Ethernet cables do not work, despite continuity.

[Alex_Baker]

Hello people

Not sure if this is the right place for this but here we go. I am working on running a couple of ethernet cables around my apartment, they are no more than 35 feet (~10M) in length. I cut the cables to length and terminated them myself, which I have done a dozen times. BOTH of the cables I made today do not work at all, the connected devices and router act like there is nothing plugged in at all. Running a different cable across the floor for testing shows that there are no problems with the router or device.

I have pretty limited tools, basically just a multimeter, so I shorted the color pairs on one end of the cable and tested resistance. I found that each pair measured about 14 ohms (7 ohms one way). According to my research this is a bit high, the working cable measured about half of that.

So clearly there is a problem with either the cable or the terminals that I bought from amazon. I hope there are some people on the forum that can give their thoughts on this issue, I dont want to buy more cable and connectors without identifying the problem first. The cable, crimp connectors, and crimper are all ones I have not used before, but are all pretty standard as far as I can see.

Thanks!

[J-R]

My first thought would be a pair got flipped, especially if you only tested with the multimeter.  Do you have an ethernet cable tester?  You'll want at least a basic one that tells you if it's wired correctly.  But you could also use your multimeter to eventually figure this out if you can't see the colors clearly, assuming you can get both ends near each other.

Some other issues off the top of my head:
- not inserting the wires fully into the plug
- not using the right cable (or wire gauge) and/or plug
- sometimes one of the plastic dividers on the plug gets bent over preventing contact once inserted into the jack

To clarify on cable, solid core is only for infrastructure wiring where you punch it down to the patch panel and wall jacks, stranded is for patch cords.

For the plug, there are two types and using the wrong one can sometimes lead to issues.  For solid core there are two offset points where it essentially straddles the single copper conductor; for stranded it's typically two points that are in line with each other so they can contact multiple strands near the middle.  Putting a plug on solid core is frowned upon, but of course we do it when we have to.

[Alex_Baker]

I tested the cable by shorting pairs on one end, and using the multimeter on the other to test each pair individually. I double checked that each end was the same, unless I dont understand what you mean by a "flipped pair". I am following the T568B configuration.

[MAS3]

Hi.
There's tons of differences, even though the outer dimensions of the cables and connectors may be the same.
As you said you made cables before, i guess you are using the correct mapping of the pairs to the pins on the connector (?)
It doesn't really matter which pairs go to which "pin pair", as long as you keep the mapping correct.
Mapping means the wires do not go into the connector like 1-2, 3-4, 5-6-7-8 (more like 1-2, 3-6, 5-4, 7-8 if i'm remembering right)
There's cables with stranded or with massive cores, and there's different core sizes.
You need to match the connectors to any of these properties for the most reliable results.
There are also different ways the pins cut into the wire, mostly 2 or 3 "teeth" on it, i'm not sure if there is a better one out of these and which that better one would be.
So first thing i would do is to see whether the core sizes and the (holes in the) plug are corresponding.
If the holes are too small, the cores will not fit, and if they are too large, there will be a bad connection even though your multimeter beeps during its continuity test.
Next would be to see if you have stranded or massive cores, and if the connectors are made for the cores you've got.
I hope this helps.

[Alex_Baker]

Well I found part of my issue, I ran both of the cable runs within inches of a wireless access point 

After moving the cables there was some activity, but very sporadic and unpredictable. I may have a mismatch between cable and terminations. My terminations are designed for stranded, but I think the cable I got is too small, something like 32AWG.

[edpalmer42]

32AWG?  I've never heard of ethernet cable that small.  Is this actual ethernet cable or some other 4 pair cable?  Is it CAT5 or CAT5E?

[Alex_Baker]

Well I found part of my issue, I ran both of the cable runs within inches of a wireless access point 

No, that is not the problem.  Obviously you can run Ethernet cable to within 0 inches of an access point.

You really need to get a cable tester, or just buy a couple of premade patchcords.

Yeah you are right, how else would you hook up to an access point....  Interestingly it seemed to have an effect in my case, probably because the cable was very marginal to begin with. I originally chose to terminate my cables myself because I needed to run it through a hole, but I have since found an alternative route.

