IP Camera Long Distance Ethernet Test

By: John Scanlan, Published on Aug 12, 2016

Many often ask about running Ethernet beyond its standard distance of 100 meters for UTP, especially when cameras are 400', 500' or more away.

Will it fail?

How We Tested

We tested lengths of Category 6 and 5e cables starting at 1000' (~300m), down to 300' (~90m). Plugs were directly connected to both ends, with three tests performed:

Camera test: We first tested whether cameras would power up and maintain link with a switch, whether pings were received reliably, and whether we could access the web interface/live video. Cameras tested included: 720p, 1080p, 4MP, fixed, PTZ, integrated IR, 802.3af and 802.3at.
Laptop test: Next, we connected two laptops together via the length of cable to see if they maintained link and check connection speed using iPerf.
PoE test: Finally, we measured voltage and wattage using a PoE tester at the far end of the cable to see if power over Ethernet was being properly delivered over the length of cable.

Cat 6 (yellow, below) and 5e (blue) were both used to check for differences in performance due to category rating or simply larger gauge (23 AWG vs. 24 AWG).

See our results inside.

**** ***** *** ***** running ******** ****** *** ******** distance ** *** ****** for ***, ********** **** cameras *** ***', ***' or **** ****.

**** ** ****?

How ** ******

** ****** ******* ** Category * *** ** cables ******** ** ****' (~300m), **** ** ***' (~90m). ***** **** ******** connected ** **** ****, with ***** ***** *********:

****** ****:** ***** ****** ******* cameras ***** ***** ** and ******** **** **** a ******, ******* ***** were ******** ********, *** whether ** ***** ****** the *** *********/**** *****. Cameras ****** ********: ****, *****, ***, fixed, ***, ********** **, 802.3af *** ***.***.
****** ****:****, ** ********* *** laptops ******** *** *** length ** ***** ** see ** **** ********** link *** ***** ********** speed ***** *****.
*** ****:*******, ** ******** ******* and ******* ***** **** ******** *** *** *** of *** ***** ** see ** ***** **** Ethernet *** ***** ******** delivered **** *** ****** of *****.

*** * (******, *****) and ** (****) **** both **** ** ***** for *********** ** *********** due ** ******** ****** or ****** ****** ***** (23 *** **. ** AWG).

*** *** ******* ******.

[***************]

Key ********

** *** ***** ** cables **** *** ******** 300', ** ***** ******* key ********:

******** ***** ********* *** achieved ** ***' ** multiple ******* ** ******* specifications.
**** ******* *** **** at **** ******* *******, as *** ******** ********** using * ***' ***** length.
********* ******* *** *********** ********* when ***** **** **** cables *** *********-********* *******.
*** *** ******** ********* at *,***' (*** ******* length ******), *** *** signal.
**** *** ** *** 6 ********* ********* ** our *****, **** ** practical *********** ******* **** at *** ********.

What ********* ***

****** *** ** ** no ****** **** **** (~330') ********* ** ******/***-*** ********. **** **** ** specified ** ******* *** horizontal ***** (********* *****, from ***** ***** ** device) ** **** ** up ** *** ** patch ******, ****** ** the **** ** ** cameras ** ****** ******* panels, ***** ****** *** often *** ****, ********* ******** ******** ** the *****. *** **** ******* ** these ******, ******* ****** see ********** **** ********* ***** *** ***** *** ** ***** Surveillance *****.

Note **** ** ******** ********* ********** ****** *** ***/*** ******** *** ******* *** **** ******** **********. This test was performed to see maximum distance possible, but by violating standards you assume risk of potential future performance issues. Distances over 100m are not tested for longevity, interference issues, etc., which have been properly tested in <100m cables.

*** ***** ******* ****** cable **** **** *** 300'/100m *****, *** *** report: **** ** ****** *** Options: *****, *** ********* and *** ********.

