1? The standards calls it PGND = Protective ground and most reference seem to indicate that this is chassis/earth ground at both ends of a cable. --Chuck

On 06/28/2018 01:42 PM, Eric Smith wrote: The original standard is very old--it dates form 1960--a very different time; DCE was strictly under the control of the telcos, and I suspect that connection to DTE had to be approved by them. In that sense, pin 1 serving as a "protective ground" might have made some sense--and has been grandfathered in. Note that the DE-9 connector version doesn't include this signal. As a matter of fact, I'm not even sure that NEMA 5-15R receptacles existed in 1960, so grounding may have been a moot issue. Mostly, I ignore pin 1 and will occasionally make use if it if I'm using shielded cable--but only on one end--usually the DCE side. But perhaps someone with a copy of the current TIA-RS232C document might see if the wording has changed. --Chuck --Chuck

On Fri, Jun 29, 2018 at 9:27 AM, Peter Coghlan via cctalk <cctalk at classiccmp.org> wrote: I thought the original use of the RS232 interface was between a terminal and a modem sitting right next to it. In which case there is not likely to be much voltage between the protective grounds of the 2 devices (well, unless there are serious problems with your mains wiring!). Longer distance serial interfaces were originally current loop which were electrically isolated at one end, originally using high-speed relays and later opto-isolators. The one I've come across is called a 'Protector, Heat Coil & Fuse'. It has a cartrdge fuse and a heat coil (basically a resisance coil with a spring-loaded soldered joint that is designed to open if the thing gets hot enough to melt the solder) in sereis with each line wire and a spark gap between each wire and ground. I think they were more trouble than they were worth, the one I have was officially modified by removing the spark gaps and replacing the (cartridge) fuses with solid metal bars, leaving only the heat coils in-circuit. There was a thing called a 'barrier box'. At one time in England you had to rent the modem from the Post Office (who had a monopoly on telephone systems [1]). You also had to have a 'barrier box' between your RS232 device and the modem. This consisted of a low-rating fuse (50mA?) in series with each RS232 wire (apart from pin 1, Protective Ground) and a pair of zener diodes (25V?) in inverse-series between each RS232 wire on the modem side of the fuse and Protective Ground. So 24 fuses and 48 zeners in all. Protective Ground was passed through it AFAIK. [1] I seem to remember that Hull had its own telephone company, but I suspct the same rules applied. Some early serial interfaces had internal fuses/zeners. The HP11206 (modem interface for the HP9830) does, although the HP11205 (RS232 interface for serial printers, etc) and HP11284 (async/sync serial interface) do not. In the case of the HP11205, I suspect it is because it was never intended to be used with a modem. in the case of the HP11284, I suspect the requirements had eased by the time it was introduced. It appears the British Post Office were very worried about excessive voltages on their lines. I have an HP telephone line analyser and the UK version (which is the one I have) has an extra PCB of zeners and fuses inside (it's desribed in the service manual as the 'A44 Protection Board Assembly' and it states that the 'A44 Assembly is only fitted to instruments intended for use in the UK' For some unknown reason, you were supposed to use a barrier box on the GPIB interface of the PERQ if you linked certain printers to it (like the Versatec electrostatic thing). The unit was just an RS232 breakoutbox with a couple of specially-wired cables to get the signal lines on protected pins and the protective ground line on the GPIB sheld pin. -tony

If anyone's wondering, here's a 1975 Bell System treatise on grounding and protection. Note that carbon-block protection was standard practice: http://bellsystempractices.org/500-/518-/518-010-105-i02_1975-09-01.pdf Note also, that the standard practice dictated that grounding at the dmarc should be tied to the power system ground, but that the grounding not use the power system ground (i.e. independent connection). Note also, if equipment frame/chassis grounds varied, bonding to a common ground e was recommended. This has loosened up quite a bit since the old days. On my home, built in 1980, the network interface box is tied to the old demarc box ground, which ran out to the pad-mounted transformer in the side yard. When the local utility moved the transformer pad to the front yard, they simply disconnected the telco ground and didn't tell anyone. When a tree in the side yard became uprooted during an ice storm a couple of years ago, it pulled up the old telco ground cable. I took out a service ticket for the incident and was told not to bother. When I asked why, I was told that buried subscriber lines are no longer required to be grounded. They didn't seem to be too perturbed about it, so I just coiled the cable up and buried it when I filled in the hole left by the tree. But yes, I remember the old Bell System 232 hookups--DB-25 connectors with phenolic backshells. I think I've still got a couple of those in my junkbox. --Chuck