32AWG?  I've never heard of ethernet cable that small.  Is this actual ethernet cable or some other 4 pair cable?  Is it CAT5 or CAT5E?

The cable calls itself CAT5E patch cable, maybe I just ended up with some really crappy amazon cable while not paying enough attention.

[BradC]

I think the cable I got is too small, something like 32AWG.

Twaddle. I have 30 and 32awg tinned copper here and it's like hair. There is absolutely no way a crimp terminal could terminate that stuff unless it was something like 30awg x 7 (minimum). Also, close to a WAP is bullcrap. It clearly sounds like you have no idea what you are doing. Know what you don't know. Employ someone who actually knows.

A TDR cable tester is bugger all on the usual crap auction/sales sites. Buy one and get a clue.

[J-R]

Ooof, harsh, Brad!


If you shorted the orange & orange/white at one end and checked the resistance at the other end, you would get a reading even if the orange from one end went to where the orange/white was supposed to be on the other end.  So it's not a valid test.  A valid test would be if you could get both ends next to each other and check the continuity on pin 1 only, etc. 

In many cases you can see through the RJ-45 plug to determine if the colors are all 100% correct, but some times you can't.  So a cheap continuity-based Ethernet cable tester is really a must-have item.


You could also post Amazon links to the products you bought so we can see for ourselves if anything is suspect (supplies and tools).

And photos wouldn't hurt either.

[BradC]

Ooof, harsh, Brad!

Not really. If you don't know what you don't know then you'll never know what you need to learn. If it was in the beginners section, maybe I was a bit, but it isn't.

If you shorted the orange & orange/white at one end and checked the resistance at the other end, you would get a reading even if the orange from one end went to where the orange/white was supposed to be on the other end.  So it's not a valid test.  A valid test would be if you could get both ends next to each other and check the continuity on pin 1 only, etc. 

Or you could use a cheap arse RJ45 tester that hits all pins in a sequence and you watch the results on LED's. Cheap, cheerful and lets you know quickly if you've fucked it up. Even Chinese TDR based units that give you an actual wiring map *and* an estimated cable length are worth maybe 10 pre-made patch cables.

[wraper]

Resistance suggests AWG30 wire, CCA on top of that. Such substandard thin wires won't crimp properly in most of Cat5e connectors and zero chance with Cat6 connectors.

[J-R]

Brad, you didn't get the hint, so I'll be more clear: go find another topic to play around in.  It's obvious you're in a bad mood.

And you've just repeated what I already suggested, which was a tester, but obviously the OP doesn't have one, so an alternate method was presented using what they have on hand.

Resistance suggests AWG30 wire, CCA on top of that. Such substandard thin wires won't crimp properly in most of Cat5e connectors and zero chance with Cat6 connectors.

Definitely plausible and with copper prices it's even more likely the cables are suspect.  But also the OP could have just guessed at the wire gauge.
Also, depending on how the other end was "shorted", this could just be some contact resistances adding up, especially if oxidation is present (at the short and/or at the DMM test leads).  So maybe a red herring.

[Berni]

Yeah Ethernet cable quality matters a lot as soon as you go longer distances.

It is more than just having continuity on the wires. The dimensions of the wire and insulation have to be exactly right to get the correct characteristic impedance of 100 Ohm. The pairs need to be each twisted a different amount to reduce crosstalk problems. The material used for insulation determines how fast the signal gets lost. The good cables have shielding (foil or braid or both) even better cables have shielding on each pair individually.

It is not worth buying the cheapest Ethernet cable they have on offer. For the distances typical in residential settings it doesn't cost that much more to buy the high quality brand name cable and know it will be reliable for decades to come.

Also as mentioned above if you have cable that is too far away from the spec then it might not work well with punch downs and crimp connectors. They are designed for the wire conductors being a specific size, If they are not then it might not make a solid reliable connection, even if it does have contact right now, it might loose it once you bend or tug on the connector.

[forrestc]

Just to help (hopefully) help you know what is important:

1) "normal" Network cable (CAT5 or later) is all somewhere around 24AWG.  Although you can find networking patch cables which are smaller, standard cable usable with standard ends will all be about that size.