PoE ** ****'

** *** *****, *** functioned **** **** * full *,***' *** ** cable. ********** * *** test ***** ** **** with *** ****** ** this ******, ** *** see **** ******* ******** was ****** *** ***** (44-57VDC), *** ******* *** the **** ** ******* drew ** *******, *********-********* distances.

**** **** *** ******* are ******** ** ******** voltage ** ****** ***** runs, **** *** ** 57V **, ** ***** to ********** *** ******* drop *** ** ********. These ******* **** ********** in **** ****, **** voltage ********** **** *** at ***' ** ** at *,***', ***** ** the ***** *****:

******* *** *******/******* ***** properly ********, ** **** unable ** ******** * link ** *,***', ****** due ** *********** ** data *******. ******** ***** were *** ********** ***** 800' *** ******* (*** *****).

Video ******** ** ***'

************, ** **** **** to *** ******** ********* at ***' ** *** camera ******, ************ *** ******(**-*********-*).

***** ******* ** **** length *** *******, *********-********* lengths *** ****** *********, seen ** *** ********** below. ** *** ** frame ****, **********, ** other ****** ** **** length, *** ** *****.

*******, **** **** **** was ******** ** **** one ******, **** ****** simply *** ********** (**** at ***'). ************, ****** resolutions *** ********** *** generally **** ****** ** experience ******, ** ********* limits *** ** ***** long ******* *** *******.

Reliable ** ****** ***** ** ***'

**** *******, ** **** able ** ******* ******* ******* without ***** ** ***' distance. **** ******** ******* options: ****, *****, ***, fixed, ***, ********** **, 802.3af *** ***.***. *** cameras ****** ** **** distance **** ** ******* pings *** ** ****** on ***** *******.

PC ** ** ***** ** ***'

** **** **** ** link *** *** ** 800' ******* ** **** Cat ** *** *, with ** ******* ***** and **** ********* ******* properly ******* *** ***. However, ****** **** **** limited, **** ***** * Mb/s, ****** **** *********** for **** ***** *** in **** *****.

Cat * **. *** **: ** **********

******* *** ****** **** gauge *** ****** *********, we *** ** ********* difference ******* *** * and ** ** *** tests. ** **** *****, Cat * ******* ***** pings **** *** ** at **** ********* (***'+), but ** ** **** did * ****** ** PC ******* ** *** 6 *** *** **.

Comments (38)

Undisclosed #1

08/12/16 12:17pm

Try UNV. they say if you use UNV NVR and Camera they can guarantee up to 250m operation. I am not sure exactly what models support that though

Stephanie Guo

In reply to Undisclosed #1 | 08/13/16 03:05am

All UNV (Uniview) products use special hardware and optimized algorithm to guarantee the performance of 950 feet PoE connection. Hope IPVM can test any UNV (Uniview) models regarding this feature.

Undisclosed Integrator #6

In reply to Stephanie Guo | 08/13/16 02:42pm

Perhaps OT:

Where can one find Uniview i the US or in the LATAM (Latin America) countries? I have read some great things abou them Results seem to be rather superb.

OnVIF compliance or at least drivers for usual/popular VMS is a necessity.

Undisclosed #1

In reply to Undisclosed Integrator #6 | 08/13/16 04:21pm

I posted link to the linkedin profile of the UNV usa country manager on my original post however it seems to have been deleted by a moderator.

@IPVM. Any issues posting his details again? He is my competition so no benefits for me from doing that. Let me know.

John Honovich

IPVM | In reply to Undisclosed #1 | 08/13/16 06:52pm

This is a test of regular Ethernet. I don't want this to turn into a discussion on specific manufacturer proprietary implementations.

There's lot of proprietary equipment that allows for longer Ethernet runs.

6, I will email you offline with specific contact information for Uniview. It's not appropriate for this post.

