| General > General Technical Chat |
| Big Test MFG says our Ethernet Chip violates the IEEE standard - are they right? |
| << < (3/8) > >> |
| ve7xen:
I think people (myself included!) have misinterpreted your post to mean that a test house failed you on compliance and provided a report saying so, not that they simply had trouble using your product with a Cat3 cable. I don't think it's really reasonable to expect a Cat3 cable to 'just work' with modern equipment, it won't meet spec for 100base-TX or better, and almost all Ethernet interfaces are 'or better' these days. So outside of curiosity, this seems like a don't care to me. Does your product support 1000base-T (Gigabit Ethernet)? Pretty typical for a Cat3-era cable to only be two-pair, and if they were connecting it to 1000base-T equipment on both sides, a 1000mbps link will be negotiated, but fail to function at all with the missing pairs. It wouldn't meet the integrity requirements for > 10base-T anyway, even if it did link, so there would be no expectation of it working properly (though it likely would, at least for a short run), and most modern equipment is going to want to link up at 1000base-T. If it's 100baseTX or 10baseT, it could be a crossover issue, depending on whether your/their side does AutoMDIX or not. Almost any PC these days will do it, but an old switch or whatnot that's been sitting in their equipment rack for 20 years they might be plugging it into may not. You can mess up the PHY implementation, of course, but not really in a way where an improved cable spec would help (it mostly affects crosstalk and the like), at least that I can think of. This is typically all handled in the COTS PHY and you simply wire it to magnetics and the jack that are also purpose designed. I'd guess it's one of the above things (missing pairs, crossover, out-of-spec) or that the cable is simply bad. |
| metrologist:
Thanks - yes - sorry - my OP was poorly constructed. Trying to squeeze something in between doing other things. I did mean to state in the title, Big Test Equipment MFG... I don't have access to the specification but it seems reasonable that it would state backward compatibility, so I'd expect the system to fall back to 10Base-T, and if our implementation does not do that then I'd be inclined to agree with the customer. The interface supports gigabit, but would only be used for sending settings, like for a function generator. We even supply PC software with a simple interface for remote operation - as simple as entering the IP address. |
| wraper:
--- Quote from: metrologist on October 20, 2021, 08:22:20 am ---Thanks - yes - sorry - my OP was poorly constructed. Trying to squeeze something in between doing other things. I did mean to state in the title, Big Test Equipment MFG... I don't have access to the specification but it seems reasonable that it would state backward compatibility, so I'd expect the system to fall back to 10Base-T, and if our implementation does not do that then I'd be inclined to agree with the customer. The interface supports gigabit, but would only be used for sending settings, like for a function generator. We even supply PC software with a simple interface for remote operation - as simple as entering the IP address. --- End quote --- If the device on the other end supports more that 10Base-T, it's not surprising that connection would fail. I guess in such case you would need to manually limit your device to 10Base-T for it to work over crappy cable. Or neither or devices supports Auto MDI-X and require crossover cable to work (and they had a straight-through one). |
| tszaboo:
Maybe they told you that, because your specification for the test wasn't appropriate (technical sense). Should it even support Cat 3 in the first place? Is there a sectional part of the standard that you should've asked for for the certification? Some testing house will know if your product will fail some obscure, outdated, or rarely used part of a standard, and only test that, and charge you full price for the testing. Because they will know, that 95% of the DUTs will fail and it is less effort for more money. |
| tooki:
A few thoughts/summaries in addition to what’s been said already: - I don’t have access to standards to check, but I suspect that backwards-compatibility is customary, but not required, because I think I’ve seen uplink ports on (gigabit or lower) switches that don’t support all lower speeds. - 10Gbps Ethernet ports seem to have dropped support for 10Mbps Ethernet. (I’m not sure about 2.5Gbps ports.) - 10Mbps Ethernet requires Cat 3 cable (using two pairs), but requires the 8P connector. It cannot be used on a 6P4C connector. - 100Mbps Ethernet requires Cat 5 cable, but still only uses two pairs. - Gigabit requires Cat 5 cable, but requires 4 pairs. - Typically if two gigabit devices are connected with a two-pair cable, it’ll drop down to 100Mbps. (There’s no fallback to 500Mbps or something.) - Auto-MDIX is mandatory for gigabit ports, but optional on lower speed ports. I want to say that I’ve seen or heard of ports that will not do auto-MDIX at 10Mbps, even if they support it at higher speeds. I agree with the suspicion that both devices support at least 100Mbps, and thus expected Cat 5 cable even when just two pairs are in use, so they tried running 100Mbps over Cat 3, which is unlikely to succeed, and indeed did not. Given the fact that Cat 5 patch cables were the norm even back in the heyday of 10Mbps Ethernet, Cat 3 cable certainly isn’t a usage scenario I’d worry about. I can’t even remember if I ever saw Cat 3 patch cables in the wild other than as pack-in cables with things like DSL modems. It seems to me that most Cat 3 cable was the stuff installed in the walls. |
| Navigation |
| Message Index |
| Next page |
| Previous page |