2) Matching CAT rating and stranded/solid and shielded/unshielded are all important when selecting ends for a cable.   Using stranded cable with crimp-on-ends that are only for solid doesn't work.  Trying to use CAT6 in a CAT5 end often doesn't work due to the wire dimensions/sizes being different.  And so on.

3) The ethernet standard assume you have stranded patch cables at both ends and solid in the middle.  Although that's the standard, it's perfectly acceptable to just crimp an end on Solid cable. But it needs to be a connector designed for that.  Not having the stranded patches at each end aren't going to affect performance - using stranded for a long run including the middle might.

4) It is critical that you put each pair of wires in the right spot on the cable.  I'm not talking "same spot" at both ends, but that you have a pair on 1,2.  And another pair on 7,8, and another pair on 4,5 and finally a pair on 3,6 (note these aren't next to each other).   If you instead put a pair on 1,2, another on 3,4, another on 5,6 and the last on 7,8, the cable won't work correctly.  There are standards for which colors of wires go on each pin, follow them and this will work out correctly.

5) a properly terminated cable is pretty much impervious to any external interference you're likely to encounter in a house - power strip, extension cords, wifi router, etc.

6) Having crimped (or troubleshot someone else's crimping) way too many times in my life - the crimper and ends being of decent quality and compatible with each other (and the wire) is important.  Buying expensive tools and ends don't guarantee this.  I.E. there is very little correlation between acceptablity of a tool and it's price. 

7) I don't know about your budget here - or what you have, or if this is available in your country, but I will make a recommendation of a specific example of a tool kit I like.  Klein makes a VDV226-107 crimper which works really well with klein branded passthrough ends.  In addition, there's a klein VDV526-200 tester which will check for all of the common miswires.  You can often find a kit with the tester, crimper, and ends for around $100 total.   This should be a "lifetime" purchase unless you're using this professionally.   

8 ) Don't feel obligated to buy the specific kit in #7.  It's just one I've finally found that is a good combination of "cheap enough to be affordable" and "expensive enough to work".  It's more an example than a "you must buy this crimper".

9) The passthrough ends in #7 can be controversial as some people have had bad luck with them, but I find that I fight them a lot less than I do a standard end.  With these, when making up the cable you push each wire all the way through the connector and then can tug on them, make sure they're in the right spot, and everything before you crimp it.  The crimp cuts the ends off as part of the procedure.

10) If I was to guess what is wrong in your setup:  Either  you have the wrong cable, or wrong ends for that cable - with a small chance that you have a split pair (meaning pairs aren't in the right spot even though you have end-to-end connectivity).  The reason I say there's a small chance for the split pair is that at the lengths you're working, the pairing being acceptable is less important.  On the other hand, you saying that that the position of the cable matters is a good argument for the split pair issue.

[Jeroen3]

If you follow the rules it's not that hard.

1. No CCA for runs over 10 meters.
2. No plugs on solid core.
3. No stranded wires into LSA punch-down boxes.
4. Follow the same TIA-568 on both ends.
5. Avoid shielded cable unless you really need to and know where to tie the shield to ground.
6. Minimize the number of "turns" your have untwist a pair.
7. No sharp bends.
8. Test before use, it's so easy to accidentally have one of the green and blue mixed up in a plug, then the cable doesn't work

If this all goes okay and the cable still does not work, there is probably a defect in the cable that doesn't show up in DC but prevents the HF to work properly.
For example a tiny in-line copper break due to bending, pulling or standing on it. DC has no issues with it, but the HF will reflect and interfere.
Some switched or NIC have a cable test that can detect the distance of this defect and what pair it's on.

[forrestc]

2. No plugs on solid core.

This really isn't a requirement.

I can guarantee that almost every PoE-powered camera, outdoor radio, commercial wireless access point, and so on is connected to the network using a standard RJ45 crimped onto the end of a solid cable.

Solid is the better, more reliable cable for the signal in an Ethernet system.  But what it isn't good for is flexing and being handled regularly by humans.  As a result, the Ethernet spec allows up to 10m of patch cable total across both ends of an Ethernet link, so you can run solid to a jack on the wall and then use a stranded patch cable from there to whatever device you're using.   And potentially from the switch to the patch panel on the other end.