Undisclosed #1

In reply to John Honovich | 08/14/16 12:11am

I understand. Thanks for the clarification. :)

german bravo

In reply to Undisclosed Integrator #6 | 08/26/16 02:41am

www.artilec.cl in Chile and Peru

Undisclosed Manufacturer #8

In reply to Undisclosed #1 | 02/03/17 08:24am

Every network switch support this 250m connection will be a good news.
Otherwise it's just like analog connection. Completely waste of network cables in site.

Paul Charles

08/12/16 12:19pm

Could you test Cat6a -v- Cat6 please

Undisclosed Integrator #2

08/12/16 01:17pm

Great "what if" stuff. I wish I had more time to do things like this myself. Glad to have IPVM look into these interesting scenarios.

Gert Molkens

IPVMU Certified | 08/12/16 01:57pm

The problem is that these things might verry well work and sometimes they won't. There's no guarantee that it will work if you go behind limits.

That's the purpose of the standards: making sure it works reliably if you stick to them.

I have also seen longer than allowed lengths working just fine. I've also seen a lot of them fail, from time to time or all the time. That's the problem, you never know what it will do now and in the future

Ethan Ace

IPVM | In reply to Gert Molkens | 08/12/16 02:20pm

I've seen things work/not work, as well, and I agree it's incredibly risky to run overlength cables.

This is why we have this note in the report:

Note that we strongly recommend installers follow the TIA/EIA standard and observe the 100m distance limitation. This test was performed to see maximum distance possible, but by violating standards you assume risk of potential future performance issues. Distances over 100m are not tested for longevity, interference issues, etc., which have been properly tested in <100m cables.

Undisclosed Integrator #3

08/12/16 02:30pm

What Brand of Cable were you using for the tests?

John Scanlan

IPVM | IPVMU Certified | In reply to Undisclosed Integrator #3 | 08/12/16 02:50pm

They are both cable matters, both are solid conductors (not stranded), the Cat6 is 23 AWG / 550MHz and the Cat5E is 24AWG / 350 MHz.

Ethan Ace

IPVM | In reply to John Scanlan | 08/12/16 03:53pm

Note that we intentionally did not buy any of the bigger brands of cables (see Poll: Do You Buy Brand Name Cable?) as we wanted to make it a test of average or below average real world cable quality.

Undisclosed Integrator #4

08/12/16 03:07pm

Great article! While we've all seen this in the field, it is great to see a reputable source verify some of what is seen in the field. I agree completely that it's generally not worth the risk. That said, there have been a few times where we have a 400' run and have decided to chance it without extenders.

Undisclosed #5

IPVMU Certified | 08/12/16 03:19pm

Thanks for the test. The subject is very close to my heart, and wallet.

Undisclosed Vs Undisclosed Ethernet Challenge - Who Will Go The Distance?

Maybe had this research come out first I wouldn't have been so incredulous.

Did you use a cable verifier? It would be interesting to see what a high-end one would say.

BTW and IMHO, the biggest limitation and variable when getting out past 500' is something you can't control and varies by mfr.

Namely the twist rate of each pair. They're always different to reduce local crosstalk, which has the effect of making each pair a different total length. This length variation is left up to the mfr.

Different lengths of each pair means different timings, after some point the differences between pairs becomes so great that the interface can no longer keep the signals synced and the link fails.

This is one of the main things that Ethernet extenders either a)regenerate (if inline) or b) tolerate if at the ends, using their own prop methods.

Ethan Ace

IPVM | In reply to Undisclosed #5 | 08/12/16 03:55pm

We did not use a verifier.

Undisclosed #5

IPVMU Certified | In reply to Ethan Ace | 08/13/16 08:42pm

Come to think of it a verifier might not provide much information, as opposed to a qualifier that would actually push packets and see how they do.

The verifier though is going to show FAIL right away on at least the length and probably most other metrics as well. And with these extreme lengths, it is probably legally required to notify the IEEE. ;)

Bruce Cisne

IPVMU Certified | 08/15/16 02:17pm

John your tests implies that longer cable lengths work even though you threw in the caveat to follow industry standards. Our field experience says different when it comes to multiple cameras connected to the same POE switch. We had a junior sales engineer do exactly that and almost cost us a customer when cameras started dropping out after several months of working. The more cameras on a switch increases the draw and, in time, will experienced camera failures.