On 29/06/2018 14:31, Paul Koning via cctalk wrote: The RS-232-C standard doesn't actually specify a maximum distance. Section 3.1 *recommends* short cables up to 50 feet or 15 meters (with speeds up to 20kbps) for RS-232 (and the later-dated Foreword suggests RS-449 beyond that) but says that longer is fine so long as the load capacitance excluding source and cable is less than 2500pF. In practice, much depends on the baud rate and cable quality used. -- Pete Pete Turnbull

On 06/29/2018 02:30 PM, Ethan Dicks via cctalk wrote: That was a very common setup in the 80s. 25 or 50 pair phone cable running through the usual wiring closets and punch-down blocks. Simple, and something that installers could understand. --Chuck

From: Chuck Guzis Sent: Friday, June 29, 2018 3:43 PM In fact, both of the LOTS sites (the original, in the CERAS building, and LOTS Two in the old bowling alley at Tresidder Union) were wired this way. I learned to use a punchdown tool my first day on the job. Rich Rich Alderson Vintage Computing Sr. Systems Engineer Living Computers: Museum + Labs 2245 1st Avenue S Seattle, WA 98134 mailto:RichA at LivingComputers.org http://www.LivingComputers.org/

On 2018-Jun-29, at 1:27 AM, Peter Coghlan via cctalk wrote: Over here at least, they are still a standard part of any telephone landline. Many decades or a century ago they were a big white 4*5 inch ceramic block mounted in the basement, with two long red (fusible?) resistors in series with each pair-wire and a double carbon-block spark gap to ground underneath a black bakelite knob. Now they're built into the house-side connection box and you might not even know there is something besides the connection terminals there. How often they actually come into active play I have no idea, but they still strike me as a good idea when you have miles of overhead line back to the CO. Surely you'll often get away with such, but "surely it should be connected to pin 7 . ."? No. I don't know whether the -232 spec actually refers to pin 1 as "protective" ground, but regardless, "protective" here doesn't have to mean the same thing as "protective" in the mains wiring arena (or provide the identical functions). Protective ground in the mains arena serves both to flatten voltages and bleed off slight currents from leakage and L/C coupling, and to carry enough current to blow the mains breaker in the event of shorts to chassis. The "protective" ground on a -232 connection could be performing at least two functions: the afore-mentioned neutralisation of leakage, and noise suppression. It is not at all insensible to separate those functions into a separate wire from the wire providing signal-circuit continuity (signal common / signal ground). You don't want currents or voltages from leakage or noise upsetting the signal-circuit common level. I don't think anyone was intending the -232 protective ground to provide the 'blow the mains breaker' function. Yes, sometimes the noise/leakage/signal-common functions are all combined into one wire, or stated alternatively, sometimes the environment and needs allow one to combine them into one wire. Consumer audio for instance - where the shield is commonly both the signal common and noise/hum shield. But even in that low-requirements arena you might remember turntables - where there were two channel-cable common/ground shields but there was also a separate chassis ground wire. Equipment, esp. back in the 60's, might not have a mains ground. Even in equipment that does, signal common/ground and chassis ground may be separated. Relying on the signal ground wire to provide all these functions when connecting equipment with varying grounding policies is asking for problems. - Speaking of abusing RS-232, in our CS dept (UBC) in the 80s when we moved our project offices to another room some ways away from the machine room we were faced with getting terminal cabling between them. 'Properly' this would involve calling in the physical plant dept or the telco to run new wire or make twisted-pair connections on the telephone wiring. However, a little OCD observation of wall panels over my time there, some speculation, and some continuity testing allowed me to figure out there were some little-to-unused 100-pair cables terminating in punch-blocks in panels in various rooms, so with a little punching between blocks within a panel I was able to get continuity between the machine room and our new room. No new wire-runs and no bureaucracy involved. We used -232 over those connections. Something makes me think I reserved 3 pair per line but used 2 (4 wires): XMT, RCV, DTR to let the terminal switch know a terminal was present, and signal-GND. The university timesharing Computing Center, to provide terminals across the large campus, used 422-style balanced-line signaling. Every CC-connected terminal around campus had a little 3*4*6 box sitting with it, with 3 LEDS (power, XMT, RCV), and containing dual power supplies (232 terminal side & 422 'campus' side), two opto-isolators (XMT & RCV) and assorted drivers. Each connection was 2-pair (XMT,RCV) over the telephone wiring physical plant. There were various manufacturers (Gandalf and Develcon come to mind) that made such devices, but in UBC's case the computing centre made their own. I saved what is probably the last existent one when I ran across it in a radio-musuem donation pile years later but years ago now.