But.... the permitted use of the stranded cable at both ends is there for convenience and cable robustness not for anything affecting signal integrity (at least until you mangle the connector which is easier on solid).  It also is the correct way to do this in a commercial installation where you're going from a patch panel in a wiring closet to a fixed jack on the wall.

If you're running a cable over about 10 meters, it should be solid.  If you are plugging it into something that won't move or be regularly plugged in and unplugged, crimping an end onto a solid cable is perfectly acceptable.  Which is why it's the way that cameras and the like are done.  The solid cable gets pulled to the location the camera needs to be, an end is crimped on it, and it is plugged directly into the camera.  Once installed, it isn't touched again until the camera is replaced for any reason.

[ftg]

I think the cable I got is too small, something like 32AWG.

...
A TDR cable tester is bugger all on the usual crap auction/sales sites. Buy one and get a clue.

Would you have a suggestion for a cheap ethernet TDR on aliexpress or ebay?
Most of what I see are not in the usual impulse buy pricing one would consider for a single job.

I'll also echo the question about the exact cable and connectors bought on Amazon.
To figure out what it is and if it could be the issue.
I have seen the flat, unshielded cables being sold in 25meter lengths or more with "cat6a" prints on them cause problems with 1GbE.

Do the latest Ethernet standards finally give any definitions or requirements for CCA (Copper Coated Aluminium) and CCS (Copper Coated Steel) "Ethernet" cable?
Or is the standard still copper only?

[amyk]

I have a cheap "Ethernet installation kit" I got off Aliexpress for around $30 (stripper, cutter, crimper, continuity checker, etc.) and a big bag of "crystal heads" for $5, and have terminated around a dozen cables with it that are perfectly usable at 1Gbps, with the longest being around 150'. Cable was cat5e, no real brand, also from the same place.

The OP needs to post some pictures of what he has.

[Haenk]

Don't go cheap on the cable. Buy a known brand, not cheap chinese crap on Amazon. Solid core is best. If in doubt, buy a used cable on ebay and cut to length.
Or go simple and buy a 10m cable.
And you really want a crimping tool, even a cheap one is good enough. You can't do the plugs without a tool, that rarely works at all.

[Kean]

Would you have a suggestion for a cheap ethernet TDR on aliexpress or ebay?
Most of what I see are not in the usual impulse buy pricing one would consider for a single job.

The Uni-T UT685B seems to get good reviews at USD100-150.  AliEx obviously has better deals than Amazon or eBay, but not as convenient.  I have not tried one myself as I have a Fluke CableIQ.

I do have a few of the cheap sequenced continuity testers, with my most recent one supporting both RJ45 and HDMI connections.  So it will cycle through either 9 or 20 conductors (including shield), and I imagine this means it could be easily repurposed for testing your own cable harnesses.  Powered from single AAA or a microUSB supply.  If anyone is interested: https://www.aliexpress.com/item/1005009191148668.html

I agree with others that cable quality is going critical to getting good crimped connections, and for longer runs you will want a higher gauge especially if doing PoE.

[Haenk]

Honestly, I don't think you need a tester. I installed multiple kilometers of networking cables, lots of terminals and crimped quite a few plugs. I think I had a non-working crimp only twice or so. Easily diagnosed without tester.
Just stick to the colors, push them all the way in, crimp, done. For only a couple of meters and low-ish speed, you don't even need to take care of shielding, making this even easier. If the cable does not work, cut off the plug, repeat. Modern switches (newer than, say, 20 years old) even do auto-crossover, so not even required to have different cables made.
And again, for proper installation you need proper cable, and a crimping tool.

[electric_cheeze]

If you follow the rules it's not that hard.