The problem is when an inexperienced installer, after reading a report like yours, ignores the standards and thinks that he can regularly exceed cable length will eventually have failures. The TIA/EIA standard is there for a reason and for you to even suggest that it can be done doesn't mean it should.

Luis Carmona

IPVMU Certified | In reply to Bruce Cisne | 08/15/16 03:21pm

Speaking only for myself, Bruce, I'd have to disagree on your general point that IPVM would share responsibility of someone exceeding cable length specs. It is enough to through in there that people should "follow industry standard". Because that is a hard rule that the person doing the job should know. Someone doing a job and not knowing spec's should always be followed despite something being shown to work outside of specs, is poor training, knowledge or discipline on the technician/engineer side, not the fault of a website doing a test. I can show how you can install electrical lines without using safety gear, but it doesn't excuse someone who is supposed to be a properly trained and licensed electrician from not using safety equipment.

It's not IPVM fault if they show equipment can work beyond cable specs and someone uses that as excuse to always do it. They were told to follow specs and as a professional they should know better to do it.

Robb Larsen

In reply to Luis Carmona | 08/15/16 04:21pm

Agreed. I think the article shows, "Hey everyone, in case you're curious, this works; but definitely shouldn't be your standard."

Fun article, thanks for sharing.

Undisclosed #5

IPVMU Certified | In reply to Bruce Cisne | 08/15/16 05:25pm

The problem is when an inexperienced installer, after reading a report like yours, ignores the standards and thinks that he can regularly exceed cable length will eventually have failures.

Ok, but the problem is that the information is out there regardless of IPVM, so it becomes more of a question of who responsibly educates them in such matters than if they can be kept in the dark forever.

There's always that tech around with a little more "knowledge" (and a lot more imagination to fill the gaps) eager to "explain" things to the impressionable newbie, that must be considered. Peer-to-peer pressure can be tough to resist.

And inexperienced techs are naturally curious about such things; remember yourself at such a point in your own career... Moreover, some questioning, and even a little experimentation (strictly in-house), is a good sign of an analytical mind and attitude, and something you should nurture.

Setting boundaries is key; And although they may not always agree with you, if they respect you, you can have confidence that you wishes will be heeded, when they're on their own and faced with the inevitable on-site temptation.

Because it's too easy for them if they are not properly schooled in the consequences, to just walk out of there today with everything working fine, but then 9 months down the line you get the shocking news...

So have that "talk" with your techs today, before somebody else does, and by all means support that with factual information from sources like IPVM.

You'll reap the benefits of a more open and productive relationship with your tech.

Undisclosed Manufacturer #7

08/15/16 05:41pm

I'm glad they tested and posted. This won't change my answer to the question.

As a manufacturer, all I can tell them is we ONLY support a standard installation, using industry standards.

I know for a fact at 900' it doesn't work because I had a guy install a camera at that distance without understanding there is a distance limitation on Ethernet.

He felt we should have listed Ethernet distances on the product box.

justin kozak

08/16/16 03:00pm

Extending runs past standards, 100 meter rule is like driving with your feet. You can do it, but it's probably not a good idea.

Brandon Knutson

IPVMU Certified | In reply to justin kozak | 08/16/16 03:45pm

It's like doing 69mph in a 60mph zone, it works until it doesn't

Undisclosed #5

IPVMU Certified | In reply to Brandon Knutson | 08/16/16 04:49pm

Ma'am, is this your current MAC address?

Undisclosed Integrator #6

08/17/16 09:32am

Lost in all this is that there is at least one company claiming that their gears is (perdon the pun) geared to exceed the 100 meters limitation.

The thing is we should (and likely will ) see this more often. The introduction of extenders to go past the 100 meter limitation is frustrating to say the least: These are not cheap and don't always "play nice" with switches.