On 06/26/2018 06:02 PM, Paul Koning via cctalk wrote: I wouldn't call it "confusion" on my part. I'd choose "ignorance" and I believe it's more accurate, at least for me. Coming from nearly 20 years of networking and many colleagues that did things because they were told without asking, these types of minutia, yet important, details get lost. I've got to say that I think a hybrid of computers and amateur radio is a wonderful thing. Especially when training for one crosses over and supports the other. Throw in a healthy dose of DC & AC (low Hz RF) and I think you've got a winning combination. Certainly more than the sum of the parts. -- Grant. . . . unix || die

On 2018-06-26 08:02 PM, Paul Koning via cctalk wrote: And if you want to see the reflection on a scope or so you can look at https://www.youtube.com/watch?v=y8GMH7vMAsQ . That is a follow up to the theoretical explanation given at https://www.youtube.com/watch?v=rN9zjVK3Nrk ------------------------------------------------------------------- Techwiz, Peter Sjoberg GPG key (42DDDDDD) on keyserver & homepage Key fingerprint =? EB81 3135 1636 576A DA83? 826B 2455 0E88 42DD DDDD Homepage: http://www.techwiz.ca/~peters Pictures: http://www.flickr.com/photos/henahadu/ Enigma: http://meinEnigma.com/

It's clear to me: the person making the claim to Jonathan simply didn't know enough EE to understand that his statement was invalid. No, the terminator placement has nothing to do with it. Collision detect is a very simple process: it involves measuring the signal on the wire and comparing that with what the transceiver is currently sending. If the two differ by enough, it means another sender is active at the same time. There are no special cases in this. Proper collision detection depends on the segment length rule and the minimum packet length. If you run the numbers, you will see that the minimum packet length and max segment length were chosen to ensure that a collision between two stations, at opposite end of a max length cable, will be detected by both senders. If you violate either limit by a significant amount this will no longer be true, and you've reduced the system to basically Aloha, which works fine at low load but maxes out at 16% of wire rate, or thereabouts. paul

Yes, to avoid false alarms. The purpose of the spacing rule is to ensure that there is enough signal integrity that you do not get spurious collision indications due to reflections off the impedance variations along the cable. On a segment with few transceivers, there is enough margin that the rule doesn't matter. This is why 10Base-2 doesn't have that rule: the station count limit is low enough that it isn't needed. A tap is a stub, and on a transmission line a stub is an impedance change -- a small one if the stub is short in proportion to the wavelength. Similarly, connectors will show up as small bumps because the devices aren't ideal. If these bumps are spaced at multiples of the wavelength, the reflections that occur at any impedance variation will combine to form larger reflections. If enough of these add up, the reflected signal can appear like another transmitter to the collision sense circuitry. So the spacing rule (for taps) and the cable section length recommendations (distance between connectors) are both set to place these perturbations at points that are NOT multiples of the wavelength or small fractions thereof. (Sometimes people say that the spacing rule is there to place things at multiples of the wavelength; that is exactly backwards.) So the spacing rule doesn't have anything to do with detecting real collisions, but it is necessary to ensure no false collosions. paul

I can believe that. But the limit for 10Base-5 is 500 meter segment length, 100 transceivers. For 10Base-2, the limits are 185 meters and 30 stations. Note that's per segment; the whole network can be larger when repeaters are used. But the segment limits are the ones that relate to the electrical and signal integrity aspects of collision detect. I hope your 100-node thinwire network wasn't a single segment; such a config would run only if you're very lucky. The first Ethernets I saw were at DEC, in engineering buildings. They had long runs of 10Base-5 coax going all over the building, with transceivers for every node and a whole lot of nodes total. Those were probably pushing the limits. This reminds me of a research project done at DEC that at one point was discussed as a possible product but didn't happen: an Ethernet segment mapping device. It was called "packet voltmeter". The idea was that you'd have one at each end of the cable (it replaced the terminator). It would build a table of source addresses and packet signal amplitude. You could then combine the measurements at the two endpoints, plus the known cable attenuation, to make a physical map (with tap placement) of each Ethernet node. The arrival of thinwire with its short segments and star wiring, and especially twisted pair with point to point wiring, made this idea not suitable for a product. But the concept was pretty amazing, and it did work. paul