1. No CCA for runs over 10 meters.
2. No plugs on solid core.
3. No stranded wires into LSA punch-down boxes.
4. Follow the same TIA-568 on both ends.
5. Avoid shielded cable unless you really need to and know where to tie the shield to ground.
6. Minimize the number of "turns" your have untwist a pair.
7. No sharp bends.
8. Test before use, it's so easy to accidentally have one of the green and blue mixed up in a plug, then the cable doesn't work

If this all goes okay and the cable still does not work, there is probably a defect in the cable that doesn't show up in DC but prevents the HF to work properly.
For example a tiny in-line copper break due to bending, pulling or standing on it. DC has no issues with it, but the HF will reflect and interfere.
Some switched or NIC have a cable test that can detect the distance of this defect and what pair it's on.

This is great advise
One nit pick, thing I would add is in the US standards twisted pair category cable should be 100% copper, and no smaller than 26 AWG.  Typically you want 24 AWG copper. Any Copper Clad Aluminum (CCA) would be a violation of the NEC (National Electrical Code) section 800.179. Probably other national standards have the save verbiage. EN 50288-1:2003, 4.1. states "conductor shall be annealed solid copper", just FYI there is a lot of CCA cat5/6 sold online and few places where it would be legal to install. 
Cheers

[Alex_Baker]

If you shorted the orange & orange/white at one end and checked the resistance at the other end, you would get a reading even if the orange from one end went to where the orange/white was supposed to be on the other end.  So it's not a valid test.  A valid test would be if you could get both ends next to each other and check the continuity on pin 1 only, etc. 

I should have been more specific describing my testing. To test each pair I took a PCB mount RJ45 socket and shorted pin together, specifically; 1 & 2, 3 & 6, 4 & 5, and 7 & 8. Since I used a PCB socket with the crimped end plugged into it I was sure that the pairs were going to the pins that I expected, and that each pair had continuity all the way through the crimp. I did eventually pull the cable from where it was installed and measure continuity with the ends next to each other, testing pin to pin, I didn't learn anything I didn't already know.

Unfortunately I do actually have a tester, it just happens to be at my parents house currently where it doesn't do me much good here. I rarely crimp ethernet cables so buying a tool kit is not really worth it, the main reason that I do it instead of buying pre-built cables is because I wired all the bedrooms in my parents house with ethernet, so I had all the tools and a large spool of decent cable at their house.

Don't go cheap on the cable. Buy a known brand, not cheap chinese crap on Amazon.

As many others have said I think this is where I failed. This is the cable I bought: https://www.amazon.com/dp/B01G0LUS90?ref=ppx_yo2ov_dt_b_fed_asin_title&th=1
Not sure why I chose this one, I think I specifically wanted round white cable, when most of the ones that popped up were flat. Either way it apparently was "cheap Chinese crap on Amazon", despite pretty good amazon reviews. When you scroll all the way down to the description it does actually say that it is copper clad wire but not what gauge it is, which is seriously somewhere around 30 - 32 awg. It should be noted that I cut the 100ft cable I bought into two 30(ish) foot sections and re-terminated it. This does make me curious to know if the cable worked right out of the box.

Hopefully this will answer some questions, but so far it looks like the cable quality is the most suspect. I really appreciate everybody's input, thanks!

Also I hope Brad's day got better  . I had a brain fart with the access point thing, it threw me off that a cable I was messing with started to work when I moved it, however in hindsight  I think the cable actually was not working, but the indicator lights on the computer came on ,which they never did before moving it.

TLDR: crappy cable, poor judgement on my part. 

[aeg]

Crimping your own Ethernet cables is a false economy. If you buy preterminated patch cords from Cable Leader or Bestlink Netware, you won't have problems.

[J-R]

There are cases where you have put a new end on, such as if the hole isn't large enough for the plug to fit through.  Also, if you have a very nice, expensive patch cord but break the retention clip off, why throw away the cable when you can just put a new plug on it?

With regard to shorting the pins as a test, my point is that if you shorted pins 1 and 2 at one end, but at the other end they were inadvertently reversed (2 and 1 instead), then you'd have continuity, but the cable wouldn't work because it's wired incorrectly.  So it's not a valid test.  Testing each pin individually will be accurate.

The smaller gauge wire and the fact it is copper-clad are problems, but as with most of those cheap items it will work for a lot of people.  So that probably explains the good reviews (along with probably a lot of fakes).  The smaller wire might be an issue with higher power PoE devices, but I don't think it would prevent link.  Really just a thought experiment at this point, though.