We are also of the advice that IPVM in no way encourages >100 meter lengths. The discussion is a good one: some gears do work past the 100 meter limitation. If someone is naive and inexperienced enough to see the article has an endorsement it cannot be attributed to IPVM.

Craig Mc Cluskey

08/17/16 06:35pm

Thank you for doing the test, and including lots of links in the article. That makes for a good reference for things in the future.

Geoff Collins

11/08/16 05:56am

Very interesting study/experiment.

As you rightly point out following the standards is the only way to go, by way of an example I recently discovered at a coal processing plant located in Australia, 4 IP PTZ cameras with CAT5e cable runs at 150m+ that were not working at all. These cameras had initially been working for a number of months then 1 by one they stopped.

David Johnson

11/08/16 02:46pm

I've seen this as well where a temp installation that marginally exceeded 100m (10m+) worked fine for a couple of months and then stopped completely. I'd be interested to figure out what process caused the failure by aging as the physical installation appeared fine. The cam/nvr combo worked fine with a short test jumper and identical new devices refused to work in the field. Had to install a field switch to shorten distance.

Robert Shih

Independent
In reply to David Johnson | 11/08/16 03:19pm

Judging from my personal experience and what some other members have been mentioning, I would like to request an update to this test with actual wattage draw numbers for the distances and grades of cable used.

Undisclosed Integrator #2

02/02/17 07:17pm

Same as other people here who did not run into an immediate problem going past the max distance, but problems developed later. We've had a couple cases of cameras just occasionally dropping off for no good reason and they'd come up after a switch reboot done remotely in the office. We even replaced a switch and a camera in one instance.

We finally got a tech out who knew what they were doing and in one case the cable was around 360ft and the other around 400ft.

Undisclosed Manufacturer #8

02/03/17 08:59am

Thanks for the test.
BTW. Network latency already included in your test?
Camera frame rate remain the same after using 300M cable??
Long distance should be drop frame and increase latency in theory.

Undisclosed Integrator #9

08/23/17 04:20pm

I'd like to see fluke cable test results for those long runs

Undisclosed

08/23/17 09:07pm

If you violate the cable length spec you're in trouble. Doesn't matter how many magic tools you use to test it. One screw-up and you'll get to explain it to the customer. Get your cables "certified" (means you burned a grand on a second vendor with actual testing gear who came in, did a test, and delivered an actual report.)

Undisclosed Integrator #2

In reply to Undisclosed | 08/23/17 09:32pm

I agree, Rod, but I think UND 9 was just asking out of curiosity, what the meter would say at the long run. Maybe things like impedance and voltage (POE).

Undisclosed Integrator #9

In reply to Undisclosed Integrator #2 | 08/23/17 09:34pm

there're much more than just impedance. I'd definitely recommend to use professional cable tester next time they conduct research like that.

IP Camera Long Distance Ethernet Test

Comments (38)

Undisclosed #1

Stephanie Guo

Undisclosed Integrator #6

Undisclosed #1

John Honovich

Undisclosed #1

german bravo

Undisclosed Manufacturer #8

Paul Charles

Undisclosed Integrator #2

Gert Molkens

Ethan Ace

Undisclosed Integrator #3

John Scanlan

Ethan Ace

Undisclosed Integrator #4

Undisclosed #5

Ethan Ace

Undisclosed #5

Bruce Cisne

Luis Carmona

Robb Larsen

Undisclosed #5

Undisclosed Manufacturer #7

justin kozak

Brandon Knutson

Undisclosed #5

Undisclosed Integrator #6

Craig Mc Cluskey

Geoff Collins

David Johnson

Robert Shih

Undisclosed Integrator #2

Undisclosed Manufacturer #8

Undisclosed Integrator #9

Undisclosed

Undisclosed Integrator #2

Undisclosed Integrator #9

Login to read this IPVM report.

Member Login

Related Reports

Recent Reports